Datasets:

eswardivi

/

books

like 0

e PDN connection with non-IP PDN type corresponds to PDU session type of Ethernet or unstructured, respectively. This is to ensure that the appropriate PDU session type will be used if the UE transfers back to the 5GS. 162 Chapter 1 1.4.2 5G Deployments Scenarios and Architecture Options This section describes 5G network architecture deployment options supporting stand-alone and non-stand-alone mode of operation. These deployment options were extensively dis- cussed in 3GPP and have been prioritized based on their viability and practicality. Each option provides NR access to sufficiently capable UEs either through direct access to gNB/ 5GC or indirectly via LTE interworking. 3GPP defined both stand-alone and non-stand- alone deployment configurations for NR in 3GPP Rel-15. A stand-alone NR deployment would not require an associated LTE network. The NR-capable UE could use random access to directly establish a radio link with a gNB and attach to the 5GC in order to use network services. The stand-alone NR deployment required a complete set of specifications from 3GPP for all entities and interfaces in the network, which was subsequently defined in 3GPP Rel-15 specifications. In stand-alone operation (shown as option 2 in Fig. 1.80), the network access procedures closely follow the LTE counterparts. The additional requirements mainly include broadcast of NR system information, which includes a minimum set of parameters and extended set of parameters where the former is periodically broadcast and comprises basic information required for initial access and the scheduling information for other system information; and the latter encompasses other system information that are not transmitted via the broadcast channel, which may either be broadcast or provisioned in a dedicated manner which can be triggered by the network configuration or upon request from the UE. In comparison to LTE system information broadcasting scheme, on-demand broadcasting is a new mechanism introduced in NR to deliver other system information by U

E request. For UEs in RRC_CONNECTED state, dedicated RRC signaling is used for the request and delivery of the other system information. For UEs in RRC_IDLE and RRC_INACTIVE states, making the request will trigger a random access procedure [16]. As an interim step for NR deployments, 3GPP has defined a set of non-stand-alone deploy- ment configurations using dual connectivity between NR gNB and LTE eNB (or ng-eNBs). Since initial NR networks will not have full coverage, dual connectivity can be used to combine the coverage advantage of the existing LTE networks with the throughput and latency advantages of the NR. However, it requires more complex UE implementations to allow simultaneous connections with both LTE and NR networks, potentially increasing the cost of the UEs. This will require more complex UE radio capabilities including the ability to simultaneously receive downlink transmissions from NR and LTE base stations on sepa- rate frequency bands. The non-stand-alone NR deployments use architectures where NR gNBs are associated with LTE eNBs and do not require separate signaling connections to the 5GC. These architectures are classified based on the control-plane and user-plane con- nections used between eNB, gNB, EPC, and 5GC. 5G Network Architecture LTE eNB Option 2 Option 3 LTE eNB ng-eNB Option 3A Option 4 ng-eNB ng-eNB Option 4A Option 5 ng-eNB ng-eNB Option 7 Option 7A Figure 1.80 NR deployment options 34]. The deployments based on architecture option 3 use EPC as the core network as shown in Fig. 1.80. In this case, S1-C control-plane interface for the UE is established between the LTE eNB and the EPC. The gNB acts as a secondary node connected to the master node represented by the eNB. Control-plane information is exchanged between the LTE eNB and the NR gNB, and no direct control-plane interface exists between the gNB and the EPC. User-plane bearers are supported between eNB and EPC over S1-U. In option 3A, the gNB also terminates user-plane bearers with the EPC. In option 3A, those g NB terminate

d S1-U bearers may further split and carried over the X2/Xn interface to the eNB and over the LTE air interface. The deployments based on architecture option 3 do not require interface with the 5GC, and allow service over the NR air interface with Uu (i.e., the interface between the UE and the serving gNB) and X2/Xn interfaces defined. As such, this is seen as the most likely architectural scenario for early NR deployments. From control-plane perspec- tive, there is only one RRC state in the UE, which is based on the LTE RRC protocols and there is only one control-plane connection toward the core network. 164 Chapter 1 The deployments based on architecture option 4 are essentially the opposite of option 3, with the gNB and the eNB representing the MCG and the SCG, respectively. The control- plane connection is established between the gNB and the 5GC over the NG-C, and the eNB exchanges control-plane information with gNB over Xn. In option 4A, direct user-plane bearers with the 5GC are terminated at the eNB. In architecture option 5, the ng-eNB (i.e., an LTE eNB compliant with LTE Rel-15 onward) is connected to the 5GC. The deployments based on architecture option 7 use the same topology as option 3, with the eNB acting as MCG and the gNB acting as SCG. The difference between the two options is that the 5GC serves as the core network instead of the EPC, requiring the eNB upgrade to ng-eNB interfaces with the 5GC. In this scenario, each radio node has its own RRC entity which can generate RRC PDUs to be sent to the UE. Note that RRC PDUs generated by the gNB (SN) can be transported via the LTE Uu interface or NR Uu interface to the UE, if configured. The eNB (MN) always sends the initial SN RRC configuration via MCG SRB (SRB1); however, subsequent reconfigurations may be transported via the MN or SN. Furthermore, the UE can be configured to establish an SRB with the SN (i.e., SRB3) to enable RRC PDUs for the SN to be sent directly between the UE and the SN. 1.5 Network Aspects of Mobility and Power Management Mobil

ity and power management continue to be the most important aspects of any new cel- lular standard to ensure seamless connectivity and sustainable power consumption of user terminals as well as overall energy efficiency of the network. In 4G/5G, the states of a UE with respect to mobility and connection establishment are described by NAS and RRC states. Fig. 1.81 shows and compares the EPS and 5GS/NG-RAN NAS and RRC states. There are three states shown in the EPS model, that is, EPS mobility management (EMM), EMM-DEREGISTERED RM-DEREGISTERED RM-REGISTERED ECM-IDLE CM-IDLE CM-CONNECTED RRC_IDLE RRC_IDLE EMM-REGISTERED ECM-CONNECTED RRC_CONNECTED RRC_CONNECTED EMM-REGISTERED RM-REGISTERED RRC_INACTIVE ECM-IDLE CM-IDLE RRC_IDLE RRC_IDLE EPS state model 5GS state model Figure 1.81 Comparison of mobility management states of EPS and 5GS [3]. 5G Network Architecture 165 EPS connection management (ECM), and RRC. The EMM and ECM states are managed by the core network, where the EMM state represents whether a UE is registered in the EPC and the ECM state indicates if NAS signaling connection between the UE and the MME is established. On the other hand, the RRC state is managed by E-UTRAN, and it represents whether there is a connection between the UE and the serving eNB. A UE in the ECM- CONNECTED state needs to be in the RRC_CONNECTED state, because a radio link con- nection is required in order to establish NAS signaling [3,16]. 1.5.1 Mobility Management In 5GS, the mobility management state of a UE can be either RM-REGISTERED or RM- DEREGISTERED depending on whether the UE is registered in 5GC, which is very similar to EMM-REGISTERED and EMM-DEREGISTERED states in EPC. Two connection man- agement states are used to reflect the NAS signaling connectivity of the UE with the AMF namely CM-IDLE and CM-CONNECTED. A UE in CM-IDLE state has no NAS signaling connection established with the AMF over N1. A UE in CM-CONNECTED state has a NAS signaling connection with the AMF over N1. A NAS signaling connection uses an RRC connecti

on between the UE and the NG-RAN and an NGAP UE association between the access network and the AMF for 3GPP access. NR RRC protocol states consist of three states, where in addition to RRC_IDLE and RRC_CONNECTED states, a third state has been introduced, RRC_INACTIVE, as an intermediate state prior to transition to RRC_IDLE state in order to save UE power and to allow fast connection setup [3,16]. As shown in Fig. 1.81, in EPS, when a UE is in the RRC_CONNECTED state, the serving eNB evaluates the received signal strength measurements from the UE and initiates a hand- over procedure when the UE's received signal strength goes below a threshold. However, in the RRC_IDLE state, where the eNB is not aware of the UE's location, the UE decides whether to camp on the current cell or to reselect a neighboring target cell based on received signal strength measurements. This procedure is referred to as cell reselection. In EPS, the mobility procedures, that is, as handover in the RRC_CONNECTED state and cell reselection in the RRC_IDLE state, are not flexible, whereas in 5GS, the core network is able to flexibly control whether to perform a handover or cell reselection for a UE in CM- CONNECTED state [3,16]. In EPC, the location of a UE is tracked by the MME. The granularity of location tracking is different depending on the RRC state of the UE. In the RRC_CONNECTED state, the UE's location is tracked at the cell level, whereas in the RRC_IDLE state, its location is tracked at the tracking area level, which is a set of cells belonging to a paging group that simulta- neously transmit paging messages. Similarly, 5GC can track the location of a UE at the tracking area level in the CM-IDLE state, whereas the UE's location is known at the level of the serving cell to the core network in the CM-CONNECTED state. In 5GC, when a UE 166 Chapter 1 registers with the network over the 3GPP access, the AMF allocates a set of tracking areas in TAI list to the UE. The AMF takes into account various information (e.g., mobility pat- tern a

nd allowed/non-allowed area, etc.), when it allocates registration area to the UE, that is, the set of tracking areas in TAI list. In 5GS; however, NG-RAN also supports the loca- tion tracking for the UEs in RRC_INACTIVE state. In that state, the core network knows that the UE is somewhere within the NG-RAN, but the NG-RAN node needs a new location tracking functionality to determine the exact location of the UE because the connection between the UE and the NG-RAN node is not active. In EPS, when downlink traffic for a UE in the RRC_IDLE state arrives at the SGW, the MME performs a paging procedure based on the detected location of the UE. However, 5GS supports two types of paging, namely core-network-initiated paging and access- network-initiated paging. The UE in RRC_IDLE and RRC_INACTIVE states may use DRX in order to reduce power consumption. While in RRC_IDLE, the UE monitors 5GC- initiated paging, in RRC_INACTIVE, the UE is reachable via RAN-initiated paging and 5GC-initiated paging. The RAN and 5GC paging occasions overlap and the same paging mechanism is used. In core-network-initiated paging, the default paging procedure is requested by the core network when the UE is in the CM-IDLE state. The newly introduced RAN-initiated paging mode is used for the UEs in the RRC_INACTIVE state. Since a UE in the RRC_INACTIVE state is in the CM-CONNECTED state (see Fig. 1.81), the core net- work simply forwards the data or the signaling messages to the corresponding RAN when data or signaling messages arrive for the UE. Therefore, RAN itself generates the paging message and performs paging to find the updated location of the UE, and then sends the data or signaling messages to the UE. The 5GC can transmit additional assistance informa- tion for RAN paging [16]. Service area restriction also known as mobility-on-demand, which defines areas where the UE may or may not initiate communication with the network, is a concept to selectively support mobility of devices on a need basis. It includes supporting UE's mobility at

a cer- tain level classified as mobility restriction and mobility pattern (or mobility level). The for- mer addresses mobility restriction in terms of allowed, non-allowed, and forbidden areas. The minimum granularity of the area is at tracking area level. In the allowed area, UE can communicate through the control or user planes. The UE cannot send service request and session management signaling in the non-allowed area. However, periodic registration update is possible. It can also respond to the paging messages from the core network. Furthermore, emergency calls or multimedia priority services are allowed. In the forbidden area, UE is not allowed to have any communication with the network except for the emer- gency services. The mobility pattern is used as a concept to describe the expected mobility of UE in 5GC. Mobility pattern may be used by the AMF to characterize and optimize the UE mobility. The AMF determines and updates mobility pattern of the UE based on sub- scription of the UE, statistics of the UE mobility, network local policy, and the UE-assisted 5G Network Architecture information, or any combination of these parameters. The statistics of the UE mobility can be history-based or expected UE moving trajectory. The UE mobility pattern can be used by the AMF to optimize mobility support provided to the UE [16]. This procedure is used for the case where the UE moves from one gNB-DU to another gNB-DU within the same gNB-CU during NR operation. The Internet of things is an important 5G service category. IoT devices mostly send mobile- originated data. For this type of devices, the mobile-originated only mode is defined in 5GS where the core network determines whether to apply the mobile-originated only mode to a UE during the registration procedure based on the UE subscription data and the network policy. The mobile-originated only mode is allocated to a UE, which does not require mobile-terminated traffic. Therefore, the UE in mobile-originated-only mode does not listen to the paging messages. The cor

e network does not need to manage the UE's location while it is registered in the 5GC. For optimization, the core network may decide to deregister the UE after the mobile-originated data communication is finished, without transferring the UE's state into the CM-IDLE state in the RM-REGISTERED state, because most functions supported in the CM-IDLE state are not relevant to the UE in mobile-originated only mode, for example, UE location tracking and reachability management. In such cases, the UE needs to perform attach procedure whenever the mobile-originated data transmission is nec- essary to communicate with the core network. A UE receives services through a PDU session, which is a logical connection between the UE and the data network. In 5GS, various PDU session types are supported, for example, IPv4, IPv6, Ethernet, etc. Unlike EPS, where at least one default session (i.e., default EPS bearer) is always created while the UE attaches to the network, 5GS can establish a session when service is needed irrespective of the attachment procedure of UE, that is, attachment without any PDU session is possible. 5GS also supports UE establishing multiple PDU ses- sions to the same data network or to different data networks over a single or multiple access networks including 3GPP and non-3GPP access. The number of UPFs for a PDU session is not specified. The deployment with at least one UPF is essential to serve a given PDU ses- sion. For a UE with multiple PDU sessions, there is no need for a convergence point like SGW in the EPC. In other words, the user-plane paths of different PDU sessions are completely disjoint. This implies that there is a distinct buffering node per PDU session for the UE in the RRC_IDLE state. In order to ensure slice-aware mobility management when network slicing is supported, a slice ID is introduced as part of the PDU session information that is transferred during mobility signaling. This enables slice-aware admission and conges- tion control [34]. LTE-NR interworking is important for stand-a

lone mode of operation between LTE and NR unlike for dual connectivity where there is simultaneous transmission across both RATs most of the time. Interworking between LTE and NR is not expected to be significantly 168 Chapter 1 E-UTRA Handover RRC_CONNECTED NR RRC_CONNECTED RAN paging/RAN update/ connection resume Connection establishment Connection establishment and release NR RRC_INACTIVE and release RAN-based notification area Cell reselection update E-UTRA RRC_IDLE Cell reselection NR RRC_IDLE Figure 1.82 LTE-NR mobility state diagram [16,35]. different from what is defined in LTE specifications for interworking with other 3G net- works. The inter-RAT mobility is expected to be supported both in idle mode as well as in the connected mode. Fig. 1.82 illustrates possible mobility scenarios across LTE and NR. RRC_INACTIVE is a new RRC state in NR, in addition to RRC_IDLE and RRC_CONNECTED. It is a state where a UE remains in CM-CONNECTED and is able to move within an area configured by NG-RAN (i.e., RAN-based notification area or RNA) without notifying NG-RAN. The RNA can cover a single cell or multiple cells. In RRC_INACTIVE, the last serving NG-RAN node retains the UE context and the UE- associated NG connection with the serving AMF and UPF. The UE notifies the network via RAN-based notification area update procedure, if it has moved out of the configured RNA. If the last serving gNB receives downlink data from the UPF or downlink signaling from the AMF while the UE is in RRC_INACTIVE state, it pages the UE in the cells correspond- ing to the RNA and may forward the paging message to the neighbor gNB(s) if the RNA includes cells of neighbor gNB(s) [16]. Connected mode mobility is enabled using Xn inter- face between LTE eNB and NR gNB where both eNB and gNB are connected to 5G core network. S1/N2-based handover is supported when LTE eNB is connected to EPC and NR gNB is connected to 5GC. Xn and core network handover in 5GC, when both eNB and gNB are connected to 5GC, is transparent to the UE. Seamless handove

r is possible based on tight interworking between radio access technologies when anchored at 5GC. 1.5.2 Network-Controlled Power Management Power management schemes are important to sustaining UE services and prolonged UE bat- tery life. In this section, we discuss the power management schemes and strategies used by 5G networks to optimize UE power consumption while satisfying user expectations and 5G Network Architecture 169 QoS requirements of the applications. During the normal operation, a 5G UE can be in one of the three states of connected, inactive, or idle as explained in the previous section. If there is no data to be transmitted/received, the UE stays in the energy-efficient idle state. In contrast, the connected state is the energy-consuming state, as the UE needs to continuously monitor the link quality of the serving and neighboring cells and to provide periodic reports on the quality and status of the radio link. Mobile devices in connected mode may be configured with discontinuous reception (DRX) for power saving purposes. The parameters of DRX configuration can be optimized to either maximize power saving or minimize latency performance based on the UE's active applica- tions/services. The DRX cycles are broadly configurable to support a wide range of services with different requirements in terms of power consumption and accessibility delays. The implementation of DRX allows the UE to avoid frequently monitoring the physical down- link control channel, except during specific time intervals configured by higher layers. A typical DRX cycle can be divided into an active time interval and an inactive period. During active time interval, the UE is awake and performs continuous reception, while the inactivity timer has not expired. During this time, the UE is performing continuous recep- tion while waiting for a downlink transmission. The UE then enters the inactive period, if there is no traffic activity longer than the inactivity timer duration, after expiration of on- duration, or if the UE receives

a MAC control message and is instructed to enter the DRX mode. When the DRX is configured and the UE is in an active state while in the on- duration interval (see Fig. 1.83), the UE monitors the relevant control channels to detect any downlink allocations and to receive pending transmissions from the serving base sta- tion. If no allocations including paging messages are detected within the on-duration, the UE will enter the inactive interval and will follow the DRX configuration to wake up in the next DRX cycle. During inactivity period, the UE first follows the short DRX cycle, if con- figured, and starts the DRX short cycle timer. After the short cycle timer expires, the UE follows the long DRX cycle. If no short DRX cycle is configured, the UE directly enters the long DRX cycle. It is necessary to optimize the DRX parameters according to the QoS requirements of vari- ous services such as VoIP, web browsing and video streaming, where each has a different traffic model and specific QoS requirements, which may significantly impact the configura- tion of DRX parameters. The services with low delay requirement can be dealt with by acti- vating the UE more frequently to monitor downlink allocations with short DRX cycle setting. While delay-tolerant services can achieve high energy-saving gains by waking up the UE to monitor downlink allocations with long DRX cycle configuration. Therefore, the best trade-off between power saving and responsiveness of the UEs should be made to opti- mally configure DRX parameters. 3GPP releases define the DRX configuration per UE. When multiple data bearers are established, DRX is enabled only when all the data bearers met their corresponding DRX inactivity timer condition, and the shortest DRX cycles 170 Chapter 1 DRX concept UE traffic Inactivity Traffic timer Traffic UE state Active DRX sleep DRX sleep Active On-duration On-duration On-duration DRX cycle DRX sleep DRX sleep DRX cycle DRX cycle DRX cycle LTE DRX Connected Connected C-DRX Connection Data/ signaling User-inactivity

timer Connected longer DRX opportunity period setup transmit/ NR DRX receive Connected Connected Data/signaling transmit/receive Connected shorter DRX opportunity period Figure 1.83 DRX concept and comparison of LTE and NR DRX schemes (example) 16,82]. among all the data bearers are followed. This solution is simple and effective for non-CA sce- narios. In CA scenarios, the UE may operate over several component carriers and supports separate RF transceivers for each RF carrier. The baseline UE-specific DRX mechanism is no longer suitable to achieve higher energy-saving gains since the same DRX setting has to be configured for all component carriers according to the delay requirement of all applications running on the UE. The UE has to wake up the RF circuitries for all RF carriers at the same time to monitor possible downlink allocations. When bandwidth adaptation is configured in NR, the UE only has to monitor the downlink control channel on the [single] active band- width part, thus it does not have to monitor downlink control channels on the entire downlink frequency bands used in the cell. A bandwidth part inactivity timer (independent of DRX inactivity-timer described earlier) is used to switch the active bandwidth part to the default one, that is, the timer is restarted upon successful downlink control channel decoding and switching to the default bandwidth part occurs upon its expiration (see Fig. 1.83). 5G Network Architecture 171 The percentage of time spent in the connected and idle states depends on a number of parameters controlled by the network including paging occasions, DRX cycles, user- inactivity timer, etc. The user-inactivity timer determines how long the UE stays in the connected state after it receives or transmits the last data packet. When the timer expires, the gNB releases the RRC connection and the UE immediately transitions to the idle state. The shorter the user-inactivity timer, the longer the UE battery life. However, if a new packet arrives the gNB queue shortly after the UE transi

tions to the idle state, the core network has to page the UE with network and radio signaling, causing extra service latency to transition to the connected state. In other words, the length of the user- inactivity timer determines a trade-off between UE energy consumption and connection latency as well as network control signaling overhead. Whenever the latency requirement can be relaxed, the DRX can provide further power savings. This will reduce the active duty cycle for downlink control channel monitoring, and if cross-slot scheduling is config- ured, then data channel reception is only needed when UE data is present. With this con- figuration the UE modem/transceivers may spend approximately 90% of time in a low- power sleep mode, but the penalty would be the substantially increased latency. The UE in RRC_IDLE and RRC_INACTIVE states may use DRX in order to reduce power con- sumption. While in RRC_IDLE, the UE monitors 5GC-initiated paging, in RRC_INACTIVE the UE is reachable via RAN-initiated paging and 5GC-initiated paging. The RAN and 5GC paging occasions overlap and same paging mechanism is used. The UE monitors one paging occasion per DRX cycle to receive the paging message. Paging DRX cycle length is configurable and a default DRX cycle for core-network-initiated pag- ing is sent via system information. A UE-specific DRX cycle for core-network-initiated paging can be configured via UE dedicated signaling. The NG-RAN can configure a UE with a DRX cycle for RAN-initiated paging, which can be UE specific. The number of paging occasions in a DRX cycle is configured and signaled via system information. A network may assign UEs to the paging occasions based on UE identities when multiple paging occasions are configured in the DRX cycle. When DRX is configured, the UE does not have to continuously monitor downlink control channels. If the UE detects a rele- vant downlink control channel, it stays awake and starts the inactivity timer. The DRX mechanism is characterized by several parameters such as on-duration

