LOGICAL LINKS

Field of the invention

This invention relates to managing radio link failure of multiple logical links in a communications system.

Background of the invention

A fifth-generation ("5G") telecommunications standard is being developed.

Any abbreviations used in this document which are not defined in this document should be interpreted in the context of 5G standardisation.

There are currently wireless communication systems in which user equipment (UE) can support multiple logic links wherein each logical link may be optimized for certain reliability and latency requirements. An example of such a communication system is a 5G eV2x (enhanced vehicle to x communication) communication system. 5G eV2x is intended for supporting communications between vehicles (especially road-going vehicles) and other entities. Communications over eV2x may provide non-safety-related services such as dynamic digital map updates and high data rate entertainment, and safety-related services which may assist automated driving, car platooning etc. These types of services have different quality of service requirements. Safety-related services may have to fulfil stringent safety requirements such as providing highly reliable, low latency communications, whereas non- safety-relates services may be more capable of tolerating unreliability or delays. One way to divide these quality requirements at a logical level is to implement network slicing. In network slicing, different logical links are provided on the network side to support different types of service. E2E (end-to-end) slice differentiation may affect any part of a slice, including the RAN.

There is a need for a mechanism to make slices less likely to fail when a UE is supporting multiple slices. It would be preferred that network slicing isolation and/or network slice specific functions could be provided so that a failure on a slice to one UE should not compromise a session or service running on another slice to the same UE.

Summary

According to one aspect there is provided a communication entity capable of supporting a plurality of logical links over a single physical link to a network, the communication entity storing, for each logical link, a respective quality threshold, the communication entity being configured to: periodically measure the quality of the physical link; and transmit to the network a request to modify the configuration of that logical link, if the quality of the physical link violates the threshold for a logical link. According to a second aspect there is provided a method for modifying the configuration of a logical link established in a network to serve a communication entity communicating with the network over a physical link, the method comprising: storing, for each logical link, a respective quality threshold; periodically measuring the quality of the physical link; and transmitting a request to modify the configuration of that logical link, if the quality of the physical link violates the threshold for a logical link.

The request may be a request to modify a subset of the logical links currently serving the UE, which subset includes the said logical link. The subset is such that there is at least one logical link serving the UE that is not included in the subset.

At least one of the quality thresholds may be based on a block error rate. That block error rate may be a block error rate for communications supported for the respective logical link.

The quality threshold for one of the logical links may be different from the quality threshold for another of the logical links. Those quality thresholds may differ in terms of reliability or based on reliability of the respective logical links.

Each logical link may comprise a 5G slice. Each logical link may employ a respective session management function of the network and a respective user plane function of the network.

The communication entity may be capable of supporting the physical link to a single physical access point of the network.

The communication entity may be configured to support the said plurality of logical links simultaneously.

The communication entity may be configured to transmit the said request as a radio resource control message. The said request to modify the configuration of that logical link may comprise a slice and/or service specific radio link failure request for the slice and/or service associated with that logical link. The request may comprise an indication of the identity of the logical link whose configuration is to be modified. The indication of the identity of the logical link whose configuration is to be modified may comprise a slice identifier (ID) which may be the core network part of the slice ID or the RAN part of the slice ID associated with that logical link.

The request may be a request to deactivate the logical link whose configuration is to be modified. The communication entity may be a user equipment entity. The physical link may be a radio link.

According to a third aspect there is provided a network entity for interfacing between (a) a user equipment entity over a physical link between the user equipment entity and a network comprising the network entity and (b) and a plurality of sets of network entities each supporting a respective logical link for the user equipment entity, the network entity being configured to: receive from the user equipment entity a request to modify the configuration of one of the logical links; and in response to that request, communicate with at least one member of the set of network entities supporting that link so as to modify the configuration of that link.

A set of network entities may comprise a respective session management function of the network and a respective user plane function of the network. The network entity may be configured to, in response to the said request, communicate with the respective session management function of the network and the respective user plane function of the network.

The network entity may be an access node of the network.

The network entity may be configured to support the said plurality of logical links between the sets of network entities and the user equipment entity simultaneously.

