WO2022128300A1

WO2022128300A1 - Method in a telecommunications network, computer program, computer readable carrier medium and device

Info

Publication number: WO2022128300A1
Application number: PCT/EP2021/081988
Authority: WO
Inventors: Stephen Johnson; Francis SCAHILL
Original assignee: British Telecommunications Public Limited Company
Priority date: 2020-12-17
Filing date: 2021-11-17
Publication date: 2022-06-23
Also published as: GB2602076B; GB2602076A; GB202019989D0

Abstract

This invention provides a method in a telecommunications network, and a device for implementing said method, the telecommunications network having a first access network, a second access network, and a User Equipment, UE, wherein the second access network includes an access point, the UE has a UE cellular access connection to the first access network and a UE non-cellular access connection to the access point of the second access network, and the access point has an access point cellular access connection to the first access network, the method comprising the steps of: determining a property of the second access network; modifying a traffic policy for the UE and/or a traffic policy for the access point relating to at least one of the UE cellular access connection and the access point cellular access connection, wherein the modification is to meet a network operator's goal and is based on the determined property of the second access network; and initiating enforcement of the modified traffic policy for the UE and/or the modified traffic policy for the access point.

Description

METHOD IN A TELECOMMUNICATIONS NETWORK, COMPUTER PROGRAM, COMPUTER READABLE CARRIER MEDIUM AND DEVICE

Field of the Invention

The present invention relates to a telecommunications network.

Background

The 3^rd Generation Partnership Project (3GPP) define a set of standards that relate to the 5^th Generation (5G) cellular telecommunications network. 5G cellular networks are designed so that the 5G core network may be accessed by multiple forms of access network (that is, they are “agnostic” to the access network). Accordingly, User Equipment (UE) may connect to the 5G core network via cellular or non-cellular access networks using the same signalling mechanisms defined in the set of 5G standards. Any access network that is used to connect to the 5G core network that is not standardised by the 3GPP is known as a non-3GPP network (a common example being Wireless Local Area Networks, WLANs).

Furthermore, UE may connect to the 5G core network via more than one access network, such as both a cellular and non-cellular access network, in a mode commonly known as “hybrid access”. An example of hybrid access includes a first access connection via a cellular access network and a second access connection via a wired access network. A wired access network may be, for example, a Digital Subscriber Line, DSL, access network, or a Fibre To The Premises, FTTP, access network. In hybrid access mode, the two network accesses may be bonded together so that data traffic for the UE is split and delivered on a packet- by- packet basis across the plurality of access networks. The hybrid access function in the 3GPP 5G standards is realised in the Mutli-Access Packet Data Unit (MA-PDU) Session capability. For any IP flow that is mapped to a MA-PDU Session, the 5G System ensures that an application running on the UE and its associated application server running in the Data Network (DN) see a single IP address for the UE regardless of which access the traffic is delivered over.

A residential gateway is a form of customer premises equipment that typically resides in the user’s home or office and provides an access connection between the UE and the 5G core network. The residential gateway supports both wireless access networks to the 5G core network and wired access networks to the 5G core network. The present inventors have identified at least one problem when a UE uses hybrid access mode in which a first access connection is via a cellular access network and a second access connection is via a residential gateway, in which the residential gateway further utilises an access connection via a cellular access network (that is, either the residential gateway uses only a single access connection that is via a cellular access network, or when the residential gateway is also in hybrid access mode and one of these access connections is via the cellular access network). These problems include:

• the UE receiving an unfair allocation of resources of the cellular access network (that is, by having an unwarranted overconsumption of resources to the detriment of other UEs) by having two access connections via the cellular access network; and

• in a scenario in which both the UE and residential gateway use the hybrid access mode (known as “nested” hybrid access) and traffic is delivered by the cellular access connections of both the UE and residential gateway, application performance may be impacted by excessive jitter and significant differences in latency between each cellular access connection.

The present invention alleviates some or all of the above problems.