, which is the time interval that the UE waits for, after waking up, to receive possible downlink con- trol channels; inactivity-timer, which measures the duration of time from the last successful detection that the UE waits to successfully decode a downlink control channel. If the UE fails to detect a relevant downlink allocation, it goes back to sleep mode; retransmission- timer measures the duration of time when a HARQ retransmission is expected; and cycle, which specifies the periodic repetition of the on-duration followed by a possible period of inactivity [16]. 172 Chapter 1 1.6 Quality-of-Service Framework We begin this section with a review of QoS framework in LTE to set the stage for the 5G QoS framework. As shown in Fig. 1.84, in EPC, the user traffic is classified into different SDFs each associated with different QoS classes based on the type of the service that is being provided through the SDFs. Different QoS rules are then applied to each SDF. Since SDFs are delivered through EPS bearers in an LTE network, the EPS bearer QoS has to be controlled in a way that SDF QoS is maintained. In an LTE network, the user traffic (IP flows or IP packets) is classified into SDF traffic and EPS bearer traffic. An SDF refers to a group of IP flows associated with a service that a user is utilizing, whereas an EPS bearer refers to IP flows of aggregated SDFs that have the same QoS class. The SDF and EPS bearers are detected by matching the IP flows against the packet filters, that is, SDF tem- plates for SDFs or traffic flow templates (TFTs) for EPS bearers. These packet filters are preconfigured by network operators in accordance with their policy and each of them typi- cally consists of 5-tuple (source IP address, destination IP address, source port number, des- tination port number, and protocol ID). In other words, in LTE network, IP flows with the same service characteristics that match the packet filters of an SDF template are designated to an SDF. SDFs that match the packet filters of a TFT are mapped to an

EPS bearer, in order to be delivered to the UE. SDFs with the same QoS class are delivered, as aggregated, through an EPS bearer, whereas the ones with different QoS class are delivered through dif- ferent EPS bearers. In LTE, there are two types of EPS bearers, default and dedicated. When a UE attaches to the LTE network, an IP address is assigned for PDN connection and a default EPS bearer is established at the same time [58,59]. In an LTE network, the QoS parameters are defined at service and bearer levels. SDF QoS parameters are service-level QoS parameters, whereas EPS bearer QoS parameters are bearer-level QoS parameters. Service level and bearer level are also called as SDF level and SDF aggregate level, respectively. An SDF aggregate refers to a group of SDFs which have PDN connection (EPS session) UE IP address Default EPS bearer SDF1 QoS policy IP flow 1 SDF2 QoS policy IP flow 2 IP flow 3 SDF3 QoS policy Dedicated EPS bearer IP flow 4 SDF4 QoS policy IP flow 5 Figure 1.84 QoS architecture and process in LTE [59]. 5G Network Architecture 173 the same QCI and ARP values and belong to one EPS session. Both QCI and ARP are the basic QoS parameters applied to all SDFs and EPS bearers. The QoS class identifier (QCI) is particularly important because it serves as a reference that indicates the performance characteristics of SDFs and EPS bearers. In addition to these two basic parameters, there are other QoS parameters, such as GBR, maximum bit rate (MBR), and aggregated maxi- mum bit rate (AMBR) that specify the bandwidth (or bit rate) characteristics of SDFs and EPS bearers. The SDF and EPS bearer QoS parameters are as follows: SDF QoS parameters (QCI, ARP, GBR, and MBR) and EPS bearer QoS parameters (QCI, ARP, GBR, MBR, APN-AMBR, and UE-AMBR). The QCI and ARP are applied to all EPS bearers. An EPS bearer is classified as a GBR bearer or a non-GBR bearer depending on the resource type specified by its QCI. A default bearer must be non-GBR, while a dedicated bearer can be either GBR or non-GBR. Other than QCI

and ARP, there are other QoS parameters for EPS bearers including MBR and GBR indicating the bandwidth (or bit rate) of an EPS bearer, and AMBR indicating the total bandwidth of multiple EPS bearers. The MBR and GBR are the maximum and the guaranteed bandwidths of an EPS bearer, respectively, and AMBR is the maximum total bandwidth of multiple EPS bearers [16,59]. Fig. 1.85 illustrates the QoS parameters applied to SDFs and EPS bearers. In this figure, the UE is connected to two PDNs. The UE has two IP addresses: IP address 1 assigned by PGW-1 for use in PDN-1, and IP address 2 assigned by PGW-2 for use in PDN-2. The UE has one default bearer and two dedicated bearers established for each PDN. The IP flows (user traffic) are filtered into SDFs in the PGW by using SDF templates. There are two groups of SDFs, each received from PDN-1 and PDN-2. For these SDFs, network resources are allocated and packet forwarding is treated according to the QoS rules set in the PGW. The SDFs are then mapped to EPS bearers based on their specified QCI and ARP. In case of PDN-1, as shown in the figure, the SDFs 1 and 2 are mapped to the default bearer, SDFs 3 and 4 are mapped to the non-GBR dedicated bearer, and SDF 5 is mapped to the GBR dedicated bearer, all forwarded to the UE. Such traffic mapped from SDF to EPS bearer is defined by using traffic filter template. All user traffic is subject to the EPS bearer QoS while being delivered through the EPS bearers. All non-GBR bearers associated with a PDN are controlled by the maximum APN-AMBR that they share while the ones associated with a UE are controlled by the maximum UE-AMBR that they share. In LTE, all QoS parameters for SDFs are provisioned by policy and charging rules func- tion of the EPC [59]. The NG-RAN general QoS framework, both for NR connected to 5GC and for LTE con- nected to 5GC scenarios, is depicted in Fig. 1.86. For each UE, 5GC establishes one or more PDU sessions and the NG-RAN establishes one or more data radio bearers per PDU session. The NG-RAN maps packets

belonging to different PDU sessions to different DRBs and establishes at least one default DRB for each PDU session. The NAS-level packet filters in the UE and in the 5GC associate uplink and downlink packets with QoS flows. PDN connection 1 (EPS session 1) PGW-1 Resource type QoS parameters of EPS bearer address Dedicated EPS bearer (PDN-1) MBR (DL/UL) GBR (DL/UL) MBR (DL/UL) GBR (DL/UL) MBR (DL/UL) Dedicated EPS bearer (PDN-1) Non-GBR APN-AMBR MBR (DL/UL) (DL/UL) MBR (DL/UL) Default EPS bearer (PDN-1) Non-GBR MBR (DL/UL) UE-AMBR (DL/UL) Default EPS bearer (PDN-2) Non-GBR MBR (DL/UL) APN-AMBR (DL/UL) MBR (DL/UL) Dedicated EPS bearer (PDN-2) Non-GBR MBR (DL/UL) MBR (DL/UL) GBR (DL/UL) Dedicated EPS bearer (PDN-2) MBR (DL/UL) GBR (DL/UL) MBR (DL/UL) GBR (DL/UL) PGW-2 PDN connection 2 (EPS session 2) Figure 1.85 QoS parameters in LTE 59]. 5G Network Architecture 175 NG-RAN PDU session Radio bearer NG-U tunnel QoS flow QoS flow Radio bearer QoS flow Radio interface Figure 1.86 QoS framework in NR [16]. The AS-level mapping rules in the UE and in the NG-RAN associate uplink and downlink QoS flows with DRBs. At NAS level, the QoS flow is the finest granularity for QoS differentiation in a PDU session. A QoS flow is identified within a PDU session by a QFI transferred in encapsulated format over NG-U. NG-RAN and 5GC ensure quality of service (e.g., reliability and maximum toler- able delay) by mapping packets to appropriate QoS flows and DRBs. There is a two-step mapping of IP-flows to QoS flows (NAS level) and from QoS flows to DRBs (AS level). At NAS level, a QoS flow is characterized by a QoS profile which is provided by 5GC to NG- RAN as well as a set of QoS rule(s) which are provided by 5GC to the UE. The QoS profile is used by NG-RAN to determine the treatment on the radio interface while the QoS rules define the mapping between uplink user-plane traffic and QoS flows in the UE. A QoS flow may either be GBR or non-GBR depending on its profile. The QoS profile of a QoS flow contains QoS parameters, that is, a 5G Q

oS identifier (5QI) and an ARP. In case of a GBR QoS flow, the QoS parameters are guaranteed flow bit rate (GFBR) and maximum flow bit rate (MFBR) for uplink and downlink. In case of non-GBR, the QoS parameters include the newly defined reflective QoS attribute (RQA). The RQA, when included, indicates that some and not necessarily all traffic carried on this QoS flow is subject to reflective QoS at NAS level. At AS level, the DRB defines the packet treat- ment on the radio interface. A DRB serves packets with the same packet forwarding treat- ment. Separate DRBs may be established for QoS flows requiring different packet 176 Chapter 1 forwarding treatments. In the downlink, the NG-RAN maps QoS flows to DRBs based on NG-U marking (QFI) and the associated QoS profiles. In the uplink, the UE marks uplink packets over the radio interface with the QFI for the purposes of marking forwarded packets to the core network. When reflective QoS is used, a 5G UE can create a QoS rule for the uplink traffic based on the received downlink traffic without generating control-plane sig- naling overhead, as shown in Fig. 1.87 [16]. In the uplink, the NG-RAN may control the mapping of QoS flows to DRB in two different ways. Reflective mapping where for each DRB the UE monitors the QFI(s) of the downlink packets and applies the same mapping in the uplink, that is, for a DRB, the UE maps the uplink packets belonging to the QoS flows(s) corresponding to the QFI(s) and PDU session observed in the downlink packets for that DRB. To enable reflective mapping, the NG- RAN marks downlink packets over the air interface with QFI. In addition to the reflective mapping, the NG-RAN may configure an uplink QoS flow to DRB mapping via RRC sig- naling. The UE always applies the latest update of the mapping rules regardless of whether it is performed via reflective mapping or explicit configuration. For each PDU session, a default DRB is configured. If an incoming uplink packet matches neither an RRC config- ured nor a reflective QoS flow ID to DRB ma

pping, the UE maps that packet to the default DRB of the PDU session. Within each PDU session, it is up to NG-RAN to decide how to map multiple QoS flows to a DRB. The NG-RAN may map a GBR flow and a non-GBR flow, or more than one GBR flow to the same DRB. The time when a non-default DRB between NG-RAN and UE is established for QoS flow can be different from the time when the PDU session is established. It is up to NG-RAN to decide when non-default DRBs are established. In dual connectivity scenarios, the QoS flows belonging to the same PDU ses- sion can be mapped to different bearer types and, consequently, there can be two different SDAP entities configured for the same PDU session, that is, one for MCG and another one for SCG [16]. 3-Drive and insert 2-N3 marking in 1-RQI marking UE packet filters packet filter encapsulation header downlink packet for applying uplink QoS RQI 1 IP packet (III) QoS flows Radio bearers NG-RAN Server 4-Uplink packet for the same QoS flow Figure 1.87 Illustration of the reflective QoS concept (example) [82]. 5G Network Architecture 177 As we mentioned earlier, the 5G QoS model is based on QoS flows. The 5G QoS model supports both QoS flows that require GBFR QoS flows and QoS flows that do not require guaranteed flow bit rate (non-GBFR QoS flows). The 5G QoS model also supports reflec- tive QoS. The QoS flow is the finest granularity to differentiate QoS classes in the PDU ses- sion. A QFI is used to identify a QoS flow in the 5G system. User-plane traffic with the same QFI within a PDU session receives the same traffic forwarding treatment (e.g., sched- uling and admission control). The QFI is carried in an encapsulated header format on N3 (and N9), that is, without any changes to the end-to-end packet header. The QFI is used for all PDU session types and is unique within a PDU session. The QFI may be dynamically assigned or may be equal to the 5QI. Monitoring of user-plane traffic (e.g., MFBR enforce- ment) is not considered as QoS differentiation and is done by UPFs on an SDF-lev

el basis. Within the 5GS, a QoS flow is controlled by the SMF and may be preconfigured, estab- lished via the PDU session establishment procedure, or the PDU session modification proce- dures. Any QoS flow is characterized by a QoS profile provided by the SMF to the AN via the AMF over N2 reference point or is preconfigured in the AN; one or more QoS rule(s) which can be provided by the SMF to the UE via the AMF over N1 reference point and/or derived by the UE by applying reflective QoS control; and one or more SDF templates pro- vided by the SMF to the UPF. In 5GS, the QoS flow of the default QoS rule is required to be established for a PDU session and to remain active throughout the lifetime of the PDU session. The QoS flow of the default QoS rule is a non-GBR QoS flow [3]. A QoS flow may be either GBR or non-GBR depending on its QoS profile, which contains the corresponding QoS parameters. For each QoS flow, the QoS profile includes the follow- ing parameters [3]: 5G QoS identifier (5QI) For each non-GBR QoS flow, the QoS profile may include RQA For each GBR QoS flow, the QoS profile includes GFBR and MFBR for uplink and downlink In case of a GBR QoS flow only, the QoS parameters may include notification control and maximum packet loss rate for uplink and downlink Each QoS profile has one corresponding QFI, which is not included in the QoS profile itself. The 5QI value may indicate that a QoS flow has signaled QoS characteristics, and if so, the QoS characteristics are included in the QoS profile. The UE performs classification and marking of uplink user-plane traffic, that is, the association of uplink traffic to QoS flows based on the QoS rules. These QoS rules may be explicitly pro- vided to the UE via the PDU session establishment/modification procedure, preconfigured in the UE or implicitly derived by the UE by applying reflective QoS. A QoS rule contains a QoS rule identifier which is unique within the PDU session, the QFI of the associated QoS flow and 178 Chapter 1 except for the default QoS rule a pac

ket filter set for uplink and optionally for downlink and a precedence value. Furthermore, for a dynamically assigned QFI, the QoS rule contains the QoS parameters relevant to the UE (e.g., 5QI, GBR and MBR, and the averaging window7. There are more than one QoS rule associated with the same QoS flow. A default QoS rule is required for each PDU session and associated with the QoS flow of the default QoS rule. The default QoS rule is the only QoS rule of a PDU session that may contain no packet filter set in which case, the highest precedence value is used. If the default QoS rule does not contain a packet filter set, the default QoS rule defines the treatment of packets that do not match any other QoS rules in a PDU session. If the default QoS rule does not contain a packet filter, the reflective QoS is not applied to the QoS flow of the default QoS rule. The SMF performs binding of SDFs to QoS flows based on the QoS and service require- ments. The SMF assigns the QFI for a new QoS flow and derives its QoS profile from the information provided by the PCF. The SMF provides the QFI along with the QoS profile and a transport-level packet marking value for uplink traffic to the AN. The SMF further provides the SDF template, that is, the packet filter set associated with the SDF received from the PCF together with the SDF precedence value, the QoS-related information, and the corresponding packet marking information, that is, the QFI, the transport level packet mark- ing value for downlink traffic and optionally the reflective QoS indication to the UPF enabling classification, bandwidth enforcement and marking of user-plane traffic. For each SDF, when applicable, the SMF generates a QoS rule. Each of these QoS rules contain the QoS rule identifier, the QFI of the QoS flow, the packet filter set of the uplink part of the SDF template, and optionally the packet filter set for the downlink part of the SDF tem- plate, as well as the QoS rule priority value set to the SDF precedence value. The QoS rules are then provided t

o the UE. The principle of classification and marking of user-plane traf- fic and mapping of QoS flows to AN resources is illustrated in Fig. 1.88 [3]. A packet filter set is used in the quality of service rules or service data flow template to identify a quality of service flow. The packet filter set may contain packet filters for the downlink direction, the uplink direction, or packet filters that are applicable to both directions. There are two types of packet filter set, that is, IP packet filter set and Ethernet packet filter set, corresponding to those protocol data unit session types. The quality of service rule precedence value and the service data flow template precedence value determine the order in which a quality of service rule or a service data flow template, respectively, is evaluated. The evaluation of the quality of service rules or service data flow templates is performed in the increasing order of their precedence value. The averaging window is defined only for guaranteed bit rate quality of service flows and represents the duration over which the guaranteed flow bit rate and maximum flow bit rate are calculated. The averaging window may be signaled along with 5 QoS identifiers to the access network and user-plane function, and if it is not received, a predefined value will be applied. Policy control function provides policy framework incorporating network slicing, roaming and mobility man- agement and is equivalent to policy and charging rules function in evolved packet core. Application/Service Layer Data packets from applications QoS Rules (Mapping UL packets to QoS flows and apply QoS flow marking) QoS Flow (All packets marked with the same QoS flow ID) Mapping QoS flows to AN Resources Packet Detection Rule (PDR) (classify packets for QoS AN Resources PDU Session flow marking and other AN Resources actions) Figure 1.88 Classification and user-plane marking of QoS flows and mapping to access network resources [3]. 180 Chapter 1 In the downlink direction, the incoming data packets are classi

matches the uplink packet). The UE uses the QFI in the corresponding matching QoS rule to bind the uplink packet to a QoS flow. The UE then binds QoS flows to the AN resources. If no match- ing QoS rule is found, the UE discards the uplink data packet. The UPF maps user-plane traffic to QoS flows based on the PDRs [3] and performs session-AMBR enforcement and PDU count- ing for charging. The UPF further transmits the PDUs of the PDU session in a single tunnel between 5GC and access network. The UPF includes the QFI in the encapsulation header and it may further include an indication for reflective QoS activation. The UPF performs transport-level packet marking in downlink, which is based on the 5QI and ARP of the associated QoS flow. The access network maps PDUs from QoS flows to access-specific resources based on the QFI and the associated 5G QoS characteristics and parameters. The UE uses the stored QoS rules to determine mapping between uplink user-plane traffic and QoS flows. The UE marks the uplink PDU with the QFI of the QoS rule containing the matching packet filter and transmits the uplink PDUs using the corresponding access- specific resource for the QoS flow based on the mapping provided by RAN. The access net- work transmits the PDUs to UPF. The access network includes the QFI value, in the encap- sulation header of the uplink PDU when sending an uplink packet from the access network to the core network. The access network performs transport-level packet marking in the uplink, which may be based on the 5QI and ARP of the associated QoS flow. The UPF veri- fies whether QFIs in the uplink PDUs are aligned with the QoS rules provided to the UE or implicitly derived by the UE (e.g., in case of reflective QoS) and performs session-AMBR enforcement and counting of packets for charging. The 5G QoS parameters can be further described in detail as follows [3]: 5QI is a scalar that is used as a reference to 5G QoS characteristics, that is, access node-specific parameters that control QoS forwarding treatment for

the QoS flow (e.g., scheduling weights, admission thresholds, queue management thresholds, link layer 5G Network Architecture Table 1.4: 5QI to QoS characteristics mapping [3]. Resource Default Packet Delay Packet Default Default Example Services Value Priority Budget (ms) Error Maximum Data Averaging Level Burst Volume Window (Bytes) Conversational Voice Conversational Video (Live Streaming) Real-time Gaming, V2X Messages, Electricity Distribution, Process Automation Non-Conversational Video (Buffered Streaming) Mission-critical User-plane Push-to-Talk Voice Non-mission-critical User- plane Push-to-Talk Voice Mission-critical Video Non-GBR IMS Signaling Video (Buffered Streaming) TCP-based (e.g., www, e-mail, chat, ftp, p2p file sharing, progressive video, etc.) Voice, Video (Live Streaming) Interactive Gaming Video (Buffered Streaming) TCP-based (e.g., www, e-mail, chat, ftp, p2p file sharing, progressive video, etc.) Mission-critical Delay Sensitive Signaling Mission-critical Data V2X Messages Low-latency eMBB Applications, Augmented Reality Delay Discrete Automation Critical Discrete Automation Intelligent Transport Systems Electricity Distribution protocol configuration, etc.). The 5QI values have one-to-one mapping to standardized combination of 5G QoS attributes, as shown in Table 1.4. The 5G QoS characteristics for preconfigured 5QI values are preset in the AN, whereas the dynamically assigned 5QI values are signaled as part of the QoS profile. 182 Chapter 1 The ARP contains information about the priority level, the preemption capability and the preemption susceptibility. The priority level defines the relative importance of a resource request. This allows deciding whether a new QoS flow may be accepted or should be rejected in case of resource limitation, which is typically used for admission control of GBR traffic. It may also be used to decide which of the existing QoS flows to preempt in limited resource scenarios. The range of the ARP priority level is from 1 to 15, with 1 as the highest priority le

vel. The preemption capability information defines if an SDF may use the resources that were already assigned to another SDF with a lower priority level. The preemption susceptibility information defines whether an SDF may lose the resources assigned to it in order to admit an SDF with higher priority level. The RQA is an optional parameter, which indicates that certain traffic carried in this QoS flow is subject to reflective QoS. The access network enables the transfer of the RQI for AN resource corresponding to this QoS flow when RQA is signaled. The RQA may be signaled to the NG-RAN via N2 reference point upon UE context establishment in NG-RAN and upon QoS flow establishment or modification. The notification control indicates whether notifications are requested from the RAN when the GFBR requirement can no longer be satisfied for a QoS flow during its life- time. If the notification control is enabled for a given GBR QoS flow and the NG-RAN determines that the GFBR cannot be satisfied, the AN sends a notification to the SMF, where 5GC upon receiving the notification may initiate an N2 signaling to modify or remove the QoS flow. Each PDU session of a UE is associated with a session aggregate maximum bit rate (session-AMBR). The subscribed session-AMBR is a subscription parameter which is retrieved by the SMF from UDM. The SMF may use the subscribed session-AMBR or modify it based on local policy or use the authorized session-AMBR received from PCF. The session-AMBR limits the aggregate bit rate that can be expected across all non-GBR QoS flows for a specific PDU session. For GBR QoS flows, the 5G QoS profile includes DL/UL GFBR and DL/UL MFBR. The GFBR denotes the bit rate that may be expected to be provided by a GBR QoS flow. The MFBR limits the bit rate that may be expected to be provided by a GBR QoS flow, which means that the excess traffic may be discarded by a rate shaping function. The GFBR and MFBR parameters are signaled to the AN in the QoS profile and sig- naled to the UE as QoS flow level for each