The network entity may be configured to modify the logical link by deactivating it.

The said storing step may be performed by the communication entity. The method may comprise, in response to the said request, modifying the configuration of the logical link. Modifying the configuration of the logical link may include deactivating the PDU session to the respective session management function and a session modification request with the respective user plane functions.

Brief description of the drawings

Figure 1 shows the existing RAN RLF procedure;

Figure 2 shows the 5G core network session management architecture;

Figure 3 shows the architecture of a UE;

Figures 4 and 5 show signalling for modifying the configuration of slices; Figure 6 shows signalling for modifying the configuration of slices where the UE communicates with the core network via a secondary eNB or a secondary gNB;

Figure 7 illustrates an architecture implementing slice-specific RLF notification; and Figure 8 illustrates a connection management process.

Detailed description

The present invention will now be described by way of example with reference to the accompanying drawings.

The design implementation of the 5G RAN (radio access network) is structured into common UE functionality and slice/service specific functionality. It is defined that only one RRC (radio resource configuration) instance per UE will be used to support the RAN part of slices/services. This single RRC procedure implies that the mobility management procedure of a UE supporting multiple slices will be handled with only one AMF (access and mobility management function) entity. This arrangement allows independence between network slices while a common functionality manages a UE's context information with the mobility management entity. A UE monitors the quality of the links it is supporting. It may do this by measuring the block error rate (BLER) of each link. As is known to those of skill in the art, the block error rate (BLER) of a link is the ratio of the number of erroneous blocks received over the link to the total number of blocks sent over the link. In some applications, an erroneous block may be a block in which the cyclic redundancy check (CRC) of a block is incorrect. The UE may have radio link failure (RLF) thresholds for one or more links, which represent a minimum desired link quality. These may comprise, or may be expressed in terms of, the BLER of a hypothetical PDCCH (physical downlink control channel) transmission from the serving cell. The RLF thresholds may be different for different logical links. That may result in radio link failures occurring at different times. For example: (1 ) for an eMBB (enhanced mobile broadband) link Qout (quality out of synch) may correspond to a 10% BLER while Qin (quality in synch) may correspond to a 2% BLER; (2) for a URLLC (ultra-reliable low latency communications) link the Qin and Qout thresholds can be expected to be lower than for eMBB; and (3) the 5G standardisation document R1 -1714962 proposes that NR (new radio) supports configurable in-sync and out-of-sync thresholds for a UE based on service.

The overall 5G system architecture, service based interface model operates as follows: · UEs register to a single AMF for a set of network slices

• A UE can simultaneously establish sessions with multiple network slices, each one involving different instances of other network functions (NFs), e.g. SMF (session management function)

• session management is handled on a per slice basis by the SMF

· the SMF establishes/modifies/tears down PDU sessions and corresponding user plane entities on a per slice basis

• the SMF can provide different level of service continuity (SSC)

As the RAN part is currently proposed, a radio link failure (resulting in a UE context release message) would impair all the slices that a UE is registered to or in service with, because a single RRC connection/ RRC state machine is defined. When a RLF (radio link failure) happens, UE context release is triggered. That releases the connection to all active network slices to which the UE in question is connected. As a result, the connections to all active slices are disrupted. This process is illustrated in figure 1 . Even though slices having less stringent quality requirements are likely to be more tolerant of breaks in communication due to sessions being modified or re-started, it would be preferred that a failure (e.g. a falling below the minimum quality threshold) on one slice should not compromise sessions/services running on other slices. For example: eMBB and URLLC have different QoS (quality of service) thresholds and reliability requirements as to both control channel and data channel, and it is desirable that the failure of one of them should not cause a failure in the other one. Figure 2 shows the 5G core network session management architecture. A UE 1 is connected by a physical, radio link to an access node (AN or RAN) 2. A single access management function (AMF) 3 serves the UE. To support a single slice to the UE, the AMF communicates with a session management function (SMF) 4 which supports a logical link representing that slice with a respective user plane function (UPF) 5. The UPF communicates onwards with any other entity as required to support the service being provided over the slice. When multiple slices are being supported to a single UE, the AMF serving that UE communicates with multiple SMFs, each of which communicates with a respective UPF. There is one SMF and one UPF per slice. Each SMF or UPF may be a respective physical device, or more conveniently it may be a virtualised entity.