individual QoS flow. For each PDU session, the SMF retrieves the default 5QI and ARP from UDM. The SMF may change the default 5QI/ARP based on local configuration or interaction with the PCF. The default 5QI is derived from the standardized range of values for non-GBR 5QIs. 5G Network Architecture 183 The DL/UL maximum packet loss rate indicates the maximum rate for lost packets of the QoS flow that can be tolerated in the downlink or uplink, which is provided to the QoS flow, if it is compliant to GFBR. The 5G QoS characteristics describe the packet forwarding treatment that a QoS flow receives between the UE and the UPF, which are described in terms of the following perfor- mance metrics [3]: Resource type (GBR, delay critical GBR, or non-GBR) determines whether dedicated net- work resources related to QoS flow-level GFBR value are permanently allocated, for exam- ple, by an admission control function in a base station. The GBR QoS flow is often dynamically authorized, which requires dynamic policy and charging control. A non-GBR QoS flow may be pre-authorized through static policy and charging control. There are two types of GBR resource types, GBR and delay critical GBR, where both resource types are treated in the same manner, except that the definition of PDB and packet error rate (PER) are different. Priority level indicates the resource scheduling priority among QoS flows. The priority levels are used to differentiate between QoS flows of the same UE and they are also used to differentiate between QoS flows from different UEs. Once all QoS requirements are satisfied for the GBR QoS flows, additional resources can be used for any remaining traffic in an implementation-specific manner. In addition, the scheduler may prioritize QoS flows based on other parameters such as resource type, radio condition, etc. in order to optimize application performance and network capacity. The packet delay budget defines an upper bound for the time that a packet may be delayed between the UE and the UPF that terminates N6 i

nterface. The value of the PDB is the same in downlink and uplink for a certain 5QI. The PDB is used to support the configuration of scheduling and link layer functions (e.g., configuration of schedul- ing priority weights and HARQ target operating points). For a delay-sensitive GBR flow, a packet delayed more than PDB is considered as a lost packet, if the data burst is not exceeding the MDBV within the period of PDB and the QoS flow is not exceeding the GFBR. For GBR QoS flows with GBR resource type, the PDB is interpreted as a maximum delay with a confidence interval of 98%. The packet error rate defines an upper bound for the rate of PDUs or IP packets that have been processed by the sender of a link layer protocol, but are not successfully delivered by the corresponding receiver to the upper layers. Therefore, the PER defines an upper bound for non-congestion related packet losses. The purpose of the PER is to find appro- priate link layer protocol configurations. For some 5QI values, the target PER is the same in the downlink and uplink. For QoS flows with delay-sensitive GBR resource type, a packet which is delayed more than PDB is dropped and included in the PER calculation, unless the data burst is exceeding the MDBV within the period of PDB or the QoS flow is exceeding the GFBR. 184 Chapter 1 Averaging window is defined only for GBR QoS flows and denotes the time interval over which the GFBR and MFBR are calculated. The averaging window may be sig- naled with 5QIs to the AN and the UPF and if it is not received a predefined value is used. The maximum data burst volume (MDBV) is associated with each GBR QoS flow with delay-sensitive resource type. The MDBV denotes the largest amount of data that a 5G-AN is required to serve within a period of 5G-AN PDB (i.e., 5G-AN part of the PDB). Each standardized 5QI of delay-sensitive GBR resource type is associated with a default value for the MDBV. The MDBV may also be signaled together with a stan- dardized or pre-configured 5QI to the AN. Reflective QoS enables

a UE to map uplink user-plane traffic to QoS flows without SMF provided QoS rules, which is applied to IP and Ethernet PDU sessions. The support of reflective QoS over the access network is controlled by 5GC. The UE derives the reflective QoS rule from the received downlink traffic. It must be noted that it is possible to apply reflective QoS and non-reflective QoS concurrently within the same PDU session. For user traffic that is subject to reflective QoS, the uplink packets are assigned the same QoS mark- ing as the reflected downlink packets. Reflective QoS is controlled on per-packet basis using the RQI in the encapsulation header on N3 reference point together with the QFI and a reflective QoS timer (RQ timer) value that is either signaled to the UE upon PDU session establishment or set to a default value. To summarize this section, as we discussed 5G session management supports a PDU con- nectivity service that provides PDU exchange between a UE and a data network. In 5GC, the SMF is responsible for handling session management procedures. There is a notable difference in session management between EPC and 5GC. In EPC, the entire ses- sion is maintained by a single MME in a centralized manner, SO that the user-plane path is established via a centralized PGW. This potentially results in congestion of backhaul traffic at the PGW, whereas in 5GC, different PDU sessions can be maintained by conceivably dif- ferent SMFs, and their user-plane paths are established via multiple UPFs. This can distrib- ute the cellular operator's backhaul traffic within the 5GC and reduce the perceived latency by the user. Compared to LTE's QoS framework, which is bearer-based and uses control- plane signaling, the 5G system adopts the QoS flow-based framework, and uses both control-plane and user-plane (i.e., reflective QoS) signaling in order to satisfy various appli- cation/service QoS requirements. The QoS flow-based framework enables flexible mapping of QoS flows to DRB(s) by decoupling the QoS flow and the radio bearer, allowi

ng more flexible QoS characteristics. When reflective QoS is used, the 5G UE can create a QoS rule for the uplink traffic based on the received downlink traffic without generating control- plane signaling overhead. Table 1.5 summarizes and compares the EPS and 5GS QoS and session management features. 5G Network Architecture 185 Table 1.5: Comparison of LTE and NR QoS and session management characteristics [71]. RAN-Level Interworking 5G SA With Core-Network-Level With EPC With 5GC Interworking Network Per device Per service (enabling third-party Per service (enabling third-party slicing (dedicated core) service) service) Session Limited and Flexible and distributed (lower Flexible and distributed (lower management centralized cost, lower latency) cost, lower latency) Per-bearer Per-flow Per-flow Network-initiated UE/network-initiated (dynamic UE/network-initiated (dynamic 1.7 Security Framework We begin this section with a review of security framework in LTE to set the stage for the 5G security framework discussion. Fig. 1.89 shows the scope and overall concept of the LTE security architecture. In LTE, the authentication function performs mutual authentica- tion between the UE and the network. The NAS security performs integrity protection/veri- fication and ciphering (encryption/decryption) of NAS signaling between the UE and the MME. The AS security is responsible for integrity protection/verification and ciphering of RRC signaling between the UE and the eNB and further performs ciphering of user traffic between the UE and the eNB. In 3GPP networks, authentication refers to the process of determining whether a user is an authorized subscriber to the network that it is trying to access. Among various authentica- tion procedures, EPS AKA procedure is used in LTE networks for mutual authentication between users and networks. The EPS AKA procedure consists of two steps: (1) the home subscriber server (HSS) generates EPS authentication vector(s) (RAND, AUTN, XRES, KASME) and delivers them to the MME and (2) the MME

selects one of the authentication vectors and uses it for mutual authentication with the UE and shares the same authentication key (KASME). Mutual authentication is the process in which a network and a user authenti- cate each other. In LTE networks, since the identification of the user's serving network is required when generating authentication vectors, authentication of the network by the user is performed in addition to authentication of the user by the network. Access security man- agement entity (ASME) is an entity that receives top-level key(s) from the HSS to be used in an AN. In EPS, the MME serves as the ASME and KASME is used as the top-level key in the AN. The MME conducts mutual authentication with the UE on behalf of the HSS using KASME. Once mutually authenticated, the UE and MME share the same KASME as an authentication key. USIM/AuC MeNB/UE SeNB/UE CK, IK SCG counter S-KeNB UE/HSS KUPenc KASME UE/MME Key derivation function (KDF) UP-enc-alg, Alg-ID (User-plane encryption Trunc algorithm) KNAS enc KNAS int Truncation UE/eNB KUPenc KUP enc K RRC enc K RRC int KenB* Figure 1.89 Key derivation in LTE and dual connectivity [56,57]. 5G Network Architecture 187 NAS security is designed to securely deliver signaling messages between the UE and the MME over the radio link and to perform integrity protection/verification as well as cipher- ing of NAS signaling messages. Different keys are used for integrity verification and ciphering. While integrity verification is a mandatory function, ciphering is an optional function. The NAS security keys, such as integrity key (KNASint) and ciphering key (KNASenc), are derived by the UEs and the MMEs from KASME (see Fig. 1.89). In Fig. 1.89, next hop (NH) key is used by the UE and eNB in the derivation of Kenb* for provisioning forward security. The NH is derived by UE and MME from KASME and KeNB when the security context is established, or from KASME and the previous NH. The NH chaining count (NCC) is a counter related to NH, that is, the number of key chaining that

has been performed, which allows the UE to be synchronized with the eNB and to deter- mine whether the next KeNB* needs to be based on the current KeNB or a fresh NH value [56,57]. AS security is used to ensure secure delivery of data between a UE and an eNB over the radio interface. It includes both integrity check and ciphering of RRC signaling messages over the control plane, and only ciphering of IP packets over the user plane. Different keys are used for integrity check/ciphering of RRC signaling messages and ciphering of IP pack- ets. Integrity verification is mandatory, but ciphering is optional. AS security keys, such as KRRCint, KRRCenc, and KUPenc, are derived from KeNB by a UE and an eNB. KRRCint and KRRCenc are used for integrity check and ciphering of control-plane information (i.e., RRC signaling messages), and KUPenc is used for ciphering of user-plane data (i.e., IP packets). Integrity verification and ciphering are performed at the PDCP sublayer. Key derivation for dual connectivity SCG bearers is depicted in Fig. 1.89, where SCG counter is a counter used as freshness input into S-Kenb derivations. For SCG bearers in dual connectivity, the user-plane keys are updated upon SCG change by conveying the value of the SCG counter to be used in key derivation to the UE via RRC signaling. When KeNB is refreshed, SCG counter is reset and S-KeNB is derived from the Kenb. The SCG bearers in dual connectivity scenarios share a common pool of radio bearer identities (DRB IDs) with the MCG bearers. When no new DRB ID can be allocated for an SCG bearer without guaranteeing COUNT reuse avoidance, the MeNB derives a new S-Kenb. The SeNB informs MeNB when the uplink or downlink PDCP COUNTs are about to wrap around. In that case, the MeNB updates the S-KeNB. To update the S-KeNB, the MeNB increases the SCG counter and uses it to derive a new S-Kenb from the currently active KeNB in the MeNB. The MeNB sends the freshly derived S-KeNB to the SeNB. The newly derived S-KeNB is then used by the SeNB in computing a new

encryption key KUPenc which is used for all DRBs in the SeNB for the target UE. Furthermore, when the SCG counter approaches its maximum value, the MeNB refreshes the currently active KeNB, before any further S-KeNB is derived [56,57]. 188 Chapter 1 Integral AN-CN CP CP confidentiality CP integrity AN-CN-UP UP confidentiality UP integrity NR Uu CP: Control plane UP: User plane Figure 1.90 5GS security architecture and termination points [54]. Fig. 1.90 shows 5G security architecture including the new security entities: SEAF, AUSF, authentication credential repository and processing function (ARPF), security context man- agement function (SCMF), and security policy control function (SPCF). 5GC has introduced a new security anchor called SEAF, which may be co-located with the AMF. The SEAF will create, for the primary authentication, a unified anchor key KSEAF (common for all access links) that can be used by the UE and the serving network to protect the subsequent communications. It is possible to generate two anchor keys for certain sce- narios where a UE is connected to 3GPP access (visited network) and to a non-3GPP access (home network). For normal roaming scenarios, the SEAF is located in the visited network. The AUSF terminates requests from the SEAF and further interacts with the ARPF. Depending on how the authentication functionality is split, the AUSF and the ARPF may be co-located, but an interface similar to SWx interface is defined for EAP-AKA and EAP- AKA'. The ARPF is co-located with the UDM and stores long-term security credentials such as the key K in EPS AKA or EAP-AKA for authentication. It can run cryptographic algorithms using long-term security credentials as input and can create the authentication vectors. Another new functional entity is the SCMF, which may be co-located with the SEAF in the AMF and retrieve a key from the SEAF, which is used to derive further access network specific keys. The SPCF provides the security policy to the network entities (e.g., SMF, AMF) and/or to the UE dependi

ng on the application-level input from the AF and may be stand-alone or co-located with the PCF. The security policy may include 5G Network Architecture Network side UE side CK, IK 5G AKA EAP-AKA' CK', IK' KAUSF KSEAP KNASint KNASenc KN3IWF KgNB, NH KRRCint KRRCenc KUPint KUPenc N3IWF Mobile equipment (ME) Figure 1.91 Key hierarchy in 5GS [9]. information about AUSF selection, confidentiality protection algorithm, integrity protection algorithm, key length and key life cycle. Fig. 1.91 shows the key hierarchy in 5GS [54]. The new security anchor key KSEAF is used to further derive the access network key KAN and the NAS keys KNAS. There is only one NAS security termination entity, which is the AMF. The user-plane data on the radio bearer can be secured on a per session basis with the key KUP. A session can belong to the same or different network slices. All key sets for NAS, RRC, and user-plane consist of an integrity key and a confidentiality key for encryp- tion. Recall that the termination point of the user-plane security is at eNB in LTE networks. However, the gateway location may vary in order to provide different type of services, and the gNB may be located at the edge, that is, an exposed environment. Thus, the termination point of user-plane security should be reconsidered with the principle that security termina- tion is in the entity where the traffic terminates. The user-plane security terminates at PDCP sublayer of gNB. This is aligned with LTE security framework that radio interface security 190 Chapter 1 is provided by the PDCP layer for control and user planes. This mechanism allows provi- sioning the security termination point in a CU of the gNB that typically resides in a secure location. The key hierarchy shown in Fig. 1.90 includes the following keys: KSEAF, KAMF, KNASint, KNASenc, KN3IWF, KgNB, KRRCint, KRRCenc, KUPint and KUPenc, which can be divided into the following groups [9]: Keys for NAS signaling include KNASint, which is a key that is used for the protection of NAS signaling with a par

ticular integrity algorithm, as well as KNASenc which is a key that is used for the protection of NAS signaling with a particular encryption algorithm. Keys for user-plane traffic include KUPenc, which is a key that is used for the protection of user-plane traffic with a particular encryption algorithm, as well as KUPint which is a key that is used for the protection of user-plane traffic between mobile equipment and gNB with a particular integrity algorithm. Keys for RRC signaling include KRRCint which is a key that is used for the protection of RRC signaling with a particular integrity algorithm, as well as KRRCenc which is a key that is used for the protection of RRC signaling with a particular encryption algorithm. Intermediate keys include next hop (NH), which is a key that is derived by mobile equipment and AMF to provide forward security as well as KgNB* which is a key that is derived by mobile equipment and gNB when performing a horizontal or vertical key derivation. When a UE obtains services in RRC idle mode, it does not validate the eNB, which may result in camping on a wrong base station, ultimately leading to denial of service attack. In current LTE systems, the RAN security has been focused on RRC_CONNECTED state, which has been improved in 5G security framework. The purpose of the primary authentication and key agreement procedures is to enable mutual authentication between the UE and the network and to provide key derivation mate- rial that can be used between the UE and network in subsequent security procedures. The keying material generated by the primary authentication and key agreement procedure results in an anchor key called the KSEAF provided by the AUSF to SEAF. Keys for more than one security context can be derived from the KSEAF with no need for a new authentication. In 5G systems, the storage of credentials and identities for both human and machine type devices is required in the UE. The credentials and identities can be stolen through hard- ware/software attacks. Such security threats

can impact the subscriber and/or the network. 3GPP UE security framework provides the following features for storage of UE credentials: integrity protection of the subscription credential(s); confidentiality protection of the 5G Network Architecture long-term key(s) of the subscription credential(s); and execution of the authentication algo- rithm(s) that make use of the subscription credentials. These features must be implemented in the UE, with using a tamper resistant secure hardware component. The subscriber identity module (SIM) functions for 5G are inherited from previous standards. Similar to LTE sys- items, the 5G USIM will be able to generate symmetric keys. It may also be able to generate new asymmetric key pairs and even new trusted public keys. Network slicing requires basic security from the UE side when accessing a slice; however, this is not trivial and there are new security challenges. The slice isolation must be ensured for network slices, without which attackers who access to one slice may attempt an attack to other slices. Proper isolation will enable integrity and confidentiality protection. Moreover, it should be ensured that resources of the network infrastructure or an NSI are not impacted from another slice instance, to minimize attacks and to provide availability. 5G UE can simultaneously access to different network slices for multiple services. Such access can be via various type of RANs including both 3GPP and non-3GPP access. When the network slice selection data is tampered, unauthorized UEs may use such information to establish connection with the network slice and consume network resources. On the other hand, the advantage of network slicing is that operators are able to provide customized security for each slice. Different access authentication and authorization can be provided within different network slice tenants that can be extended to host applications. In order to support network-controlled privacy of slice information for the slices that the UE can access, the UE must be ma

de aware or configured with privacy considerations that apply to NSSAI. The UE must not include NSSAI in NAS signaling or unprotected RRC signaling unless it has a NAS security context setup [3]. References 3GPP Specifications [1] 3GPP TS 23.214, Architecture enhancements for control and user plane separation of EPC nodes, Stage 2 (Release 15), September 2017. [2] 3GPP TS 23.402, Architecture enhancements for non-3GPP accesses (Release 15), March 2018. [3] 3GPP TS 23.501, System architecture for the 5G system (Release 15), December 2018. [4] 3GPP TS 23.502, Procedures for the 5G system (Release 15), April 2019. [5] 3GPP TR 23.714, Study on control and user plane separation of EPC nodes (Release 14), June 2016. [6] 3GPP TR 23.799, Study on architecture for next generation system (Release 14), December 2016. [7] 3GPP TR 28.801, Study on management and orchestration of network slicing for next generation network (Release 15), January 2018. [8] 3GPP TS 29.244, Interface between the control plane and the user plane of EPC nodes, Stage 3 (Release 15), March 2019. [9] 3GPP TS 33.501, Security architecture and procedures for 5G system (Release 15), December 2018. [10] 3GPP TR 33.899, Study on the security aspects of the next generation system (Release 14), August 2017. 3GPP specifications can be accessed at the following URL: http://www.3gpp.org/ftp/Specs/archive/ Chapter 1 [11] 3GPP TS 36.104, Evolved universal terrestrial radio access (E-UTRA), Base Station (BS) Radio Transmission and Reception (Release 15), June 2018. [12] 3GPP TR 36.932, Scenarios and requirements for small cell enhancements for E-UTRA and E-UTRAN (Release 14), March 2017. 3GPP TS 37.340, Multi-connectivity, Stage 2 (Release 15), January 2018. [14] 3GPP TS 38.104, NR; base station (BS) radio transmission and reception (Release 15), January 2018. 3GPP TS 38.401, NG-RAN, architecture description (Release 15), December 2018. [16] 3GPP TS 38.300 NR, Overall description, Stage-2 (Release 15), December 2018. [17] 3GPP TS 38.410 NG-RAN, NG general aspects a

nd principles (Release 15), December 2018. 3GPP TS 38.411 NG-RAN, NG layer 1 (Release 15), December 2018. [19] 3GPP TS 38.412 NG-RAN, NG signaling transport (Release 15), December 2018. [20] 3GPP TS 38.413 NG-RAN, NG application protocol (NGAP) (Release 15), December 2018. [21] 3GPP TS 38.414 NG-RAN, NG data transport (Release 15), December 2018. [22] 3GPP TS 38.420 NG-RAN, Xn general aspects and principles (Release 15), December 2018. 3GPP TS 38.421 NG-RAN, Xn layer 1 (Release 15), December 2018. [24] 3GPP TS 38.422 NG-RAN, Xn signaling transport (Release 15), December 2018. [25] 3GPP TS 38.423 NG-RAN, Xn application protocol (XnAP) (Release 15), December 2018. 3GPP TS 38.424 NG-RAN, Xn data transport (Release 15), December 2018. [27] 3GPP TS 38.425 NG-RAN, Xn interface user plane protocol (Release 15), December 2018. [28] 3GPP TS 38.470 NG-RAN, F1 general aspects and principles (Release 15), December 2018. 3GPP TS 38.471 NG-RAN, F1 layer 1 (Release 15), December 2018. 3GPP TS 38.472 NG-RAN, F1 signaling transport (Release 15), December 2018. [31] 3GPP TS 38.473 NG-RAN, F1 application protocol (XnAP) (Release 15), December 2018. [32] 3GPP TS 38.474 NG-RAN, F1 data transport (Release 15), December 2018. 3GPP TS 38.475 NG-RAN, F1 interface user plane protocol (Release 15), December 2018. [34] 3GPP TR 38.801, Study on new radio access technology: radio access architecture and interfaces (Release 14), March 2017. 3GPP TR 38.804, Study on new radio access technology radio interface protocol aspects (Release 14), March 2017. [36] 3GPP TR 38.806, Study of separation of NR control plane (CP) and user plane (UP) for split option 2 (Release 15), December 2017. ETSI Specifications8 [37] ETSI GS NFV-SWA 001, Network functions virtualization (NFV), virtual network functions architecture, December 2014. [38] ETSI GS NFV-IFA 001, Network functions virtualization (NFV), acceleration technologies, report on acceleration technologies & use cases, December 2015. [39] ETSI GS NFV-IFA 002, Network functions virtualization (NFV) Rele