The UE may be any suitable mobile or fixed entity, for example a mobile phone or a data module in a vehicle. The entities other than the UE in figure 2, i.e. on the network side of the physical link 6 are considered to be network entities.

Figure 3 shows schematically the architecture of the UE 1 . The UE comprises an antenna 10. The antenna is coupled to a radio front end 1 1 which handles analogue aspects of transmit and receive signal processing. The radio front end is coupled to a digital signal processor (DSP) 12. The DSP handles digital aspects of transmit and receive signal processing. That includes implementing the protocol steps to be discussed below, for example storing quality thresholds, measuring error rates, comparing those rates with the threshold and determining to transmit the appropriate messages. The DSP is configured to implement the protocol steps to be discussed below. That may be achieved by it storing (in a non-transitory way) code that can be executed by a processor in the DSP and that, when executed, implements the appropriate steps. The DSP also decodes incoming traffic signals and provides them to a consumer 13 for processing or presentation to a user. The consumer could, for example, be a mechanism for displaying data to a user, a mechanism for playing sounds to a user or a mechanism for implementing autonomous driving. The consumer 13 could be external to the UE itself, in which case the UE could terminate in a data connection 14 to which the external consumer could be connected.

The present UE can support multiple slices. As indicated above, each slice is supported over a common physical link between the UE and a single AN, and via that AN a communication link to a single AMF. Each slice has a respective SMF which communicates with that common AMF. The link between the UE and the AMF is termed N1. The link between the AMF and a respective SMF is termed N1 1 . (See figure 2). For each slice the UE stores one or more criteria that define a minimum level of quality for that slice. The minimum level of quality could be defined by, or based on, any suitable metric, for example bit error rate, block error rate, received signal strength, delay or any combination thereof. Specifically, when the minimum level of quality involves multiple criteria it may be satisfied through a defined combination of those criteria on any suitable basis, for example for the link to be of greater than the minimum level of quality it may be that all those criteria are required to be satisfied, or that any of them is satisfied. The criteria may be stored in a non- transient way. They may be stored by the DSP 12. The criteria for a slice of a specific type (e.g. an eMBB slice or a URLLC slice) may be provided in any suitable way. For example, they may be pre-stored in the UE at manufacture, or they may be provided by the network, e.g. when a slice is set up. The minimum quality criteria may correspond to 5G Qout criteria.

The UE monitors the quality being experienced on slices. It may do this by periodically making measurements of the relevant criteria (e.g. by periodically measuring bit error rate, block error rate, received signal strength and/or delay of the slices). The term "periodically making measurements" is used herein to mean the measurements are made by the UE from time to time and includes making the measurements at regularly occurring intervals or at irregularly occurring intervals. The UE compares the current quality as indicated by those measurements with the stored minimum quality criteria for the slices that are being used. In the present UE, if it is found that the current quality does not satisfy the minimum requirement for a particular slice then the UE transmits to the network a message that requests reconfiguration of that particular slice. The message is of a predetermined type to which the network is responsive so as to implement such reconfiguration. The message may include a field that specifies the slice that is to be reconfigured and a reason (e.g. radio link failure (RLF)) for the reconfiguration. It may specify multiple slices that are to be reconfigured. The network responds to that message by reconfiguring the or each slice that is specified in the message. That reconfiguration is preferably done by the access management function of the network. In this way, when there is a failure of a service being provided to a UE by a particular slice, that slice can be reconfigured without the other slices that go to the same UE necessarily being affected. For example, when a URLLC radio link problem happens in one or more slices, the UE concerned signals that to the network. That signalling can go via a gNB (g node B) of the network (or via a secondary gNB and a master gNB). The network can then modify, e.g. by suspending, the session(s) associated with the or each corresponding slice. In that way, the modification can leave other slices serving that UE unaffected. This provides a degree of network isolation for the slices.

Figure 4 shows one signalling scheme to support this. A UE is communicating with a network. The UE's interface to the network is a gNB. The gNB communicates with an AMF serving the UE. The UE is supporting multiple slices simultaneously. The network is providing multiple slices simultaneously to the UE. Each slice is provided by a respective pair of SMF and UPF, as designated in figure 4 by "slice 1 " and slice 2". Each slice can be identified by its S-NSSAI (single network slice selection assistance information).

Suppose during operation there is a radio link problem, as indicated in figure 4. In this example, that problem results in a failure of uRLLC services being provided over slices 1 and 2. The UE detects the failure in that the currently measured quality does not meet the stored quality thresholds for slices 1 and 2. At step 1 of figure 4 the UE signals that failure to the gNB. The UE may do this by means of a message of a format that the network will understand as indicating failure of one or more slices. In the embodiment of figure 4, the UE designates the slice or slices that have failed by indicating the service type(s) of that/those slices. That or those service type(s) can be indicated as a field of the message indicating failure that is sent from the UE to the network. In this example it is the uRLLC service that has failed, so the message indicates the uRLLC service. The gNB receives the failure message of step 1 . It identifies the identities of the slices providing to the UE the service or services indicated as having failed. (Step 2). The gNB signals the identities of those slices (i.e. their S-NSSAIs or RAN part of the slice ID) to the AMF serving the UE. It does that in a message requesting reconfiguration of those slices. (Step 3). The message may indicate a cause of failure, e.g. radio link failure (RLF). The AMF is configured to be responsive to that message to reconfigure the designated slices. It transmits PDU session messages to the SMFs providing the designated slices, and they transmit session modification messages to their respective UPFs. The sessions are modified. Then the session modification messages are acknowledged, and then the PDU session messages are acknowledged. Finally, the AMF signals the gNB to acknowledge the context release command as having been actioned, and the gNB signals the UE to indicate that the data bearers for slices 1 and 2 have been released. (Step 4).

Figure 5 shows a signalling mechanism that is similar to that of figure 4 except in how the UE indicates to the network which slices have failed. In figure 5, the message sent by the UE to the network to indicate the failure of one or more slices includes as a field the identities (e.g. S-NSSAIs) of that or those slices. The methods and techniques described above may also be applied to networks where the UE communicates with the main network via a secondary gNB (SgNB) or secondary eNB (SeNB), and a MeNB (master e node B) or a MgNB (master g node B). Figure 6 shows a signalling mechanism that is similar to that of figure 5 except in that the UE sends the message indicating the failure of one or more slices to a SgNB and the SgNB forwards the message to the core network via an MgNB (or alternatively an MeNB).

The message sent by the UE to indicate failure of one or more slices may be an RRC connection suspend request. It may include as its fields one or more of: (a) the identity of the UE transmitting the message, (b) an indication of the slice or slices that have failed (either through their identities or their service type(s)), (c) an indication of cause of failure. A service type could be indicated by indicating a specific data radio bearer (DRB). That could, for example be eMMB, uRLLC or mIOT. In the arrangement of figure 4, the radio access node (RAN) has, or can acquire from elsewhere in the network, knowledge of the identities of the slices that are providing specific service types to a specific UE. With that information it can determine which S-NSSAI is to be modified in response to a request to modify a particular data radio bearer. A CM state change is not triggered when the message of step 3 is sent. This avoids other slices being affected.

A slice may be modified by being taken down / deprovisioned / deactivated / destroyed. A slice may be modified by alteration of its communication parameters: for example, it may be modified to adopt greater error protection.

Figure 7 shows the architecture of a system that could implement the procedures described above. A UE 20 receives eMBB, uRLLC and mIOT signals from a gNB. The gNB is coupled to a NAS, which in turn is coupled to the AMF. The AMF has N1 1 links to respective SMFs. Each SMF provides a respective slice for one of the services (eMBB, uRLLC, mIOT) being provided to the UE.

If all active slices for a UE fail then the gNB could send a UE-specific context release message to the AMF thereby changing the CM state to idle. That would cause all active slices to that UE to be modified. This process is illustrated in figure 8.