ase 2, Acceleration Technologies, VNF Interfaces Specification, August 2017. [40] ETSI GS NFV-INF 001, Network functions virtualization (NFV), Infrastructure Overview, January 2015. [41] ETSI GS NFV 002, Network functions virtualization (NFV), Architectural Framework, December 2014. [42] ETSI, Network functions virtualization, White Paper on NFV Priorities for 5G, February 2017. [43] ETSI, Network functions virtualization, Introductory White Paper, October 2012. [44] ETSI, Network functions virtualization, White Paper, October 2014. [45] ETSI GS MEC-IEG 004, Mobile-edge computing (MEC), Service Scenarios, November 2015. [46] ETSI GS MEC 003, Mobile edge computing (MEC), Framework and Reference Architecture, March 2016. ETSI specifications can be accessed at the following URL: http://www.etsi.org/deliver/ 5G Network Architecture Articles, Books, White Papers, and Application Notes [47] China Mobile Research Institute, Toward 5G C-RAN: Requirements, Architecture and Challenges, November 2016. [48] SDN, NFV, and MEC on SDxCentral. available at: <https://www.sdxcentral.com/>. [49] Open Networking Foundation, OpenFlow Switch Specification, version 1.5.1, March 2015. [50] Y. Chao Hu, et al., Mobile edge computing: a key technology towards 5G, ETSI White Paper No. 11, September 2015. [51] 5G network architecture, a high-level perspective, Huawei Technologies Co., Ltd., 2016. [52] K. Miyamoto, et al., Analysis of mobile fronthaul bandwidth and wireless transmission performance in split-PHY processing architecture, Optics Express 24 (2) (2016) 1261-1268. [53] D. Sabella, et al., Mobile-edge computing architecture, the role of MEC in the Internet of Things, IEEE Consumer Electron Mag. 5 (4) (2016). [54] X. Zhang, et al., Overview of 5G security in 3GPP, in: IEEE Conference on Standards for Communications and Networking (CSCN), September 2017. J. Kim, et al., 3GPP SA2 architecture and functions for 5G mobile communication system, The Korean Institute of Communications Information Sciences, 2017. [56] LTE security I: LTE sec

urity concept and LTE authentication, NMC Consulting Group, July 2013. [57] LTE security II: NAS and AS security, NMC Consulting Group, July 2013. [58] LTE network architecture, NMC Consulting Group, July 2013. [59] LTE QoS-SDF and EPS bearer QoS, NMC Consulting Group, September 2013. [60] Emergence of C-RAN: separation of baseband and radio, and baseband centralization, NMC Consulting Group, March 2014. [61] The benefits of cloud-RAN architecture in mobile network expansion, Fujitsu Network Communications Inc., 2014. [62] NGMN Alliance, 5G end-to-end architecture framework, October 2017. [63] NGMN Alliance, Service-based architecture in 5G, January 2018. [64] NGMN Alliance, Update to NGMN description of network slicing concept, October 2016. [65] NGMN Alliance, NGMN paper on edge computing, October 2016. [66] NGMN Alliance, Backhaul provisioning for LTE-advanced & small cells, October 2015. [67] NGMN Alliance, Project RAN evolution: further study on critical C-RAN technologies, March 2015. [68] J. Liu, et al., Ultra-dense networks (UDNs) for 5G, IEEE 5G Tech Focus 1 (1) (2017) 6. [69] J. Wannstrom, et al., HetNet/small cells. Available from: <http://www.3gpp.org/hetnet> [70] N.T. Le, et al., Survey of promising technologies for 5G networks, Mobile Information Systems 2016 (2016) 6 pp. [71] V.G. Nguyen, K.J. Grinnemo, SDN/NFV-based mobile packet core network architectures: a survey, IEEE Commun. Surv. Tutor 19 (3) (2017) 1567-1602. [72] J.E. Mitchell, Integrated wireless backhaul over optical access networks, J Lightw. Technol. 32 (20) (2014) 3373-3382. [73] R. Trivisonno, et al., Network slicing for 5G systems, in: IEEE Conference on Standards for Communications and Networking (CSCN), 2017. [74] 5G Service-Guaranteed Network Slicing, White Paper, Huawei Technologies Co., Ltd., February 2017. CPRI Specification V7.0, Common Public Radio Interface (CPRI): Interface Specification, October 2015. [76] eCPRI Specification V1.2, Common Public Radio Interface: eCPRI Interface Specification, June 2018. [77] IEEE Std 1914

.3-2018, Standard for radio over Ethernet encapsulations and mappings, September 2018. [78] IEEE P1914.1/D4.1, Draft standard for packet-based fronthaul transport networks, April 2019. [79] D. Anzaldo, Backhaul alternatives for HetNet small cells, Part 1 and 2, Microwaves & RF, September 2015. [80] Wireless Backhaul Spectrum Policy Recommendations and Analysis Report, GSMA, November 2014. [81] R. Ravindran, et al., Realizing ICN in 3GPP's 5G NextGen core architecture, Cornell University Library, November 2017. 194 Chapter 1 [82] 4G-5G Interworking, RAN-level and CN-level Interworking, Samsung, June 2017. [83] Nomor Research, 5G RAN Architecture Interfaces and eCPRI, September 2017. [84] R. Vaez-Ghaemi, The evolution of fronthaul networks, Viavi Solutions, June 2017. [85] Sujuan Feng and Eiko Seidel, Self-Organizing Networks (SON) in 3GPP Long Term Evolution, Nomor Research GmbH, May 2008. CHAPTER 2 New Radio Access Layer 2/3 Aspects and System Operation 2.1 Overview of Layer 2 and Layer 3 Functions The NR radio interface protocols (alternatively referred to as layer-1, layer-2, and layer-3 protocols) operate between the NG-RAN and the UE and consist of user-plane protocols, for transfer of user data (IP packets) between the network and the UE, and control-plane protocols, for transporting control signaling information between the NG-RAN and the UE. The non-access stratum (NAS) protocols terminate in the UE and the AMF entity of the 5G core network and are used for core network related functions and signaling including regis- tration, authentication, location update and session management. In other words, the proto- cols over Uu and NG interfaces are categorized into user plane and control plane protocols. User plane protocols implement the actual PDU Session service which carries user data through the access stratum. Control plane protocols control PDU Sessions and the connec- tion between UE and the network from various aspects which include requesting the service, controlling different transmission resources, h

andover etc. The mechanism for transparent transfer of NAS messages is also included. The layer 2 of the new radio protocol stack is split into four sublayers: medium access control (MAC), radio link control (RLC), packet data convergence protocol (PDCP), and service data adaptation protocol (SDAP), where each sublayer hosts a number of functions and performs certain functionalities that are con- figurable via radio resource control (RRC) at layer 3. Fig. 2.1 shows the OSI protocol layers and how they map to 3GPP radio protocol architecture. As shown in the figure (dark- shaded boxes), the data link layer of OSI network model maps to these four sublayers that constitute layer 2 of the 3GPP new radio protocols. The layer 2 protocols of NR have some similarities with the corresponding LTE protocols; however, the NR has added more config- uration flexibility and more functionalities to support the new features such as beam man- agement that have no counterpart in LTE. A significant difference in NR RRC compared to LTE RRC is the introduction of a 3-state UE behavior with the addition of the RRC_INACTIVE state. The RRC_INACTIVE provides a state with battery efficiency simi- lar to RRC_IDLE while storing the UE context within the NG-RAN SO that the transitions to/from RRC_CONNECTED are faster and incur less signaling overhead. The other signifi- cant improvements relative to LTE RRC are the support of on-demand system information 5G NR. DOI: : https://doi.org/10.1016/B978-0-08-102267-2.00002-6 © 2019 Elsevier Inc. All rights reserved. Chapter 2 NR protocol structure OSI seven-layer network Non-access stratum (NAS) Internet protocol (IP) model Application layer Radio resource control (RRC) Presentation layer QoS flows Session layer Service data adaptation Control/ protocol (SDAP) configuration Transport layer Radio bearers RRC PDUs Packet data convergence Network layer protocol (PDCP) Control/configuration RLC channels Data-Link layer Radio link control (RLC) Physical layer Control/configuration Logical channels Medium

access control (MAC) Control/configuration Transport channels Physical layer (PHY) Control/configuration/ Physical layer measurements Physical channels Figure 2.1 The layer 2/3 protocols in NR protocol stack [8]. transmission that enables the UE to request when specific system information is required instead of NG-RAN consuming radio resources to periodically broadcast system informa- tion, and the extension of the measurement reporting framework to support beam measure- ments for handover in a beamformed operation [8]. In NR layer 2 protocol structure, each sublayer provides certain services to the immediately adjacent layers by processing the incoming service data units (SDUs) and generating the proper protocol data units (PDUs). The functional processing of the protocol layers is further classified into user-plane and control-plane protocols (Fig. 2.2), where the reliability requirement of the control-plane information is much higher than that of the user-plane data. New Radio Access Layer 2/3 Aspects and System Operation User-plane Control-plane IP packets (user data) NAS messages QoS flows System QoS management information Paging RRC messages Radio bearers Header Encryption and compression, integrity encryption protection RLC channels Logical channels Transport channels Channel coding Channel coding Channel coding Channel coding Physical channels Figure 2.2 NR layer 2 user-plane and control-plane functional mapping and processing. The service data adaptation protocol is a new sublayer in layer 2 which immediately inter- faces with the network layer and provides a mapping between the QoS flows and data radio bearers (DRBs). It also marks the QoS flow identifiers (QFIs) in the downlink and uplink packets. A single-protocolo entity of SDAP is configured for each individual PDU Session. The services and functions of the PDCP sublayer on the user-plane include sequence num- bering; header compression and decompression [only supports Robust Header Compression (ROHC) protocol]; transfer of user data; reordering an

d duplicate detection; in-sequence delivery; PDCP PDU routing (in case of split bearers); retransmission of PDCP SDUs; ciphering, deciphering, and integrity protection; PDCP SDU discard; PDCP re- establishment and data recovery for RLC acknowledged mode (AM); PDCP status reporting for RLC-AM; duplication of PDCP PDUs and duplicate discard indication to lower layers. The main services and functions of the PDCP sublayer on the control-plane consist of sequence numbering; ciphering, deciphering, and integrity protection; transfer of control- plane data; reordering and duplicate detection; in-sequence delivery; duplication of PDCP PDUs and duplicate discard indication to lower layers [8]. The PDCP protocol performs (optional) IP-header compression, followed by ciphering, for each radio bearer. A PDCP header is added, carrying information required for deciphering in the other end, as well as a sequence number used for retransmission and in-sequence delivery. The RLC sublayer supports three transmission modes, which include transparent mode (TM), unacknowledged mode (UM), and the acknowledged mode. The main difference between these 198 Chapter 2 modes is the presence or absence of ARQ function to provide higher reliability required for RRC and NAS messages. The RLC configuration is per logical channel with no dependency on numerologies and/or transmission duration, and ARQ can operate on any of the numerologies and/or transmission durations that the logical channel is configured to support. The radio bearers in NR are classified into two groups of DRBs for user-plane data and signaling radio bearers (SRBs) for control-plane data. The RLC-TM mode is used for transport of SRBO, paging, and broadcast system information (SI), whereas for other SRBs, RLC-AM mode is used. For trans- port of DRBs, either RLC-UM or RLC-AM mode is used. Some of the services and functions of the RLC sublayer depend on the transmission mode. Those services include transfer of upper layer PDUs; sequence numbering independent of the one in PDCP (RL

C-UM and RLC-AM); error correction through ARQ function (AM mode); segmentation (in RLC-AM and RLC-UM modes) and re-segmentation (in RLC-AM mode) of RLC SDUs; reassembly of SDU (in RLC- AM and RLC-UM modes); duplicate detection (in RLC-AM mode); RLC SDU discarding (in RLC-AM and RLC-UM modes); RLC re-establishment; and protocol error detection (in RLC- AM mode). The ARQ function within the RLC sublayer performs ARQ retransmission of RLC SDUs or SDU segments based on the RLC status reports. It further sends requests for RLC sta- tus reports when needed, and triggers an RLC status report after detecting a missing RLC SDU or SDU segment [8]. The RLC protocol performs segmentation of the PDCP PDUs, if neces- sary, and adds an RLC header containing a sequence number used for handling retransmissions. Unlike LTE, the NR RLC is not providing in-sequence delivery of data to higher layers due to the additional delay incurred by the reordering mechanism. In-sequence delivery can be pro- vided by the PDCP layer, if necessary. The services and functions of the MAC sublayer include mapping between logical and transport channels, as well as multiplexing/demultiplexing of MAC SDUs belonging to one or different logical channels on the transport blocks, which are mapped to the transport channels depending on the required physical layer processing. The MAC sublayer also per- forms scheduling of measurement reporting; error correction through HARQ [one HARQ entity per cell when carrier aggregation is utilized]; priority handling between user equip- ments through dynamic scheduling; and priority handling between logical channels of one UE through logical channel prioritization. A single MAC instantiation can support multiple numerologies, transmission timings, and cells. The mapping restrictions in logical channel prioritization can control which numerology(ies), cell(s). and transmission timing(s) a logi- cal channel can use. The main difference between MAC and RRC control lies in the signal- ing reliability. The signaling correspo

nding to state transitions and radio bearer configurations should be performed by the RRC sublayer due to higher signaling reliability requirement. The RLC PDUs are delivered to the MAC sublayer, which multiplexes a num- ber of RLC PDUs and adds a MAC header to form a transport block. It must be noted that in NR, the MAC headers are distributed across the MAC PDU, such that the MAC header corresponding to a particular RLC PDU is located immediately prior to it (see Fig. 2.3). New Radio Access Layer 2/3 Aspects and System Operation IP packet IP packet IP packet Radio bearer x Radio bearer y SDAP SDU SDAP SDU SDAP SDU PDCP SDU PDCP SDU PDCP SDU RLC SDU RLC SDU SDU segment SDU segment MAC SDU MAC SDU MAC SDU MAC SDU MAC PDU MAC PDU PHY SDU (transport block) PHY SDU (transport block) H: Header or subheader Figure 2.3 Layer 2 packet processing [8]. This is different compared to LTE, in which all header information are located at the begin- ning of the MAC PDU. The NR MAC PDUs can be assembled as soon as the RLC PDUs become available; thus there is no need to assemble the full MAC PDU before the header fields can be computed. This reduces the processing time and the overall latency 8,18]. The physical layer provides information transfer services to the MAC and higher layers. The physical layer transport services are described by how and with what characteristics data is transferred over the radio interface. This should be clearly distinguished from the classification of what is transported which relates to the concept of logical channels at the MAC sublayer. It was mentioned earlier that each sublayer in layer 2 radio protocols receives SDUs from the previous layer, processes the information according to the configured functions and parameters of the sublayer, and generates PDUs that are delivered to the next layer. In this process, a unique header or subheader is attached to the SDU by each sublayer. An example is shown in Fig. 2.3, where a transport block is generated by MAC sublayer by means of concatenating two RLC

PDUs from radio bearer X and one RLC PDU from radio bearer y. The RLC PDUs from radio bearer X each corresponds to one IP packet (n and n + 1) while the RLC PDU from radio bearer y is a segment of an IP packet (m) [8]. The services and functions of the RRC sublayer include broadcast of SI related to access stra- tum (AS) and NAS, as well as paging initiated by 5GC or NG-RAN (paging initiated by NG- RAN is a new NR feature). The RRC sublayer services further includes establishment, mainte- nance, and release of an RRC connection between the UE and NG-RAN that consist of addi- tion, modification, and release of carrier aggregation; addition, modification, and release of 200 Chapter 2 dual connectivity (DC) between LTE and NR. The security functions including key manage- ment, establishment, configuration, maintenance, and release of signaling and DRBs as well as mobility management, which comprises handover and context transfer; UE cell selection and reselection and control of cell selection and reselection; inter-RAT mobility; QoS management functions; UE measurement reporting and control of the reporting; detection of and recovery from radio link failure (RLF); and NAS message transfer are among other services provided by the RRC sublayer [8,15]. An NR UE at any time is in one of the three RRC states that are defined as follows [8]: RRC_IDLE which is characterized by PLMN selection; broadcast of system informa- tion; cell reselection; paging for mobile terminated data is initiated by 5GC; and discon- tinuous reception (DRX) for core-network paging configured by NAS. RRC_INACTIVE which is characterized by PLMN selection; broadcast of system infor- mation; cell reselection; paging initiated by NG-RAN (RAN paging); RAN-based notifi- cation area (RNA) is managed by NG-RAN; DRX for RAN paging configured by NG- RAN; 5GC and NG-RAN connection (both control and user-planes) establishment for the UEs; storage of UE AS context in NG-RAN and the UE; and NG-RAN knowledge of UE location at RNA-level. RRC_CONNECTED which is c

haracterized by 5GC and NG-RAN connection (both control and user-planes) establishment for the UEs; storage of the UE AS context in NG-RAN and the UE; NG-RAN knowledge of UE location at cell level; transfer of uni- cast data between the UE and gNB; and network-controlled mobility including measurements. The SI consists of a master information block (MIB) and a number of system information blocks (SIBs), which are divided into minimum SI and other SI. The minimum SI comprises basic information required for initial access and information for acquiring any other SI. The other SI encompasses all SIBs that are not broadcast in the minimum SI. Those SIBs can either be periodically broadcast on downlink shared channel (DL-SCH), broadcast on- demand on DL-SCH upon request from UEs in RRC_IDLE or RRC_INACTIVE states or sent in a dedicated manner on DL-SCH to UEs in RRC_CONNECTED state. A UE is not required to acquire the content of the minimum SI of a cell/frequency that is considered for camping from another cell/frequency. This does not preclude the case that the UE applies stored SI from previously visited cell(s). A cell is barred, if a UE cannot determine/receive the full content of the minimum SI of that cell. In case of bandwidth adaptation the UE only acquires SI on the active BWP. In the next sections, we will discuss layers 2 and 3 functions and procedures in more detail. We will further discuss UE states, state transitions, and important procedures such as idle, New Radio Access Layer 2/3 Aspects and System Operation 201 inactive, and connected mode procedures, random-access procedure, mobility and power management, UE capability, and carrier aggregation. 2.2 Layer 2 Functions and Services 2.2.1 Medium Access Control Sublayer The MAC sublayer performs logical channel multiplexing and controls HARQ retransmis- sions. It also handles scheduling functions and is responsible for multiplexing/demultiplex- ing data packets across multiple component carriers when carrier aggregation is configured. The services of MAC

sublayer to the RLC sublayer are in the form of logical channels. A logical channel is defined by the type of information it carries and it is generally classified either as a control channel, for transmission of control and configuration information, or a traffic channel, for transmission of user data. The MAC sublayer provides services to the physical layer in the form of transport channels (see Fig. 2.4). A transport channel is defined by how and with what characteristics the information is transmitted over the radio interface. The information traversing a transport channel is organized in the form of trans- port blocks. In each physical layer transmission time interval (TTI), one transport block with dynamic size is transmitted over the radio interface to a device. In the case of spatial QoS flows QoS flows QoS flow QoS flow QoS flow handling handling handling Radio bearers Radio bearers Security Security Security Security Security Security RLC channels RLC channels Segmentation Segmentation Segmentation Segmentation Segmentation Segmentation Logical channels Logical channels Scheduling/Priority handling Scheduling Multiplexing UE1 Multiplexing UE, Multiplexing Transport channels Transport channels Downlink Uplink Figure 2.4 NR layer 2 functions in the downlink and uplink [8]. 202 Chapter 2 Upper layers Upper layers MAC-control MAC-control Logical channel prioritization (uplink) Logical channel prioritization (uplink) (De) Multiplexing Control (De) Multiplexing Control Random Random access control access control PCH of BCH of DL-SCH of UL-SCH of RACH of BCH of DL-SCH of UL-SCH of RACH of Lower layer of MCG Lower layer of SCG Figure 2.5 Example MAC structure with two MAC entities [9]. multiplexing when more than four layers are configured, there are two transport blocks per TTI. There is a transport format associated with each transport block, specifying how the trans- port block is transmitted over the radio interface. The transport format includes information about the transport block size, the modulation an

d coding scheme, and the antenna mapping. By varying the transport format, the MAC sublayer can realize different data rates, which is known as transport format selection. The MAC entity in the UE manages the broadcast channel (BCH), DL-SCH, paging channel (PCH), uplink shared channel (UL-SCH)], and random-access channel (RACH). When the UE is configured with a secondary cell group (SCG), two MAC entities (instantiations) are configured where one of them is associated with master cell group (MCG) and another one is related to the SCG. Different MAC entities in the UE may operate independently, for example, the timers and parameters used in each MAC entity are independently configured. The serving cells, cell radio network temporary identifier (C-RNTI), radio bearers, logical channels, upper and lower layer entities, logical channel groups (LCGs), and HARQ entities are separately mapped to each MAC entity. If the MAC entity is configured with one or more SCells, there are multiple instances of DL-SCH, UL-SCH, and RACH per MAC entity; how- ever, there is one instance of DL-SCH, UL-SCH, and RACH on the special cell (SpCell), one DL-SCH, zero or one UL-SCH, and zero or one RACH for each SCell. If the MAC entity is not configured with any SCell, there is one instance of DL-SCH, UL-SCH, and RACH per MAC entity. Fig. 2.5 illustrates an example structure of the MAC entities when MCG and SCG are configured [9]. In the context of dual connectivity, the SpCell refers to the primary cell of the MCG or the primary SCell of the SCG depending on whether the MAC entity is associated with the MCG or the SCG. Otherwise, the SpCell refers to the PCell. A SpCell supports PUCCH transmission and contention-based random-access pro- cedure and it is always activated [9]. New Radio Access Layer 2/3 Aspects and System Operation 203 The MAC sublayer provides data transfer and radio resource allocation services to the upper layers and at the same time receives certain services from the physical layer which include data transfer, signaling o

f HARQ feedback, and scheduling request (SR), as well as con- ducting certain link-level measurements. The MAC sublayer provides a mapping between logical channels and transport channels; multiplexing of MAC SDUs from one or different logical channels to transport blocks to be delivered to the physical layer on transport chan- nels; demultiplexing of MAC SDUs to one or different logical channels from transport blocks delivered from the physical layer on transport channels; scheduling measurement reporting; forward error correction through HARQ; and logical channel prioritization. The MAC sublayer further provides data transfer services via logical channels. To accommodate different types of data transfer services, various logical channels are defined, each support- ing transfer of a particular type of information. Each logical channel is defined by the type of information which is transferred [9]. The MAC entity maps logical channels to transport channels in the uplink and downlink. This mapping depends on the multiplexing that is con- figured by RRC sublayer. The priority handling among multiple logical channels, where each logical channel has its own RLC entity, is supported by multiplexing the logical channels to one transport chan- nel. The MAC entity at the receiving side handles the corresponding demultiplexing and forwards the RLC PDUs to their respective RLC entity. To enable the demultiplexing function at the receiver, a MAC header is used. In NR, instead of putting the entire MAC header information at the beginning of a MAC PDU as LTE does, which implies that the assembly of a MAC PDU cannot start until the scheduling decision is available, the sub- header corresponding to a certain MAC SDU is placed immediately in front of the SDU, as shown in Fig. 2.6. This allows the PDUs to be processed before a scheduling decision is received. If necessary, padding can be used to align the transport block size with those supported in NR. A MAC subheader contains the identity of the logical channel (LCID) from which

Uplink MAC PDU R/F/LCID/L R/LCID Fixed-sized R/F/LCID/L Variable-sized MAC SDU subheader subheader MAC CE subheader MAC CE MAC sub-PDU MAC sub-PDU MAC sub-PDU including MAC sub-PDU including MAC sub-PDU including including MAC SDU including MAC SDU MAC CE 1 MAC CE 2 padding (optional) Figure 2.6 Example structures of downlink/uplink MAC PDUs [9]. A MAC PDU is a bit string that is octet-aligned and consists of one or more MAC sub- PDUs. Each MAC sub-PDU may consist of a subheader only (including padding); a subheader and a MAC SDU; a subheader and a MAC CE; or a subheader and padding. The MAC SDUs have variable sizes, where each MAC subheader corresponds to either a MAC SDU, a MAC CE, or padding. A MAC subheader typically consists of four header fields R/F/LCID/L (see Fig. 2.7). However, a MAC subheader for fixed-sized MAC CE, padding, and a MAC SDU containing uplink common control channel (CCCH), consists of two header fields R/LCID. MAC control elements are placed together. The downlink MAC subPDU(s) with MAC CE(s) is placed before any MAC subPDU with MAC SDU and MAC subPDU with padding. The uplink MAC subPDU(s) with MAC CE(s) is placed after all MAC subPDU(s) with MAC SDU and before the MAC subPDU with padding in the MAC PDU. Note that the size of padding can be zero. At most one MAC PDU can be transmitted per transport block per MAC entity. The aforementioned subheader fields are defined as follows [9]: LCID: The LCID field identifies the logical channel instance of the corresponding MAC SDU or the type of the corresponding MAC CE or padding. There is one LCID field per MAC subheader whose size is 6 bits. L: The length field indicates the length of the corresponding MAC SDU or variable- sized MAC CE in bytes. There is one L field per MAC subheader except for subheaders corresponding to fixed-sized MAC CEs, padding, and MAC SDUs containing uplink CCCH. The size of the L field is indicated by the F field. New Radio Access Layer 2/3 Aspects and System Operation 205 Octet 1 Octet 1 Octet 2 Octet 1 Octet 2 Octet 3

Figure 2.7 Structure of various MAC subheaders [9]. F: The format field indicates the size of the length field. There is one F field per MAC subheader except for subheaders corresponding to fixed-sized MAC CEs, padding, and MAC SDUs containing uplink CCCH. The size of the F field is 1 bit, where the 8-bit and 16-bit Length fields are indicated by F=0 and F=1 1, respectively. R: Reserved bit, which is set to zero. The MAC subheaders consist of the following fields: E: The extension field is a flag indicating whether the MAC sub-PDU, including the MAC subheader is the last MAC sub-PDU in the MAC PDU. It is set to one to indicate at least another MAC sub-PDU follows; otherwise, it is set to zero. T: The type field is a flag indicating whether the MAC subheader contains a random- access preamble identifier (RAPID) or a backoff indicator (BI). It is set to zero to indi- cate the presence of a BI field in the subheader; otherwise it is set to one to indicate the presence of a RAPID field in the subheader. R: Reserved bit, which is set to zero. BI: The BI field identifies the overload condition in the cell. The size of the BI field is 4 bit. RAPID: The RAPID field identifies the transmitted random-access preamble. The size of the RAPID field is 6 bits. If the RAPID in the MAC subheader of a MAC sub-PDU corresponds to one of the random-access preambles configured for SI request, MAC random-access response (RAR) is not included in the MAC sub-PDU. The MAC sub- header is octet aligned. There are a number of MAC CEs that are specified for the following purposes [9]: Duplication activation/deactivation The duplication activation/deactivation MAC CE consists of one octet that is identi- fied by a MAC PDU subheader with LCID index 56. It has a fixed size and contains eight D-fields, where the Di field indicates the activation/deactivation status of the 206 Chapter 2 PDCP duplication of the ith DRB associated with the RLC entities currently assigned to this MAC entity. SCell activation/deactivation The SCell activation/deactiva

tion MAC CE consists of either one or four octets that are identified by MAC PDU subheaders with LCID indices 58 or 57, respectively. It has a fixed size consisting of either a single octet or 4 octets and contains 7 or 31 C-fields (mapped to individual component carriers Cj) and one R-field. The DRX and long DRX command MAC CEs are identified by MAC PDU subhea- ders with LCID indices 60 and 59, respectively, both of which have a fixed size of zero bits. Timing advance command The timing advance command MAC CE is identified by a MAC PDU subheader with LCID index 61. It has a fixed size of one octet and includes timing advance group (TAG) identity (TAG ID) which indicates the TAG ID of the addressed TAG (2 bit); and timing advance command which indicates the index value TA = (0, 1, 2, , 63) that is used to control the amount of timing adjustment that MAC entity has to apply (6 bits). UE contention resolution identity The UE contention resolution identity MAC CE is identified by a MAC PDU sub- header with LCID index 62. It has a fixed 48 bit size and consists of a single field containing UE contention resolution identities. Semi-persistent (SP) CSI-RS/CSI-IM resource set activation/deactivation The network may activate and deactivate the configured SP CSI-RS/CSI-IM resource sets of a serving cell by sending the SP CSI-RS/CSI-IM resource set acti- vation/deactivation MAC CE. The configured SP CSI-RS/CSI-IM resource sets are initially deactivated upon configuration and after a handover. Aperiodic CSI trigger state sub-selection The network may select among the configured aperiodic CSI trigger states of a serv- ing cell by sending the aperiodic CSI trigger state sub-selection MAC CE. Transmission configuration indicator (TCI) states activation/deactivation UE-specific PDSCH The network may activate and deactivate the configured TCI states for physical downlink shared channel (PDSCH) of a serving cell by sending the TCI states acti- vation/deactivation for UE-specific PDSCH MAC CE. The configured TCI states for PDSCH a

re initially deactivated upon configuration and after a handover. TCI state indication for UE-specific PDCCH The network may indicate a TCI state for physical downlink control (PDCCH) reception for a CORESET of a serving cell by sending the TCI state indication for UE-specific PDCCH MAC CE. New Radio Access Layer 2/3 Aspects and System Operation 207 SP CSI reporting on physical uplink control channel (PUCCH) activation/deactivation The network may activate and deactivate the configured SP CSI reporting on PUCCH of a serving cell by sending the SP CSI reporting on PUCCH activation/ deactivation MAC CE. The configured SP CSI reporting on PUCCH is initially deactivated upon configuration and after a handover. SP SRS activation/deactivation The network may activate and deactivate the configured SP SRS resource sets of a serving cell by sending the SP SRS activation/deactivation MAC CE. The config- ured SP SRS resource sets are initially deactivated upon configuration and after a handover. PUCCH spatial relation activation/deactivation The network may activate and deactivate a spatial relation for a PUCCH resource of a serving cell by sending the PUCCH spatial relation activation/deactivation MAC CE. SP zero power (ZP) CSI-RS resource set activation/deactivation The network may activate and deactivate the configured SP ZP CSI-RS resource set of a serving cell by sending the SP ZP CSI-RS resource set activation/deactivation MAC CE. The configured SP ZP CSI-RS resource sets are initially deactivated upon configuration and after a handover. Recommended bit rate The recommended bit rate procedure is used to provide the MAC entity with infor- mation about the physical layer bit rate which the gNB recommends. An averaging window with default size of 2 seconds is applied. The gNB may transmit the recom- mended bit rate MAC CE to the receiver-side MAC entity to indicate the recom- mended bit rate for the UE for a specific logical channel and downlink/uplink direction. Single and multiple-entry power headroom report (PHR) The

single-entry PHR MAC CE is identified by a MAC PDU subheader with LCID index 57. It has a fixed size of two octets, including power headroom which indi- cates the PH level (6 bits) and PCMAXIm denoting the maximum permissible UE transmit power for carrier l and serving cell m. The multiple-entry PHR MAC CE is identified by a MAC PDU subheader with LCID index 56. It has a variable size and may include a bitmap, a Type 2 PH field and an octet containing the associated PCMAXIm field for the SpCell of the other MAC entity, as well as a Type 1 PH field and an octet containing the associated PCMAXIm for the primary cell (PCell). It may further include one or more Type X PH fields (x is either 1 or 3) and octets containing the associated PCMAX|m fields for serv- ing cells other than PCell indicated in the bitmap. The MAC entity determines whether power headroom value for an activated serving cell is based on a real trans- mission or a reference format by considering the configured grant(s) and the down- link control information (DCI) that has been received prior to the PDCCH occasion 208 Chapter 2 in which the first uplink grant for a new transmission is received after a PHR has been triggered. Configured grant confirmation The configured grant confirmation MAC CE is identified by a MAC PDU subhea- der with LCID index 55, which has a fixed size of zero bits. C-RNTI The C-RNTI MAC CE is identified by a MAC PDU subheader with LCID index 58. It has a fixed size and consists of a single field containing the C-RNTI. The length of the field is 16 bits. Buffer status report (BSR) The BSR MAC CEs consist of short BSR format (fixed size), long BSR format (variable size), short truncated BSR format (fixed size), and long truncated BSR for- mat (variable size). The BSR formats are identified by MAC PDU subheaders with LCID indices 59, 60, 61, and 62 corresponding to short truncated BSR, long trun- cated BSR, short BSR, and long BSR MAC CEs, respectively. The fields in the BSR MAC CE contain an LCG ID, denoting the LCG ID which ide

ntifies the group of logical channels whose buffer status are being reported, and one or several LCG; for the long BSR format, which indicate the presence of the buffer size field for the ith LCG. For the long truncated BSR format, this field indicates whether the ith LCG has data available. The buffer size field identifies the total amount of data available according to the data volume calculation procedure in references [10,11] across all logical channels of an LCG after the MAC PDU has been created, that is, after the logical channel prioritization procedure, which may result in setting the value of the buffer size field to zero. The amount of data is indicated in number of bytes. The size of the RLC and MAC headers are not considered in the buffer size calculation. The length of this field for the short BSR format and the short truncated BSR format is 5 bits and for the long BSR format and the long truncated BSR for- mat is 8 bits. For the long BSR format and the long truncated BSR format, the buffer size fields are included in ascending order based on the LCGi. For the long truncated BSR format the number of buffer size fields included is maximized, as long as it does not exceed the number of padding bits. The structure and content of some MAC CEs are shown in Fig. 2.8. As we mentioned earlier, a [transparent] MAC PDU may only consist of a MAC SDU whose size is aligned to a permissible transport block. This MAC PDU is used for transmis- sions on PCH, BCH, and DL-SCH including BCCH. Furthermore, a random-access response MAC PDU format is defined which consists of one or more MAC sub-PDUs and may contain some padding bits. Each MAC sub-PDU consists of a MAC sub-header with backoff indicator; a MAC sub-header with random access preamble identifier (RAPID), which is an acknowledgment for SI request; or a MAC sub-header with RAPID and MAC New Radio Access Layer 2/3 Aspects and System Operation 209 LCG ID Buffer size Octet 1 C-RNTI Octet 1 Short BSR and short truncated BSR MAC CE C-RNTI Octet 2 C-RNTI MAC CE Octet

1 Buffer size 1 Octet 2 UE contention resolution identity Octet 1 Buffer size 2 Octet 3 UE contention resolution identity Octet 2 UE contention resolution identity Octet 3 Buffer size m Octet m+1 UE contention resolution identity Octet 4 Long BSR and long truncated BSR MAC CE UE contention resolution identity Octet 5 UE contention resolution identity Octet 6 PH (Type 1, PCell) Octet 1 UE contention resolution identity MAC CE PCMAX Octet 2 Single entry PHR MAC CE Octet 1 SCell activation/deactivation MAC CE Octet 1 Duplication activation/deactivation MAC CE Figure 2.8 Some MAC control element formats [9]. RAR. A MAC sub-header with backoff indicator consists of five header fields E/T/R/R/BI (see Fig. 2.9). A MAC sub-PDU with backoff indicator is placed at the beginning of the MAC PDU, if included. The 'MAC sub-PDU(s) with RAPID' and 'MAC sub-PDU(s) with RAPID and MAC RAR' can be placed anywhere between MAC sub-PDU with backoff indicator and the padding bits. A MAC sub-header with RAPID consists of three header fields E/T/RAPID. Padding bits are placed at the end of the MAC PDU, if necessary. The presence and the length of the padding bits are based on the transport block size and the size of the MAC sub-PDU(s). The MAC entity is further responsible for distributing IP packets corresponding to each flow across different component carriers or cells, when carrier aggregation is utilized. The carrier aggregation relies on independent processing of the component carriers in the physi- cal layer, including control signaling, scheduling, and HARQ retransmissions. The MAC sublayer makes multi-carrier operation transparent to the upper sublayers/layers. The logical 210 Chapter 2 Octet 1 Timing advance command Octet 1 E/T/R/R/BI MAC subheader Timing advance command Uplink grant Octet 2 Uplink grant Octet 3 RAPID Octet 1 Uplink grant Octet 4 E/T/RAPID MAC subheader Uplink grant Octet 5 Temporary C-RNTI Octet 6 Temporary C-RNTI Octet 7 MAC random-access response E/T/R/R/BI E/T/RAPID E/T/RAPID MAC RAR subheader subheader subh

eader MAC sub-PDU 1 MAC sub-PDU 2 MAC sub-PDU 3 MAC sub-PDU 4 MAC sub-PDU n Padding (optional) (BI only) (RAPID only) (RAPID and RAR) MAC PDU with MAC RARs Figure 2.9 Structure of MAC subheaders and example MAC PDU with random-access response [9]. channels, including any MAC CEs, are multiplexed to form transport blocks per component carrier, where each component carrier has its own HARQ entity. In order to efficiently utilize the radio resources, the MAC sublayer in gNB includes sche- dulers that allocate radio resources in the downlink/uplink to the active users in the cell. The default mode of operation of a gNB scheduler is dynamic scheduling, in which the gNB makes scheduling decisions, typically once per slot, and sends the scheduling informa- tion to a group of devices. While slot-based scheduling is a common practice, neither the scheduling decisions, nor the actual data transmission are required to start or end at the slot boundaries, which is important for low-latency applications and unlicensed spectrum opera- tion. The downlink/uplink scheduling are independent and scheduling decisions can be inde- pendently made. The downlink scheduler dynamically controls the radio resources allocated to active devices in order to efficiently share the DL-SCH. The selection of transport format which includes transport block size, modulation and coding scheme, and antenna mapping, as well as logi- cal channel multiplexing for downlink transmissions are all managed by the gNB scheduler. New Radio Access Layer 2/3 Aspects and System Operation 211 The uplink scheduler serves a similar purpose, that is, dynamically controlling the transmis- sion opportunities of the active UEs and efficiently sharing of UL-SCH. The scheduling strategy of a gNB is vendor-dependent and implementation specific and is not specified by 3GPP. In general, the ultimate goal of all schedulers is to take advantage of the channel var- iations experienced by various devices and to schedule the transmissions/receptions to/from each device on the radi

o resources with advantageous channel conditions in time and fre- quency domains, that is, channel-dependent scheduling [8]. The downlink channel-dependent scheduling is supported through periodic/aperiodic chan- nel state information reports sent by the devices to the gNB, which provide information on the instantaneous downlink channel quality in time and frequency domains, as well as infor- mation to determine an appropriate antenna/beam configuration. In the uplink, the channel state information required for channel-dependent scheduling can be obtained from the sounding reference signals transmitted by each device to the gNB. In order to assist the uplink scheduler, the device can further transmit BSR (measuring the data that is buffered in the logical channel queues in the UE) and PHR (measuring the difference between the nominal UE maximum transmit power and the estimated power for uplink transmission) to the gNB using MAC CEs. This information can only be transmitted, if the device has been given a valid scheduling grant. While dynamic scheduling is the default operation mode of many base station schedulers, there are cases where semi-persistent scheduling (SPS) is pre- ferred due to reduced signaling overhead [17]. In the downlink, the gNB can dynamically allocate resources to the active UEs and identify them by their C-RNTIs on PDCCH(s). A UE always monitors the PDCCH(s) in order to find possible assignments according to its configured DRX cycles. When carrier aggregation is configured, the same C-RNTI applies to all serving cells. The gNB may preempt an ongo- ing PDSCH transmission to one UE with a delay-sensitive transmission to another UE. The gNB can configure the UEs to monitor interrupted transmission indications using INT-RNTI on a PDCCH. If a UE receives the interrupted transmission indication, it may assume that no useful information was intended for it by the resource elements included in the indication, even if some of those resource elements were already scheduled for that UE. Furthermore, the

gNB can allocate downlink resources for the initial HARQ transmissions to the UEs configured with SPS. In that case, the RRC signaling defines the periodicity of the configured downlink assignments while PDCCH addressed to CS-RNTI can either sig- nal and activate the configured downlink assignment or deactivate it, that is, a PDCCH addressed to CS-RNTI indicates that the downlink assignment can be implicitly reused according to the periodicity defined by the RRC signaling until it is deactivated [8]. The dynamically allocated downlink reception overrides the configured downlink assign- ment in the same serving cell, if they overlap in time. When carrier aggregation or band- width adaptation is configured, one configured downlink assignment can be signaled per 212 Chapter 2 serving cell or per BWP, respectively. In each serving cell, the gNB can only configure one active downlink assignment for a UE at a time; however, it can simulta- neously configure multiple active downlink assignments on different serving cells. Activation and deactivation of configured downlink assignments are independent among the serving cells [8]. In the uplink, the gNB can dynamically allocate resources to the active UEs and identify them by their C-RNTI on PDCCH(s). A UE always monitors the PDCCH(s) in order to find possible grants for uplink transmission according to its configured DRX cycles. When carrier aggregation is configured, the same C-RNTI applies to all serving cells. In addi- tion, the gNB can allocate uplink resources for the initial HARQ transmissions to the UEs with the configured grants (alternatively known as semi-persistent scheduling). Two types of configured uplink grants are specified in NR. In Type 1 configured grant the RRC sig- naling directly provides the configured uplink grant (including the periodicity), whereas in Type 2 configured grant, the RRC signaling defines the periodicity of the configured uplink grant, while PDCCH addressed to CS-RNTI can either signal and activate the con- figured uplink grant or d

eactivate it. Therefore, a PDCCH addressed to CS-RNTI indicates that the uplink grant can be implicitly reused according to the periodicity defined through RRC signaling until it is deactivated. The dynamically allocated uplink transmission over- rides the configured uplink grant in the same serving cell, if they overlap in time. The retransmissions other than repetitions are required to be explicitly allocated via PDCCH(s) [8]. If carrier aggregation or bandwidth adaptation is configured, one configured uplink grant can be signaled per serving cell or per BWP, respectively. In each serving cell, there is only one active configured uplink grant at a time. A configured grant for a serving cell in the uplink can either be Type 1 or Type 2. The activation and deactivation of config- ured grants for the uplink are independent among the serving cells for Type 2. In the case of supplemental uplink (SUL), a configured grant can only be signaled for one of the two uplink carriers of the cell [8]. When a downlink assignment is configured for SPS, the UE MAC entity will assume that the Nth downlink assignment occurs in the slot number that satisfies the following criterion [9]: where SFNstart-time start-time are the system frame number (SFN) and slot of the first transmission of PDSCH where the configured downlink assignment was (re)initialized. When a uplink grant is configured for a configured grant Type 1, the UE MAC entity will assume that the uplink grant repeats at symbols for which the following criterion is met [9]: New Radio Access Layer 2/3 Aspects and System Operation 213 When a uplink grant is configured for a configured grant Type 2, the UE MAC will assume that the uplink grant repeats with each symbol for which the following criterion is satisfied [9]: (N frame + + where start-time, slotstart-time and symbol start-time are the SFN, slot, and symbol corre- sponding to the first transmission opportunity of physical uplink shared channel (PUSCH), where the configured uplink grant was (re)initialized. Measurement

reports are required to enable the MAC scheduler to make scheduling deci- sions in the downlink and uplink. These reports include transport capabilities and measure- ments of UEs instantaneous radio conditions. The uplink BSRs are needed to support QoS- aware packet scheduling. In NR, uplink BSRs refer to the data that is buffered for an LCG in the UE. There are eight LCGs and two reporting formats in the uplink: a short format to report only one BSR (of one LCG) and a flexible long format to report several BSRs (up to eight LCGs). The uplink BSRs are transmitted using MAC CEs. When a BSR is triggered upon arrival of data in the transmission buffers of the UE, an SR is transmitted by the UE. The PHRs are needed to support power-aware packet scheduling. In NR, there are three types of reporting, that is, for PUSCH transmission, PUSCH and PUCCH transmission, and SRS transmission. In case of carrier aggregation, when no transmission takes place on an activated SCell, a reference power is used to provide a virtual report. The PHRs are trans- mitted using MAC CEs [8]. The HARQ functionality in the MAC sublayer ensures reliable transport of MAC PDUs between peer entities over the physical layer. A HARQ mechanism along with soft combin- ing can provide robustness against transmission errors. A single HARQ process supports one or multiple transport blocks depending on whether the physical layer is configured for downlink/uplink spatial multiplexing. An asynchronous incremental redundancy HARQ pro- tocol is supported in the downlink and uplink. The gNB provides the UE with the HARQ- ACK feedback timing either dynamically via DCI or semi-statically through an RRC con- figuration message. The UE may be configured to receive code-block-group-based trans- missions where retransmissions may be scheduled to carry a subset of the code blocks 214 Chapter 2 included in a transport block. The gNB schedules each uplink transmission and retransmis- sion using the uplink grant on DCI [8]. The HARQ protocol in NR uses multiple paralle

l stop-and-wait processes. When a transport block is received at the receiver, it attempts to decode the packet and to inform the transmit- ter about the outcome of the decoding process through an ACK bit indicating whether the decoding process was successful or requesting the retransmission of the transport block. The HARQ-ACKs are sent by the receiver based on a specific timing relationship between UL HARQ-ACKs and DL HARQ processes or based on the position of the ACK bit in the HARQ-ACK codebook when multiple ACKs are transmitted simultaneously. An asynchro- nous HARQ protocol is used for both downlink and uplink, that is, an explicit HARQ pro- cess number is used to identify a particular process, since the retransmissions are scheduled in the same way as the initial transmission (explicit scheduling). The use of an asynchro- nous UL HARQ protocol, instead of a synchronous one that was used in LTE, was deemed to be necessary to support dynamic TDD where there is no fixed UL/DL allocation. It also provides more flexibility in terms of prioritization of data flows and devices and is further useful for operation in unlicensed spectrum. The NR supports up to 16 HARQ processes. The larger number of HARQ processes (compared to LTE) was motivated by the disaggre- gated RAN architectures and consideration for remote radio heads and fronthaul transport delay, as well as the use of shorter slot durations at high-frequency bands. It must be noted that the larger number of maximum HARQ processes does not imply a longer roundtrip delay, since the decoding will typically succeed after a few retransmissions depending on the channel conditions. Note that the PDCP sublayer can provide in-sequence delivery; thus this function is not provided by the RLC sublayer in order to reduce the latency. A new feature of the NR HARQ protocol (compared to LTE) is the possibility for retrans- mission of code block groups, which is useful for very large transport blocks or when a transport block is partially preempted by another transmission.

As part of the channel cod- ing operation in the physical layer, a transport block is split into one or more code blocks with channel coding applied to each of the code blocks of up to 8448 bits in order to main- tain a reasonable complexity. In practice and in the presence of burst errors, only a few code blocks in the transport block may be corrupted and the majority of code blocks are correctly received. In order to correctly deliver the transport blocks to the destination MAC sublayer, it is sufficient to only retransmit the erroneous code blocks. Furthermore, to avoid the excessive control signaling overhead due to individual code block addressing by HARQ mechanism, code block groups have been defined. If per-CBG retransmission is configured, feedback is provided per-CBG and only the erroneously received code block groups are retransmitted. The CBG-based retransmissions are transparent to the MAC sublayer and are handled in the physical layer. From MAC sublayer perspective, the transport block is not correctly received until all CBGs are correctly received. It is not possible to mix the CBGs New Radio Access Layer 2/3 Aspects and System Operation 215 belonging to another transport block with retransmissions of CBGs belonging to the incor- rectly received transport block in the same HARQ process [17]. 2.2.2 Radio Link Control Sublayer The RLC sublayer is located between PDCP and MAC sublayers as shown in Fig. 2.1. The RRC sublayer controls and configures the RLC functions. The RLC functions are performed by the RLC entities. For an RLC entity configured at the gNB, there is a peer RLC entity configured at the UE and vice versa. An RLC entity receives/delivers RLC SDUs from/to the upper layer and sends/receives RLC PDUs to/from its peer RLC entity via lower layers. The RLC sublayer can operate in one of the three modes of operation defined as transparent mode, unacknowledged mode, and acknowledged mode. Depending on the mode of opera- tion, the RLC entity controls the usage of error correction, segmentation,

resegmentation, reassembly, and duplicate detection of SDUs. An RLC entity in any mode can be config- ured either as a transmitting or a receiving entity. The transmitting RLC entity receives RLC SDUs from the upper layer and sends RLC PDUs to its peer receiving RLC entity via lower layers. The receiving RLC entity delivers RLC SDUs to the upper layer and receives RLC PDUs from its peer transmitting RLC entity via lower layers. Fig. 2.10 illustrates the Upper layer RLC channel Transmitting Receiving Transmitting Receiving AM RLC entity TM RLC entity TM RLC entity UM RLC entity UM RLC entity Logical channel Lower layers Radio interface Lower layers Logical channel Receiving Transmitting Receiving Transmitting AM RLC entity TM RLC entity TM RLC entity UM RLC entity UM RLC entity RL€ channel Upper layer Figure 2.10 High-level architecture of RLC sublayer [10]. 216 Chapter 2 high-level architecture of the RLC sublayer. The [octet-aligned] RLC SDUs of variable sizes are supported for RLC entities. Each RLC SDU is used to construct an RLC PDU without waiting for notification from the MAC sublayer for a transmission opportunity. In the case of RLC-UM and RLC-AM entities, an RLC SDU may be segmented and trans- ported using two or more RLC PDUs based on the notification(s) from the lower layer [10]. In RLC-TM, no RLC headers are added. The RLC-UM supports segmentation and duplicate detection, while the RLC-AM supports retransmission of erroneous packets. A key differ- ence with LTE is that the NR RLC sublayer does not handle in-sequence delivery of SDUs to the upper layers. Eliminating the in-sequence delivery function from the RLC sublayer reduces the overall latency since the correctly received packets do not have to wait for retransmission of an earlier (missing) packet before being delivered to the higher layers. Another difference is the removal of concatenation function from the RLC protocol to allow RLC PDUs to be assembled prior to receiving the uplink scheduling grant. By eliminating the concatenation from RLC fu

nctions the RLC PDUs can be assembled in advance and upon receipt of the scheduling decision the device can forward a number of RLC PDUs to the MAC sublayer depending on the scheduled transport block size. The RLC retransmission mechanism is used to provide error correction of data delivered to higher layers. The retransmission protocol operating between the RLC entities in the receiver and transmitter monitors the sequence numbers indicated in the headers of the incoming PDUs. The receiving RLC entity can identify missing PDUs based on the RLC sequence number which is independent of the PDCP sequence number. The status reports are sent to the transmitting RLC entity, requesting retransmission of the missing PDUs. Based on the received status report, the RLC entity at the transmitter can retransmit the missing PDUs. Even though the RLC sublayer is capable of correcting transmission errors, the error correction in most cases is handled by the MAC sublayer HARQ protocol. However, RLC sublayer and MAC sublayer retransmission mechanisms are meant for dif- ferent purposes to achieve a highly reliable transmission for certain applications. The HARQ mechanism allows fast retransmissions and feedback based on the success or failure of a downlink transmission after receiving a transport block. For the uplink transmis- sions, no explicit feedback needs to be transmitted because the receiver and scheduler are located in the same node. While in theory, it is possible to attain a very low block error rate using HARQ mechanism, in practice, it comes at a cost additional signaling and radio resources as well as increased power consumption. Therefore, the target block error rate in practice is less than 1%, which inevitably results in a residual error. For many applications such as voice over IP (VoIP), this residual error rate is sufficiently low and acceptable (because voice decoders can tolerate some level of frame erasure); however, there are use cases where lower block error rates are required, for example, transmission of

RRC and NAS messages. New Radio Access Layer 2/3 Aspects and System Operation 217 A sufficiently low block error rate is not only required for URLLC services, but it also is important from a system-level perspective in terms of sustaining data rate performance. The TCP protocol (at transport layer) requires virtually error-free delivery of packets to the peer TCP layer. As an example, to obtain sustained data rates in the excess of 100 Mbps in TCP/ IP applications, a packet error rate of less than 10- is required, because the TCP protocol would consider packet errors due to network congestion and would trigger the congestion- avoidance mechanism to decrease the data rate, causing reduction of the overall data rate performance of the system. It must be noted that the infrequent transmission of RLC status reports compared to more frequent HARQ retransmission makes obtaining a reliability level 10-5 using RLC ARQ retransmissions more practical [17]. Therefore, complementing MAC sublayer HARQ protocol with RLC sublayer ARQ mechanism would achieve rela- tively lower latency and reasonable feedback overhead to satisfy the stringent requirements of URLLC applications. The RLC sublayer ARQ mechanism retransmits RLC SDUs or RLC SDU segments based on RLC status reports, where polling may be used to request RLC status reports. An RLC receiver can also trigger RLC status report after detecting a missing RLC SDU or RLC SDU segment [8]. The RLC-TM entity is configured to trans- mit/receive RLC PDUs through BCCH, DL/UL CCCH, and paging control channel logical channels. The RLC-UM entity is configured to transmit/receive RLC PDUs via downlink or uplink dedicated traffic channel (DTCH). The RLC-AM entity can be configured to trans- mit/receive RLC PDUs through downlink or uplink dedicated control channel (DCCH) or DL/UL DTCH logical channels. Functional models of RLC-AM and RLC-UM entities are illustrated in Fig. 2.11, where the functions that are only related to RLC-AM mode (ARQ functions) are marked with a dark color. The RLC P

DUs and SDUs are bit strings that are octet-aligned. The TM data (TMD PDU) consists only of a data field and does not include any RLC headers. In RLC-AM and RLC- UM modes, a sequence number is generated and attached to the incoming SDUs using 6 or 12 bits for the RLC-UM and 12 or 18 bits for the RLC-AM. The sequence number is included in the RLC PDU header as shown in Fig. 2.12. If the SDU is not segmented, the RLC PDU consists of the RLC SDU and a header, which allows the RLC PDUs to be gen- erated in advance as the header does not depend on the transport block size. However, depending on the transport block size after multiplexing at MAC sublayer, the size of the last RLC PDU in a transport block may not match the RLC SDU size, thus requiring divid- ing the SDU into multiple segments. If no segmentation is done, padding need to be used which would adversely impact the spectral efficiency. As a result, dynamically varying the number of RLC PDUs in a transport block along with segmentation to adjust the size of the last RLC PDU, ensures that the transport block is efficiently utilized. Segmentation is done by dividing the last preprocessed RLC SDU into two segments, the header of the first seg- ment is updated, and a new header is added to the second segment as shown in Fig. 2.3. Each RLC SDU segment carries the same sequence number as the original SDU and the 218 Chapter 2 RLC-AM channel RLC-AM channel [RLC-UM channel] [RLC-UM channel] Generate RLC header and store in RLC control SDU reassembly transmission buffer Remove RLC header Segmentation Retransmission modify RLC header buffer Reception buffer Add RLC header Routing RLC-AM: DTCH/DCCH RLC-AM: DTCH/DCCH RLC-UM: DTCH RLC-UM: DTCH RLC-AM mode only Figure 2.11 RLC-AM and RLC-UM entity models [10]. sequence number is part of the RLC header. To distinguish an unsegmented RLC PDU from a segmented one, a segmentation information field is included in the RLC header, indicating whether the PDU is a complete SDU, the first segment of the SDU, the last seg- ment of th

e SDU, or a segment between the first and last segments of the SDU. Furthermore, in the case of a segmented SDU, a 16 bit segmentation offset (SO) is included in all segments except the first one to indicate which byte of the SDU is represented by the segment. The RLC header may further include a poll (P) bit, which is used to request status New Radio Access Layer 2/3 Aspects and System Operation 219 Octet 1 Octet 1 Octet 2 Octet N Octet 3 TMD PDU Octet 4 Octet 5 Octet 1 Octet 2 Octet N Octet 3 UMD PDU with 12-bit SN and so Octet 4 Octet 5 ACK_SN Octet 1 ACK_SN Octet 2 Octet N E1RRRRRRR Octet 3 AMD PDU with 12-bit SN and so NACK_SN Octet 4 NACK_SN Octet 5 Octet 1 NACK_SN Octet 6 Octet 2 NACK_SN Octet 7 Octet 3 SOstart Octet 8 Octet 4 SOstart Octet 9 Octet 5 SOend Octet 10 Octet 6 SOend Octet 11 NACK range Octet 12 Octet N NACK_SN Octet 13 AMD PDU with 18-bit SN with so NACK_SN Octet 14 STATUS RLC PDU with 12-bit SN Figure 2.12 Example RLC PDU formats 10]. report for RLC-AM, and a data/control (D/C) indicator, indicating whether the RLC PDU contains data to/from a logical channel or control information required for RLC operation [10,17]. The retransmission of missing RLC PDUs is one of the main services of the RLC-AM entity. While MAC sublayer HARQ protocol provides some level of error correction, when combined with the ARQ retransmission mechanism, they can provide a higher level of reli- ability. The missing RLC PDUs can be detected by inspecting the sequence numbers of the received PDUs, and a retransmission requested from the transmitting side. The RLC-AM mode in NR is similar to LTE except the in-sequence delivery service is not supported in NR. As we mentioned earlier, eliminating the in-sequence delivery from the RLC helped reduce the overall latency. It further reduces the buffering requirements at the RLC sub- layer. In LTE, which supports the in-sequence delivery service at RLC sublayer, an RLC SDU cannot be forwarded to higher layers unless all previous SDUs have been correctly received. A single missin

g SDU can significantly delay the delivery of the subsequent 220 Chapter 2 SDUs. The RLC-AM is bidirectional which means the data may flow in both directions between the two peer entities (see Fig. 2.11), enabling the acknowledgement of the received RLC PDUs to the transmitting entity. The information about missing PDUs is provided by the receiving entity to the transmitting entity in the form of status reports. The status reports can either be transmitted autonomously by the receiver or requested by the transmitter. The PDUs in transit are tracked by the sequence number in the header. The transmitting and receiving RLC entities maintain two windows in RLC-AM, that is, the transmission and reception windows. The PDUs in the transmission window are only eligible for transmis- sion, thus PDUs with sequence numbers less than the start of the window have already been acknowledged by the receiving RLC entity. In the same way, the receiving entity only accepts PDUs with sequence numbers within the reception window. The receiver discards any duplicate PDUs and delivers only one copy of each SDU to higher layers. The concept of RLC retransmissions is exemplified in Fig. 2.13, where transmitting and receiving RLC entities are illustrated. When operating in RLC-AM mode, each RLC entity has transmitting and receiving functionality; nevertheless, in this example we only show one direction as the other direction is identical [17]. In this example, the PDUs numbered k to k + 4 are awaiting transmission in the transmis- sion buffer. At time to, it is assumed that the PDUs with sequence number SN < have been transmitted and correctly received; however, only PDUs with sequence number SN < k - 1 have been acknowledged by the receiver. The transmission window starts at k, that is, the first unacknowledged PDU, while the reception window starts at k + 1, that is, the next PDU expected to be received. Upon reception of kth PDU, the SDU is reassembled and delivered to the higher layers. For a PDU containing an unsegmented SDU the reas

sem- bly function only involves header removal, but in the case of a segmented SDU, the SDU cannot be delivered to upper layers until the PDUs carrying all segments of the SDU arrive at the receiver. The transmission of PDUs continues such that at time t1, PDUs k + 1 and k + 2 are transmitted but, at the receiving end, only PDU k + 2 has arrived. As soon as a complete SDU is received, it is delivered to the higher layers; thus PDU k + 2 is forwarded to the higher sublayer without waiting for the missing PDU k + 1, which could be undergo- ing retransmission by the HARQ protocol. Therefore, the transmission window remains unchanged, since none of the PDUs with SN > k have been acknowledged by the receiver. This could result in retransmission of these PDUs given that the transmitter is not aware of whether they have been correctly received. The reception window is not updated when PDU + 2 arrives because of the missing PDU k + 1. At this point, the receiver starts the t- Reassembly timer. If the missing PDU k + 1 is not received before the timer expires, a retransmission is requested. If the missing PDU arrives at time t2 before the timer expires, the reception window is advanced, the reassembly timer is stopped, and PDU k + 1 is deliv- ered for reassembly with SDU k + 1. The RLC sublayer is also responsible for duplicate detection using the same sequence number that is used for retransmission management. New Radio Access Layer 2/3 Aspects and System Operation 221 RLC PDUs waiting for initial RLC PDUs transmitted but RLC PDUs transmitted and transmission unackowledged acknowledged Transmitting Receiving Transmitting Receiving Transmitting Receiving entity entity entity entity entity entity Transmitting Receiving Transmitting Receiving Transmitting Receiving entity entity entity entity entity entity Figure 2.13 Example RLC-AM operation [17]. If PDU k + 2 arrives again within the reception window, it will be discarded due to exis- tence of another copy. The transmission continues with PDUs k + 3, k + 4, and k + 5, as

shown in Fig. 2.13. At time t3, PDUs with SN < k + 5 have been transmitted, but only PDU k + 5 has arrived and PDUs k + 3 and k + 4 are missing. Similar to the previous case, this causes the t-Reassembly timer to start. However, in this example no PDUs arrive prior to the expiration of the timer. The expiration of the timer at time t4 triggers the receiver to send a control PDU containing a status report indicating the missing PDUs to its peer entity. The control PDUs have higher priority than data PDUs to avoid delayed transmission of the status reports, which would adversely impact the retransmission delay. Upon receipt of the status report at time t5, the transmitter is informed that PDUs up to k + 2 have been correctly 222 Chapter 2 received; thus the transmission window is advanced. The missing PDUs k + 3 and k + 4 are retransmitted and are later correctly received. At time t6, all PDUs, including the retransmis- sions, have been transmitted and successfully received. Since k + 5 was the last PDU in the transmission buffer the transmitter requests a status report from the receiver by setting a flag in the header of the last RLC data PDU. Upon reception of the PDU with the flag set, the receiver will respond by transmitting the requested status report, acknowledging all PDUs with SN k + 5. The reception of the status report allows the transmitter to declare all PDUs as correctly received and to advance the transmission window. The status reports can be triggered for many reasons. However, to avoid transmission of excessive number of status reports, a t-StatusProhibit timer is used where status reports cannot be transmitted more than once within a time interval determined by the timer. In the above example, we assumed that each PDU carries an unsegmented SDU. Segmented SDUs are handled in the same way, but an SDU cannot be delivered to the higher layers until all segments have been received. The status reports and retransmissions operate on individual segments, and only the missing segments of a PDU are retran

smitted [17]. 2.2.3 Packet Data Convergence Protocol Sublayer 2.2.3.1 PDCP Services and Functions The services and functions of the PDCP sublayer on the user-plane include sequence num- bering; header compression and decompression; transfer of user data; reordering and dupli- cate detection; in-sequence delivery; PDCP PDU routing in multi-connectivity; retransmission of PDCP SDUs; ciphering, deciphering, and integrity protection; PDCP SDU discard; PDCP re-establishment and data recovery for RLC-AM; PDCP status reporting for RLC-AM; duplication of PDCP PDUs and duplicate discard indication to lower layers. The main services and functions of the PDCP sublayer on the control-plane consist of sequence numbering; ciphering, deciphering, and integrity protection; transfer of control-plane data; reordering and duplicate detection; in-sequence delivery; duplication of PDCP PDUs and duplicate discard indication to lower layers. The PDCP protocol performs (optional) IP-header compression, followed by ciphering, for each radio bearer. A PDCP header is added, carrying information required for deciphering in the other end, as well as a sequence number used for retransmission and in-sequence delivery [8]. In some services and applications such as VoIP, interactive gaming, and multimedia messaging, the data payload of the IP packet is almost the same size or even smaller than the header itself. Over the end-to-end connection comprising multiple hops, these protocol headers are extremely important, but over a single point-to-point link, these headers serve no useful purpose. It is possi- ble to compress these headers, and thus save the bandwidth and use the expensive radio resources more efficiently. The header compression also provides other important benefits, such as reduction in packet loss and improved interactive response time. The payload header New Radio Access Layer 2/3 Aspects and System Operation 223 compression is the process of suppressing the repetitive portion of payload headers at the sender side and restoring th

em at the receiver side of a low-bandwidth/capacity-limited link. The use of header compression has a well-established history in transport of IP-based payloads over capacity-limited wireless links where more bandwidth efficient transport methods are required. The Internet Engineering Task Force (IETF) 2 has developed several header compression proto- cols that are widely used in telecommunication systems. The header compression mechanism used in 3GPP LTE and NR standards is based on IETF RFC 5795, ROHC framework. The IP together with transport protocols such as TCP or UDP and application layer protocols (e.g., RTP) are described in the form of payload headers. The information carried in the header helps the applications to communicate over large distances connected by multiple links or hops in the network. This information consists of source and destination addresses, ports, protocol identi- fiers, sequence numbers, error checksums, etc. Under nominal conditions, most of the informa- tion carried in packet-headers remains the same or changes in specific patterns. By observing the fields that remain constant or change in specific patterns, it is possible either not to send them in each packet, or to represent them in a smaller number of bits than would have been orig- inally required. This process is referred to as header compression. The PDCP sublayer, at the transmit side, performs encryption of IP packets to protect user privacy, and additionally for the control-plane messages, performs integrity protection to ensure that control messages originate from the correct source and can be authenticated. At the receiver side the PDCP performs the corresponding decryption and decompression oper- ation. The PDCP further discards duplicate PDUs and performs in-sequence delivery of the packets. Upon handover, the undelivered downlink PDUs will be forwarded by the PDCP sublayer of the source gNB to peer entity of the target gNB for delivery to the UE. The PDCP entity in the device will also handle retransmission of all up

link packets that were not delivered to the gNB given that the HARQ buffers are flushed upon handover. In this case, some PDUs may be received in duplicate form, possibly from both the source and the target gNBs. In this case, the PDCP will remove any duplicates. The PDCP entity can also be configured to perform reordering to ensure in-sequence delivery of SDUs to higher layer protocols. In some cases, the PDUs can be duplicated and transmitted on multiple cells, increasing the likelihood of their correct reception at the receiving end, which can be useful for services requiring very high reliability. At the receiving end, the PDCP duplicate removal functionality removes any duplicates. The PDCP plays an important role in dual connectivity, where a device is connected to two cells, that is, an MCG and an SCG. The two cell groups can be handled by different gNBs. A radio bearer is typically handled by one of the cell groups, but there is also the possibility for split bearers, in which case one IETF: https://www.ietf.org/. 224 Chapter 2 radio bearer is handled by both cell groups. In this case the PDCP entity is responsible for distributing the data between the MCG and the SCG [17]. The PDCP entities are located in the PDCP sublayer. Several PDCP entities may be defined for a UE. Each PDCP entity carries the data of one radio bearer, which is associ- ated with either the control-plane or the user-plane. Fig. 2.14 illustrates the functional block diagram of a PDCP entity. For split bearers, the routing function is performed in the transmitting PDCP entity. The PDCP sublayer provides services to RRC and SDAP sub- layers. The PDCP provides transfer of user-plane and control-plane data; header compres- sion; ciphering; integrity protection services to the higher layer protocols. The maximum size of a data or control PDCP SDU supported in NR is 9000 bytes [8]. It must be noted that an NR system provides protection against eavesdropping and modification attacks. Signaling traffic (RRC messages) is encrypted and integrit

y protected. User-plane traffic (IP packets) is encrypted and can be integrity protected. User-plane integrity protection is a new feature (relative to LTE) that is useful for small-data transmissions, and particularly Octet 1 Octet 1 Octet 2 Octet N Octet 3 TMD PDU Octet 4 Octet 5 Octet 1 Octet 2 Octet N Octet 3 UMD PDU with 12-bit SN and so Octet 4 Octet 5 ACK_SN Octet 1 ACK_SN Octet 2 Octet N IRRRRRRR Octet 3 AMD PDU with 12-bit SN and so NACK_SN Octet 4 NACK_SN Octet 5 Octet 1 NACK_SN Octet 6 Octet 2 NACK_SN Octet 7 Octet 3 SOstart Octet 8 Octet 4 SOstart Octet 9 Octet 5 SOend Octet 10 Octet 6 SOend Octet 11 NACK range Octet 12 Octet N NACK_S SN Octet 13 AMD PDU with 18-bit SN with so NACK_SN Octet 14 STATUS RLC PDU with 12-bit SN Figure 2.14 Functional block diagram of the PDCP entities [11]. New Radio Access Layer 2/3 Aspects and System Operation 225 for constrained IoT devices. Data traffic including voice calls, Internet traffic, and text messages are protected using encryption. The device and the network mutually authenti- cate each other and use integrity-protected signaling. This setup makes nearly impossible for an unauthorized party to decrypt and read the information that is communicated over the air. Although integrity protection of user-plane data is supported in 5G networks, this feature is not used in E-UTRA-NR-DC (EN-DC) scenarios since LTE systems only pro- vide integrity protection of control-plane messages. When duplication is configured for a radio bearer by RRC, a secondary RLC entity is added to the radio bearer to handle the duplicated PDCP PDUs, where the logical chan- nels corresponding to the primary and the secondary RLC entities are referred to as the primary logical channel and the secondary logical channel, respectively. Therefore, the duplication function at PDCP sublayer consists of submitting the same PDCP PDUs twice, that is, to the primary RLC entity and to the secondary RLC entity. With two independent transmission paths, packet duplication would improve the transmission rel

iability and reduce the latency, which is especially advantageous to URLLC services [8]. The PDCP control PDUs are not duplicated and are always submitted to the primary RLC entity. When configuring duplication for a DRB, the RRC sublayer sets the initial state to be either activated or deactivated. After the configuration, the state can be dynamically con- trolled by means of a MAC CE. In dual connectivity, the UE applies the MAC CE com- mands regardless of their origin (MCG or SCG). When duplication is configured for an SRB, the state is always active and cannot be dynamically controlled. When activating duplication for a DRB, NG-RAN ensures that at least one serving cell is activated for each logical channel of the DRB. When the deactivation of SCells leaves no activated serving cells for the logical channels of the DRB, NG-RAN ensures that duplication is also deactivated [8]. When duplication is activated, the original PDCP PDU and the corresponding duplicate are not transmitted on the same carrier. The primary and secondary logical channels can either belong to the same MAC entity (referred to as CA duplication) or to different ones (referred to as DC duplication). In CA duplication, logical channel mapping restrictions are applied to the MAC sublayer to ensure that the primary and secondary logical channels are not sent on the same carrier. When duplication is deactivated for a DRB, the secondary RLC entity is not reestablished, the HARQ buffers are not flushed, and the transmitting PDCP entity should indicate to the secondary RLC entity to discard all duplicated PDCP PDUs. In addition, in case of CA duplication, the logical channel mapping restrictions of the primary and secondary logical channels are relaxed as long as duplication remains deac- tivated. When an RLC entity acknowledges the transmission of a PDCP PDU, the PDCP entity notifies the other RLC entity to discard that PDU. When the secondary RLC entity 226 Chapter 2 Voice datagrams with IPv6/UDP/RTP headers Voice datagrams with ROHC headers Voice

datagrams with IPv6/UDP/RTP headers compressor Radio link decompressor Reference header Reference header Figure 2.15 Example ROHC compression/decompression of RTP/UDP/IP headers for communication over a radio link 16]. reaches the maximum number of retransmissions for a PDCP PDU, the UE informs the gNB, but does not trigger RLF³ 3 [8]. 2.2.3.2 Header Compression Function There are multiple header compression algorithms referred to as profiles in 3GPP speci- fications [11]. Each profile is specific to a particular network layer, transport layer, or upper layer protocol combination, for example, TCP/IP or RTP/UDP/IP. The PDCP enti- ties associated with DRBs carrying user-plane data can be configured by upper layers to use header compression. Each PDCP entity uses at most one ROHC compressor instance and at most one ROHC decompressor instance. As we mentioned earlier, the ROHC algorithm reduces the size of transmitted RTP/UDP/IP header by removing the redundancies. This mechanism starts by classifying header fields into different classes according to their variation pattern. The fields that are classified as inferred are not sent. The static fields are sent initially and then are not sent anymore and the fields with varying information are always sent. The ROHC mechanism is based on a context, 4 which is maintained, by both ends, that is, the compressor and the decompressor (see Fig. 2.15). The context encompasses the entire header and ROHC information. Each context has a context ID, which identifies the flows. The ROHC scheme operates in one of the following three operation modes [16]: The UE declares a RLF when one of the following criteria is met: (1) Expiry of a timer started after indication of radio problems from the physical layer (if radio problems are recovered before the timer expires, the UE stops the timer), random-access procedure failure, or RLC failure. After an RLF is declared, the UE stays in RRC_CONNECTED and selects a suitable cell and then initiates RRC connection reestablishment. The UE further

where a return path from decompressor to compressor is unavailable or undesirable. Optimistic mode (O) is a bidirectional mode similar to the unidirectional mode, except that a feedback channel is used to send error recovery requests and (option- ally) acknowledgements of significant context updates from the decompressor to compressor. The O-mode aims to maximize compression efficiency and sparse usage of the feedback channel. 3. Reliable mode (R) is a bidirectional mode which differs in many ways from the previ- ous two modes. The most important differences include intensive use of feedback chan- nel and a strict logic at both compressor and decompressor that prevents loss of context synchronization between compressor and decompressor except for very high residual error rates. The U-mode is used when the link is unidirectional or when feedback is not possible. For bidirectional links, O-mode uses positive feedback packets (ACK) and R-mode use positive and negative feedback packets [ACK and NACK]. The ROHC mecha- nism always starts header compression using U-mode even if it is used over a bidi- rectional link and it does not send retransmissions when an error occurs; thus the erroneous packet is dropped. The ROHC feedback is used only to indicate to the compressor side that there was an error and probably the context is damaged. After receiving a negative feedback the compressor always reduces its compression level. The ROHC compressor has three compression states defined as follows [16]: 1. Initialization and refresh (IR), where the compressor has been just created or reset and full packet-headers are sent. 2. First order (FO), where the compressor has detected and stored the static fields such as IP addresses and port numbers on both sides of the connection. Second order (SO), where the compressor is suppressing all dynamic fields such as RTP sequence numbers and sending only a logical sequence number and partial check- sum to make the other side generate the headers based on prediction and verify the headers o

f the next expected packet. Each compression state uses a different header format in order to send the header informa- tion. The IR compression state establishes the context, which contains static and dynamic header information. The FO compression state provides the change pattern of dynamic fields. The SO compression state sends encoded values of sequence number (SN) and time- stamp (TS), forming the minimal size packets (Figs. 2.6-2.15). Using this header format, all header fields can be generated at the other end of the radio link using the previously established change pattern. When some updates or errors occur, the compressor returns to 228 Chapter 2 No static Repeated failures Success No dynamic Success Repeated failures Success Compressor states Modes of operation Decompressor states Figure 2.16 ROHC state machines [16]. upper compression states. It only transitions to the SO compression state after retransmitting the updated information and reestablishing the change pattern in the decompressor. In the U-mode the feedback channel is not used. To increase the compression level an opti- mistic approach is used for compressor to ensure that the context has been correctly estab- lished at decompressor side. This means that compressor uses the same header format for a number of packets. Since the compressor does not know whether the context is lost, it also uses two timers, to be able to return to the FO and IR compression states. The decompressor works at the receiving end of the link and decompresses the headers based on the header fields' information of the context. Both the compressor and the decompressor use a context to store all the information about the header fields. To ensure correct decompression the context should be always synchronized. The decompressor has three states as follows: (1) no context (NC) where there is NC synchronization, (2) static context (SC) where the dynamic information of the context has been lost, and (3) full context (FC) when the decompressor has all the information about hea

der fields. In FC state the decompressor tran- sitions to the initial states as soon as it detects corruption of the context. The decompressor uses the "k out of n" rule by looking at the last n packets with CRC failures. If k CRC fail- ures have occurred, it assumes the context has been corrupted and transitions to an initial state (SC or NC). The decompressor also sends feedback according to the operation mode (Fig. 2.16). The values of the ROHC compression parameters that determine the efficiency and robust- ness are not defined in ROHC specification and are not negotiated initially but are stated as implementation dependent. The values of these parameters remain fixed during the com- pression process. The compression parameters are defined as follows [16]: L: In U-mode and O-mode, the ROHC compressor uses a confidence variable L in order to ensure the correct transmission of header information. Timer_1 (IR_TIMEOUT): In U-mode, the compressor uses this timer to return to the IR compression level and periodically resends static information. New Radio Access Layer 2/3 Aspects and System Operation 229 Timer_2 (FO_TIMEOUT): The compressor also uses another timer in U-mode, and this timer is used to go downward to FO compression level, if the compressor is working in SO compression level. Sliding window width (SWW): The compressor, while compressing header fields such as sequence number (SN) and TS, utilizes window-based least significant bit (W_LSB) encoding that uses a sliding window of width equal to SWW. W_LSB encoding is used to compress those header fields whose change pattern is known. When using this encoding, the compressor sends only the least significant bits. The decompressor uses these bits to construct the original value of the encoding fields. and n: The ROHC decompressor uses a "k out of n" failure rule, where k is the num- ber of packets received with an error in the last n transmitted packets. This rule is used in the state machine of the decompressor to assume the damage of context and move downw

ards to a state after sending a NACK to the compressor, if bidirectional link is used. The decompressor does not assume context corruption and remains in the current state until k packets arrive with error in the last n packets. 2.2.3.3 Ciphering and Integrity Protection Functions The RRC confidentiality protection is provided by the PDCP sublayer between a UE and the serving gNB. The user-plane security policy indicates whether the user-plane confidenti- ality and/or user-plane integrity protection is activated for all DRBs belonging to the PDU session. The input parameters to the 128-bit NR encryption algorithm (NEA) (or alterna- tively encryption algorithm for 5G), which is used for ciphering, are 128-bit cipher key referred to as KEY (KRRCenc), 32-bit COUNT (PDCP COUNT), 5-bit radio bearer identity BEARER, 1-bit direction of the transmission, that is, DIRECTION, and the length of the keystream required identified as LENGTH. The DIRECTION bit is set to zero for uplink and one for downlink. Fig. 2.17 illustrates the use of the ciphering algorithm NEA to encrypt plain text by applying a keystream using a bit-wise binary addition of the plaintext block and the keystream block. The plaintext block may be recovered by generating the same keystream block using the same input parameters and applying a bit-wise binary addi- tion with the ciphertext block. Based on the input parameters, the algorithm generates the output keystream block keystream which is used to encrypt the input plaintext block to pro- duce the output ciphertext block. The input parameter LENGTH only denotes the length of the keystream block and not its content [2,11]. The ciphering algorithm and key to be used by the PDCP entity are configured by upper layers and the ciphering method is applied according to the security architecture of 3GPP system architecture evolution (SAE, which is the LTE system architecture). The ciphering function is activated by upper layers. After security activation, the ciphering function is applied to all PDCP PDUs indica

ted by upper layers for downlink and uplink transmissions. The COUNT value is composed of a hyper- frame number (HFN) and the PDCP SN. The size of the HFN part in bits is equal to 32 230 Chapter 2 Count Bearer Direction Length Count Bearer Direction Length Encryption algorithm for 5G Encryption algorithm for 5G (NEA) (NEA) Keystream Keystream block block Plaintext block Ciphertext block Plaintext block Transmitter Receiver PDCP SN Count Figure 2.17 Illustration of the ciphering and deciphering procedures [2,11]. minus the length of the PDCP SN. The PDCP does not allow COUNT to wrap around in the downlink and uplink; thus it is up to the network to prevent it from happening [2,8,11]. As shown in Fig. 2.18, the input parameters to the 128-bit NR integrity protection algorithm (NIA) (or alternatively integrity protection algorithm for 5G) are the RRC messages denoted as MESSAGE, 128-bit integrity key KRRCint referred to as KEY, 5-bit bearer identity BEARER, 1-bit direction of transmission denoted as DIRECTION, and a bearer specific direction-dependent 32-bit input COUNT which corresponds to the 32-bit PDCP COUNT. The RRC integrity checks are performed both in the UE and the gNB. If the gNB or the UE receives a PDCP PDU which fails the integrity check with faulty or missing message authentication code (MAC-I) after the start of integrity protection, the PDU will be dis- carded. The DIRECTION bit set to zero for uplink and one for downlink. The bit length of the MESSAGE is LENGTH. Based on these input parameters, the sender computes a 32-bit message authentication code (MAC-I/NAS-MAC)™ 5 using the integrity protection algorithm In cryptography a MAC-I is a cryptographic checksum on data that uses a session key to detect both acciden- tal and intentional modifications of the data. A MAC (not to be confused with medium access control MAC) requires two inputs: a message and a secret key known only to the originator of the message and its intended recipient(s). This allows the recipient of the message to verify the integr

ity of the message and authenticate that the message sender has the shared secret key. Any mismatch between the sender's and receiver's calcu- lated MAC-I values would invalidate the message. There are four types of message authentication codes: unconditionally secure, hash function-based, stream cipher-based, and block cipher-based. In the past, the most common approach to creating a message authentication code was to use block ciphers; however, hash- based MACs which use a secret key in conjunction with a cryptographic hash function to produce a hash, have become more widely used. New Radio Access Layer 2/3 Aspects and System Operation 231 Count Bearer Direction Message Count Bearer Direction Message Integrity algorithm for 5G Integrity algorithm for 5G (NIA) (NIA) MAC-I/NAS-MAC XMAC-I/XNAS-MAC Compare MAC-I/NAS- Valid/ Message Message + MAC-I (NAS-MAC) MAC and XMAC-I/XNAS- invalid Transmitter Receiver PDCP SN Count Figure 2.18 Integrity protection and verification procedures [2,11]. NIA. The message authentication code is then appended to the message when sent. For integrity protection algorithms, the receiver computes the expected message authentication code (XMAC-I/XNAS-MAC) on the message received in the same way that the sender computed its message authentication code on the message sent and verifies the integrity of the message by comparing it to the received message authentication code, that is, MAC-I/ NAS-MAC. The integrity protection algorithm and key to be used by the PDCP entity are configured by upper layers and the integrity protection method is applied according to secu- rity architecture of 3GPP SAE [2]. The integrity protection function is activated by upper layers. Following the security activation, the integrity protection function is applied to all PDUs including and subsequent to the PDU indicated by upper layers for downlink and uplink transmissions. As the RRC message which activates the integrity protection function is itself integrity protected with the configuration included in that RRC

message, the mes- sage must be decoded by RRC before the integrity protection verification can be performed for the PDU in which the message was received. The parameters that are required by PDCP for integrity protection are defined in reference [11] and are input to the integrity protection algorithm. The PDCP data PDU is used to convey user-plane and control-plane data, as well as MAC-I in addition to the PDU header. The PDCP control PDU is used to transport PDCP status report and/or interspersed ROHC feedback in addition to the PDU header. A PDCP SDUs and PDUs are octet-aligned bit strings. A compressed or uncompressed SDU is included in a PDCP data PDU. Fig. 2.19 shows the format of the PDCP data PDU with 12-bit SN, which is applicable to SRBs. The figure further shows the format of the PDCP data PDU 232 Chapter 2 PDCP SN Octet 1 PDCP SN Octet 1 PDCP SN(cont.) Octet 2 PDCP SN(cont.) Octet 2 Octet 3 Octet 3 MAC-I Octet N-3 MAC-I (optional) Octet N-3 MAC-I (cont.) Octet N-2 MAC-I (cont.) (optional) Octet N-2 MAC-I (cont.) Octet N-1 MAC-I (cont.) (optional) Octet N-1 MAC-I (cont.) Octet N MAC-I (cont.) (optional) Octet N PDCP data PDU format for SRBs PDCP data PDU format with 12 bits PDCP SN PDU type Octet 1 PDU type Octet 1 Octet 2 Interspersed ROHC feedback Octet 2 PDCP control PDU format for interspersed ROHC FMC (cont.) Octet 3 feedback FMC (cont.) Octet 4 FMC (cont.) Octet 5 Bitmap (optional) Octet 6 Bitmap (optional) Octet N+5 PDCP control PDU format for PDCP status report Figure 2.19 PDCP PDU formats [11]. with 12-bit SN for RLC-UM and RLC-AM DRBs. The structure of PDCP control PDU car- rying one PDCP status report, which is applicable to RLC-AM DRBs, as well as the struc- ture of PDCP control PDU transporting one interspersed ROHC feedback, which is applicable to RLC-UM and AM DRBs are shown in Fig. 2.19. In PDCP PDU formats, the sequence number (SN) is a 12- or 18-bit number which is con- figured by RRC. The data field is a variable-size field which includes uncompressed user- plane/control-plane data

or compressed user-plane data. As we stated earlier, the header compression only applies to user-plane data. The MAC-I field carries a message authenti- cation code. For SRBs the MAC-I field is always present; however, if integrity protection is not configured, the MAC-I field is still present in PDCP PDU but is padded with zeros. New Radio Access Layer 2/3 Aspects and System Operation 233 For DRBs, the MAC-I field is present only when the DRB is configured with integrity protection, which is unique to NR. The D/C field indicates whether the corresponding PDCP PDU is a PDCP data PDU or a PDCP control PDU. The PDU Type identifies the type of control information included in the corresponding PDCP control PDU, which can be a status report, interspersed ROHC feedback, or reserved. The first missing COUNT indicates the COUNT value of the first missing PDCP SDU within the reordering window. The Bitmap field indicates which SDUs are missing and which SDUs have been correctly received in the receiving entity. The interspersed ROHC feedback has a variable length and contains one ROHC packet with only feedback, that is, a ROHC packet which is not associated with a PDCP SDU. When an interspersed ROHC feedback is generated by the header compression protocol, the transmitting PDCP sends the corresponding PDCP con- trol PDU to the lower layers without associating a PDCP SN or performing ciphering. The receiving PDCP entity delivers the corresponding interspersed ROHC feedback to the header compression protocol without performing deciphering [11]. 2.2.4 Service Data Adaptation Protocol Sublayer The main services and functions of SDAP sublayer include mapping between a QoS flow and a DRB and marking QFI in downlink and uplink IP packets. A single-protocolocol entity of SDAP is configured for each individual PDU session. The SDAP sublayer was introduced in NR because of the new QoS framework compared to LTE QoS management when con- nected to the 5G core. However, if the gNB is connected to the EPC, which is the case for non-stand

alone deployments, the SDAP service/functionality is not used. As we mentioned earlier, the NG-RAN architecture supports disaggregated gNB where gNB functions are split into a central unit (gNB-CU) and one or more distributed units (gNB-DU) connected via F1 interface. In the case of a split gNB, the RRC, PDCP, and SDAP protocols, described in more detail below, reside in the gNB-CU and the remaining protocol entities (RLC, MAC, and PHY) will be located in the gNB-DU. The interface between the gNB (or the gNB-DU) and the device is denoted as the Uu interface. In the example shown in Fig. 2.4, the SDAP protocol maps the IP packets to different radio bearers, that is, IP packets n and n + 1 are mapped to radio bearer X and IP packet m is mapped to radio bearer y. The SDAP mapping function between a QoS flow and a DRB is due to the new QoS framework which is used in the new radio. The SDAP further marks the QFIs in the downlink due to the use of reflective QoS6 and in the uplink due to the use of new QoS framework. A single SDAP entity (as shown in Fig. 2.20) is configured for each individual PDU session, except for the dual connectivity scenario where two entities can be configured. Reflective QoS flow to DRB mapping is a QoS flow to DRB mapping scheme where a UE monitors the QoS flow to DRB mapping rule in the downlink and applies it to in the uplink [3]. 234 Chapter 2 PDU session PDU session QoS flows QoS flows SDAP-SAP SDAP-SAP SDAP entity SDAP entity SDAP sublayer Radio bearers SDAP-PDU PDCP-SAP PDCP-SDU PDCP-SAP entity entity PDCP sublayer entity entity UE/NG-RAN NG-RAN/UE (transmitting side) (receiving side) QoS flow QoS flow Transmitting Receiving SDAP entity SDAP entity SDAP entity SDAP entity Mapping of QoS flow to a DRB Removing SDAP header SDAP header is SDAP header is configured Reflective QoS flow to not configured DRB mapping Adding SDAP header SDAP header is SDAP header is configured not configured Radio interface (Uu) Figure 2.20 High-level SDAP sublayer functional architecture 3]. Fig. 2.20 illustra

tes one possible structure for the SDAP sublayer; however, the actual implementations may vary. The SDAP sublayer is configured by RRC. It maps the QoS flows to DRBs. One or more QoS flows may be mapped onto one DRB. However, in the uplink, one QoS flow is mapped to only one DRB at a time. The SDAP entities are located in the SDAP sublayer. Several SDAP entities may be defined for a UE. One SDAP entity is New Radio Access Layer 2/3 Aspects and System Operation 235 Octet 1 Octet 1 Octet 2 Octet N SDAP data PDU format without SDAP header Octet N Downlink SDAP data PDU format with SDAP header Octet 1 Octet 1 Octet 2 Uplink SDAP data PDU format with SDAP header Octet N Uplink SDAP data PDU format with SDAP header Figure 2.21 SDAP PDU formats [3]. configured for each individual PDU session. An SDAP entity receives/delivers SDAP SDUs from/to upper layers and transmits/receives SDAP data PDUs to/from its peer SDAP entity via lower layers. In the transmitting side, when an SDAP entity receives an SDAP SDU from upper layers, it constructs the corresponding SDAP data PDU and submits it to lower layers. In the receiving side, when an SDAP entity receives an SDAP data PDU from lower layers, it retrieves the corresponding SDAP SDU and delivers it to upper layers. Fig. 2.20 illustrates the functional block diagram of the SDAP entity for the SDAP sublayer [3]. Note that the reflective QoS flow to DRB mapping is performed at UE, if downlink SDAP header is configured. The SDAP sublayer transfers user-plane data and exclusively provides ser- vices to the user-plane upper layers. As shown in Fig. 2.20, the SDAP sublayer supports transfer of user-plane data, mapping between a QoS flow and a DRB for both downlink and uplink; marking QFI in both downlink and uplink packets; and reflective QoS flow to DRB mapping for the uplink SDAP data PDUs. The SDAP data PDU is used to convey SDAP header and user-plane data. An SDAP PDU or SDU is a bit string that is octet-aligned (Fig. 2.21). An SDAP SDU is included in an SDAP PDU. An SDAP data PDU

may only consist of a data field with no SDAP header. As shown in Fig. 2.21, the SDAP data PDU in the downlink or uplink consists of a header and data fields, where the headers for the downlink and uplink are different. For each downlink SDAP data PDU, in which reflective QoS indication (RQI) is set to one, the SDAP entity would inform the NAS layer of the RQI and QFI. The end-marker control PDU is used by the SDAP entity at UE to inform the gNB that the SDAP SDU QoS flow mapping to the DRB on which the end-marker PDU is transmitted, has stopped. The D/C bit specifies whether the SDAP PDU is an SDAP data PDU or an SDAP control PDU. The QFI field identifies the QFI to which the SDAP PDU belongs. The RQI bit indicates whether NAS should be informed of the updated SDF to QoS flow mapping rules and the reflective QoS flow to DRB mapping indication bit implies whether QoS flow to DRB mapping rule should be updated [3]. 236 Chapter 2 2.3 Layer 3 Functions and Services 2.3.1 Radio Resource Control Sublayer The RRC sublayer consists of control-plane set of protocols for connection control and setup, system configuration, mobility management, and security establishment. It is further responsible for broadcast of SI including NAS common control information and information applicable to UEs in RRC_IDLE and RRC_INACTIVE states, for example, cell selection/ reselection parameters, neighboring cell information, as well as information applicable to UEs in RRC_CONNECTED state, for example, common channel configuration information [14]. The RRC connection control functions include paging; establishment/modification/sus- pension/resumption/release of RRC connections including assignment/modification of UE identity; establishment/modification/suspension/resumption/release of SRBs except SRBO; access barring; initial security activation including initial configuration of AS integrity pro- tection (SRBs, DRBs), and AS ciphering (SRBs, DRBs); mobility management including intra-frequency and inter-frequency handover, associated secur

ity handling such as key or algorithm change, specification of RRC context information transferred between network nodes; establishment/modification/suspension/resumption/release of radio bearers carrying user data (DRBs); radio configuration control including assignment/modification of ARQ configuration, HARQ configuration, and DRX configuration. In case of dual connectivity, the RRC sublayer provides cell management functions includ- ing change of primary second cell (PSCell), addition/modification/release of SCG cell(s). In case of carrier aggregation, RRC sublayer provides cell management functions including addition/modification/release of SCell(s). The RRC sublayer further provides QoS control including assignment/modification of SPS configuration and DL/UL configured grant con- figuration, assignment/modification of parameters for uplink rate-control in the UE, that is, allocation of a priority and a prioritized bit rate for each resource block. The RRC sublayer also handles recovery from RLF condition; inter-RAT mobility including security activa- tion, transfer of RRC context information; measurement configuration and reporting which includes establishment/modification/release of measurement configuration (e.g., intra- frequency, inter-frequency, and inter-RAT measurements); setup and release of measure- ment gaps; and measurement reporting [14]. The RRC messages are sent to the UEs using SRBs, based on the same set of protocol layers that are used for user-plane packet processing except the SDAP sublayer. The SRBs are mapped to the CCCH during connection setup and to the DCCH once the connection is established. The control-plane and user-plane data can be multiplexed in the MAC sublayer and transmitted to the device within the same TTI. The MAC CEs can be used for control of radio resources in some specific cases where low latency is more important than cipher- ing, integrity protection, and reliable transport of data. New Radio Access Layer 2/3 Aspects and System Operation 237 Table 2.1: Radio resource

control (RRC) functions in standalone and non-standalone NR operation. Services Functions Differences With LTE RRC standalone Standalone Architecture Architecture System Broadcast of minimum information system information Broadcast of other system Introduction of on-demand information and area provision Connection Bearer and cell settings Introduction of split SRB and control direct SRB Connection establishment Introduction of with the core network RRC_INACTIVE state Paging Introduction of RAN level paging Access control Introduction of unified access control Mobility Handover Cell selection/reselection Measurement Downlink quality Introduction of beam measurements/reporting measurements Cell identifier measurement/reporting a 'Split SRB is a bearer for duplicating RRC messages generated by the master node for terminals in dual connectivity scenar- ios and transmitting via the secondary node. Direct SRB is a bearer whereby the secondary node can send RRC messages directly to the terminals in dual connectivity scenarios. Table 2.1 shows the functional classification of NR RRC, the relevance of each function to standalone and non-standalone operation, and the similarities and differences with the LTE RRC functions. SRBs are defined as radio bearers that are used only for the transmission of RRC and NAS messages. More specifically, the new radio has specified four types of SRBs which includes SRBO for RRC messages using the CCCH logical channel; SRB1 for RRC messages which may include a piggybacked NAS message, as well as for NAS messages prior to the estab- lishment of SRB2 using DCCH logical channel; SRB2 for NAS messages using DCCH log- ical channel (SRB2 has a lower priority than SRB1 and may be configured by the network after security activation); and SRB3 for specific RRC messages when UE is in EN-DC mode using DCCH logical channel. In the downlink, piggybacking of NAS messages is used only for bearer establishment/modification/release. In the uplink, piggybacking of NAS messages is used only for transferring

the initial NAS messages during connection setup and connection resume. The NAS messages transferred via SRB2 are also contained 238 Chapter 2 in RRC messages, which do not carry any RRC protocol control information. Once security is activated, all RRC messages on SRB1, SRB2, and SRB3, including those containing NAS messages, are integrity protected and ciphered by PDCP sublayer. The NAS indepen- dently applies integrity protection and ciphering to the NAS messages [14]. 2.3.2 System Information The system information consists of an MIB and a number of SIBs, which are divided into minimum SI and other SI. The minimum SI comprises basic information required by the UEs for initial access and acquiring any other SI. The minimum SI itself consists of MIB which contains cell barred status information and essential physical layer information of the cell required for the UEs to receive further SI, for example, CORESET#0 configura- tion. The MIB is periodically broadcast on BCH. The minimum SI further includes SIB1 which defines the scheduling of other SIBs and contains information required for initial access. The SIB1 is also referred to as remaining minimum system information (RMSI) and is periodically broadcast on DL-SCH or sent in a dedicated manner on DL-SCH to UEs in RRC_CONNECTED state. The other SI encompasses all SIBs that are not broadcast as part of minimum SI. Those SIBs can either be periodically broadcast on DL-SCH, broad- cast on-demand on DL-SCH, that is, upon request from the UEs in RRC_IDLE or RRC_INACTIVE state or sent in a dedicated manner on DL-SCH to the UEs in RRC_CONNECTED state. The other SI is divided into the following SIBs [8]: SIB2 contains cell reselection information related to the serving cell. SIB3 contains information about the serving frequency and intra-frequency neighbor cells relevant for cell reselection, including cell reselection parameters common for a frequency as well as cell-specific reselection parameters. SIB4 contains information about other NR frequencies and inter-freque

ncy neighbor cells relevant for cell reselection including cell reselection parameters common for a frequency as well as cell-specific reselection parameters. SIB5 contains information about E-UTRA frequencies and E-UTRA neighbor cells rele- vant for cell reselection, including cell reselection parameters common for a frequency as well as cell-specific reselection parameters. SIB6 contains an Earthquake and Tsunami Warning System (ETWS) primary notification. SIB7 contains an ETWS secondary notification. ETWS is a public warning system developed to satisfy the regulatory requirements for warning notifications related to earthquake and/or tsunami events. The ETWS warning notifications can either be a primary notifi- cation (short notification) or secondary notification (providing detailed information) [8]. New Radio Access Layer 2/3 Aspects and System Operation 239 Minimum system information (MIB) Periodically broadcast on BCH Minimum system information (SIB1) Periodically broadcast on DL-SCH Minimum system information (SIB1) Unicast on DL-SCH Other system information (SIBn) Periodically broadcast on DL-SCH Other system information (SIBn) Broadcast on-demand on DL-SCH Other system information (SIBn) Unicast on DL-SCH Figure 2.22 System information provisioning [8]. SIB8 contains a Commercial Mobile Alert System (CMAS)8 8 warning notification. SIB9 contains information related to Global Positioning System (GPS) time and coordi- nated universal time. Fig. 2.22 summarizes SI provisioning. For a cell/frequency that is considered for camping by the UE, the UE is not required to acquire the contents of the minimum SI of that cell/fre- quency from another cell/frequency layer. This does not preclude the case where the UE applies stored SI from previously visited cell(s). The UE would consider a cell as barred, if it cannot determine the full content of the minimum SI broadcast by that cell. The UE only acquires SI on the active BWP, when using bandwidth adaptation. The MIB is mapped to BCCH and is exclusively carried on B

CH; however, other SI mes- sages are mapped to BCCH and are dynamically carried on DL-SCH. The scheduling of SI messages is part of other SI and is signaled via SIB1. The UEs in RRC_IDLE or RRC_INACTIVE state may request other SI which would trigger a random-access proce- dure, wherein the corresponding Msg3 includes the SI request message unless the requested SI is associated with a subset of physical RACH (PRACH) resources, in that case Msgl is CMAS is a public warning system developed for the delivery of multiple, concurrent warning notifications [8]. 240 Chapter 2 used. When Msgl is used, the minimum granularity of the request is one SI message (i.e., a set of SIBs), one RACH preamble and/or PRACH resource can be used to request multiple SI messages and the gNB acknowledges the request in Msg2. When Msg3 is used, the gNB acknowledges the request in Msg4. The other SI may be broadcast at a configurable period- icity for a certain duration of time. The other SI may also be broadcast when it is requested by a UE in RRC_IDLE or RRC_INACTIVE state [8]. A UE would be allowed to camp on a cell, if it acquires the minimum SI broadcast by that cell. It must be noted that not all cells in a network broadcast the minimum SI; thus the UE would not be able to camp on those cells. The SI may be changed at the specific radio frames according to a modification period. The SI may be transmitted a number of times with the same content within the modification period defined by its scheduling. The modification period is configured by the SI. When the network parameters change (some of the SI), it first notifies the UEs about this change, that is, this may be done within a modification period. In the next modification period, the net- work transmits the updated SI. Upon receiving a change notification the UE acquires the new SI from the beginning of the next modification period. The UE applies the previously acquired SI until the UE acquires the new SI. The short message transmitted with P-RNTI via DCI on PDCCH is used to inform

UEs in RRC_IDLE, RRC_INACTIVE, or RRC_CONNECTED state about an SI change. If the UE receives a short message with SI change indication, it means that the SI will change at the next modification period boundary [8]. As we mentioned earlier, the SI is divided into the MIB and a number of SIBs. The MIB is always transmitted on the BCH with a periodicity of 80 ms and is repeated within the 80 ms. The MIB includes parameters that are needed to acquire SIB1 from the cell. The SIB1 is transmitted on the DL-SCH with a periodicity of 160 ms and variable transmis- sion repetition periodicity within 160 ms. The default repetition period of SIB1 is 20 ms; however, the actual repetition periodicity is up to network implementation. For synchroniza- tion signal/PBCH block (SSB) and CORESET multiplexing pattern 1, SIB1 repetition trans- mission period is 20 ms. For SSB and CORESET multiplexing pattern 2/3, SIB1 repetition period is the same as the SSB period. The SIB1 includes information regarding the avail- ability and scheduling (e.g., mapping of SIBs to SI message, periodicity, and SI-window size) of other SIBs with an indication whether the SIBs are only provided on-demand, and in that case the configuration needed by the UE to perform the SI request. The SIB1 is cell- specific. Other SIBs are carried in SI messages, which are transmitted on the DL-SCH. The SIBs with the same periodicity can only be mapped to the same SI message. Each SI message is transmitted within periodically occurring time domain windows referred to as SI-windows with same length for all SI messages. Each SI message is associated with an SI-window and New Radio Access Layer 2/3 Aspects and System Operation 241 the SI-windows of different SI messages do not overlap. That is, within one SI-window only the corresponding SI message is transmitted. Any SIB except SIB1 can be configured to be cell-specific or area-specific, using an indication in SIB1. The cell-specific SIB is applicable only within the cell that provides the SIB, while the area-specific SIB

is applica- ble within an area referred to as SI area, which consists of one or several cells and is identi- fied by systemInformationAreaID. For a UE in RRC_CONNECTED state, the network can provide SI through dedicated signaling using the RRCReconfiguration message, for exam- ple, if the UE has an active BWP with no common search space configured to monitor SI or paging. For PSCell and SCells, the network provides the required SI by dedicated signal- ing, that is, within an RRCReconfiguration message. Nevertheless, the UE acquires the MIB of the PSCell in order to obtain the SFN timing of the SCG which may be different from that of MCG. Upon change of the relevant SI for SCell, NG-RAN releases and adds the concerned SCell. The physical layer imposes a limit on the maximum size of a SIB. The maximum SIB1 or SI message size is 2976 bits [14]. We will explain in Chapter 3 that the PDCCH (physical downlink control channel) monitor- ing occasions for SI message are determined based on the search space indicated by searchSpaceOtherSystemInformation parameter, if the latter parameter is not set to zero. The PDCCH monitoring occasions for SI message, which are not overlapping with uplink symbols (determined according to tdd-UL-DL-ConfigurationCommon) are sequentially num- bered from one in the SI-window. The PDCCH monitoring occasion(s) [xN + K for SI mes- sage in SI-window correspond to the Kth transmitted synchronization signal block [see Chapter 3 for description], where = 0, X - 1, K = 2, number actual transmitted synchronization signal blocks determined according to ssb- PositionsInBurst in SIB1 and X = Number of PDCCH monitoring occasions in SI - window/N] [14]. 2.3.3 User Equipment States and State Transitions The operation of the RRC sublayer is guided by a state machine which defines the states that a UE may be present in at any time during its operation in the network. Apart from RRC_CONNECTED and RRC_IDLE states, which are similar to those of LTE, the NR has introduced a new RRC state referred to as RRC_INAC

TIVE state. As shown in Fig. 2.23, when a UE is powered up, it is in disconnected and idle mode. However, the UE can transi- tion to the connected mode with initial access and RRC connection establishment. If there is no UE activity for a short period of time, the UE can suspend its active session and tran- sition to the inactive mode. Nevertheless, it can resume its session by moving to the con- nected mode. 5G applications and services have different characteristics. To meet the requirements of different services, it was imperative to reduce the control-plane latency by introducing a new RRC state machine and a dormant state. The URLLC services are Power up Attach/RRC connect RRC suspend RRC_IDLE CONNECTED INACTIVE Detach/RRC release Deregistered, idle RRC resume Registered, connected Connection failure Connection failure Registration reject Registration update accept AN signaling connection established Registration accept (initial NAS message) RM-DEREGISTERED RM-REGISTERED CM-IDLE CM-CONNECTED Deregistration AN signaling registration reject connection released UE status Attach Idle/registered Connected to EPC Active UE status Attach Connected/inactive Connected/active EMM state Deregistered Registered RM state Deregistered Registered ECM state Connected CM state Connected RRC state Connected Connected RRC state Connected Inactive Connected Mobility UE-based UE-based Network-based Mobility UE-based UE-based/NW assisted Network-based Figure 2.23 NR UE states and comparison to LTE 8,14]. New Radio Access Layer 2/3 Aspects and System Operation 243 characterized by transmission of frequent/infrequent small packets that require very low latency and high reliability; thus the devices must stay in a low-activity state, and intermit- tently transmit uplink data and/or status reports with small payloads to the network. There is further a need for periodic/aperiodic downlink small packet transmissions. A UE can move to RRC idle mode from RRC connected or RRC inactive state. A UE needs to register with the network to rece

ive services that requires registration. Once registered, the UE may need to update its registration with the network either periodically, in order to remain reachable (periodic registration update); or upon mobility (mobility regis- tration update); or to update its capabilities or renegotiate protocol parameters (mobility reg- istration update). As we discussed in Chapter 1, the mobility state of a UE in 5G core network (CN) can be either RM-REGISTERED or RM-DEREGISTERED depending on whether the UE is registered with 5GC. The registration management (RM) states are used in the UE and the Access and Mobility Management Function (AMF) to reflect the registra- tion status of the UE in the selected PLMN. In the RM-DEREGISTERED state, the UE is not registered with the network. The UE context in AMF holds no valid location or routing information for the UE; thus the UE is not reachable by the AMF. However, some parts of UE context may still be stored in the UE and the AMF to avoid performing an authentica- tion procedure in each registration procedure. In the RM-REGISTERED state, the UE is registered with the network and can receive services that require registration with the net- work [1]. Fig. 2.23 shows the UE RM states and transition between the two states. Two connection management states are used to reflect the NAS signaling connectivity status of the UE with the CN, namely, CM-IDLE and CM-CONNECTED. A UE in CM-IDLE state has no NAS signaling connection established with the AMF over N1, whereas in CM- CONNECTED state the UE has a NAS signaling connection with the AMF over N1. A NAS signaling connection uses the RRC connection between the UE and the NG-RAN to encapsulate NAS messages exchanged between the UE and the CN in the RRC messages. The NR RRC protocol states consist of three states, where in addition to RRC_IDLE and RRC_CONNECTED states, a third state has been introduced, RRC_INACTIVE, as a pri- mary sleeping state prior to transition to RRC_IDLE state in order to save UE power and to allow fast connecti

on setup [1,8]. In RRC_IDLE state there is no UE context, that is, the parameters necessary for communica- tion between the device and the network, in the radio access network, and the device is not registered to a specific cell. From the CN perspective, the device is in the CM-IDLE state where no data transfer may be performed since the device is in a sleep mode to conserve the battery. The idle mode UEs periodically wake up to receive paging messages, if any, from the network. In this mode, the mobility is managed by the device through cell reselection. The uplink synchronization is not maintained in the idle mode, and thus the UE is required to per- form a random-access procedure in order to transition to the connected mode. As part of tran- sitioning to the RRC_CONNECTED state, the UE context is established in the device and the 244 Chapter 2 network. From the CN perspective the device is in the CM-CONNECTED state and registered with the network. The cell to which the device belongs is known and a temporary identity for the device, that is, C-RNTI is used to identify the UE in NG-RAN while in the connected mode. The connected mode is intended for data transfer to/from the device; however, a DRX cycle can be configured during inactive times to reduce the UE power consumption. Since there is an already-established UE context in the gNB in the connected mode, transition from DRX mode and starting data transfer is relatively fast, requiring no connection setup. In this mode, the mobility is managed by the network. The device provides neighbor cell measure- ments to the network, and the network would instruct the device to perform a handover when necessary. The UE may lose uplink synchronization and may need to perform random-access procedure to be synchronized. The LTE system only supports idle and connected modes. In practice, the idle mode serves as the primary means to reduce the device power consumption during operation in the net- work. However, intermittent transmission of small packets by some delay-sensit

ive applica- tions results in frequent state transitions which would cause signaling overhead and additional delays. Therefore to reduce the signaling overhead and the latency, a third state has defined in NR. As shown in Fig. 2.23, in RRC_INACTIVE state, the UE context is maintained in the device and the gNB and the CN connection is preserved (i.e., the device is in CM-CONNECTED state). As a result, the transitions to or from the RRC_CONNECTED state for data transfer becomes more efficient and faster. In this mode, the UE is configured with sleep periods similar to the idle mode, and mobility is controlled through cell reselection without involvement of the network. The characteristics of NR UE states are summarized in Table 2.2. One important difference between the UE states in NR is the way that the mobility is handled. In the idle and inactive states, the mobility is Table 2.2: Characteristics of RRC states in NR [19]. RRC_IDLE RRC_INACTIVE RRC_CONNECTED UE-controlled mobility based on network configuration Network-controlled mobility (cell reselection) within NR and to/from LTE DRX configured by NAS DRX configured by NAS or DRX configured by gNB Broadcast of system information Neighbor cell measurements Paging (CN-initiated) Paging (CN-initiated or Network can transmit and/or NG-RAN-initiated) receive data to/from UE UE has an CN ID that uniquely NG-RAN knows the RNA NG-RAN knows the cell which the identifies it w/in a tracking area which the UE belongs to UE belongs to No UE context stored in gNB UE and NG-RAN have the UE AS context stored, and the 5GC-NG-RAN connection (both control/user-planes) is established for the UE New Radio Access Layer 2/3 Aspects and System Operation 245 Handover E-UTRA RRC_CONNECTED NR RRC_CONNECTED Resume/release with suspend Connection Connection E-UTRA establish/ establish/ RRC_INACTIVE* RRC_INACTIVE release release reselection Release Release reselection E-UTRA RRC_IDLE NR RRC_IDLE *The E-UTRA RRC_INACTIVE state is only supported when E-UTRA Is connected to 5GC Figure 2.24 UE