Summary of the Invention

According to a first aspect of the invention, there is provided a method in a telecommunications network having a first access network, a second access network, and a User Equipment, UE, wherein the second access network includes an access point, the UE has a UE cellular access connection to the first access network and a UE non-cellular access connection to the access point of the second access network, and the access point has an access point cellular access connection to the first access network, the method comprising the steps of: determining a property of the second access network; modifying a traffic policy for the UE and/or a traffic policy for the access point relating to at least one of the UE cellular access connection and the access point cellular access connection, wherein the modification is to meet a network operator’s goal and is based on the determined property of the second access network; and initiating enforcement of the modified traffic policy for the UE and/or the modified traffic policy for the access point. Modification of the traffic policy for the UE and/or modification of the traffic policy for the access point includes modification of at least one of a group comprising: access type preference, non-seamless offload indication, and an Access Traffic Steering, Switching and Splitting, ATSSS, policy. A traffic policy relating to a non-cellular access connection of the residential gateway may be unchanged.

The property may be of the access point of the second access network and includes one or more of the following: access type, operator identity, and network performance metrics.

The method may further comprise the steps of: determining an identity of the access point; monitoring network performance to determine the network performance metrics for the identified access point.

The method may further comprise the steps of: determining that the UE and access point disconnect; and in response to said determination, reverting from the modified traffic policy for the UE to the traffic policy for the UE and/or reverting from the modified traffic policy for the access point to the traffic policy for the access point.

The network operator’s goal may be one of a network fairness goal and a performance goal. The network fairness goal may be that the UE is allocated an amount of resources proportionate to its associated subscription. In other words, the network fairness goal may be that the UE does not receive an unwarranted overconsumption of resources to the detriment of other UEs.

According to a second aspect of the invention, there is provided a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the first aspect of the invention. The computer program may be stored on a computer readable carrier medium.

According to a third aspect of the invention, there is provided a device for a telecommunications network, the device including a processor adapted to carry out the steps of the first aspect of the invention. Brief Description of the Figures

In order that the present invention may be better understood, embodiments thereof will now be described, by way of example only, with reference to the accompanying drawings in which:

Figure 1 is a schematic diagram of a telecommunications network, illustrating untrusted non-3GPP access;

Figure 2 is a schematic diagram of a telecommunications network, illustrating trusted non-3GPP access;

Figure 3 is a schematic diagram of a telecommunications network having a residential gateway, illustrating hybrid access for the residential gateway;

Figure 4 is a schematic diagram of a telecommunications network implementing Access Traffic Steering, Switching and Splitting (ATSSS);

Figure 5 is a schematic diagram of a telecommunications network, illustrating a User Equipment (UE) in hybrid access mode and a residential gateway in hybrid access mode, wherein the UE uses untrusted non-3GPP access;

Figure 6 is a schematic diagram of a telecommunications network, illustrating a User Equipment (UE) in hybrid access mode and a residential gateway in hybrid access mode, wherein the UE uses trusted non-3GPP access;

Figure 7 is a schematic diagram of a first embodiment of a telecommunications network of the present invention, illustrating a UE in hybrid access mode and a residential gateway in hybrid access mode, wherein the UE uses untrusted non-3GPP access;

Figure 8 is a flow diagram of a first process of a first embodiment of a method of the present invention;

Figure 9 is a flow diagram of a second process of the first embodiment of the method of the present invention; and

Figure 10 is a flow diagram of a third process of the first embodiment of the method of the present invention.

Detailed Description of Embodiments

An overview of the various access connections of 5G systems will now be described. As noted in the Background section, a UE may connect to a 5G core network via a non- 3GPP access network. The 3GPP defines two ways a UE may connect via a non-3GPP access network - untrusted and trusted access. Trusted access is where the access point is known to the 5G core network, and untrusted access is where the access point is not known to the 5G core network.

An example of untrusted access between a UE and a data network via the 5G core network is shown in Figure 1. The UE connects to an access point of the untrusted non- 3GPP network using any suitable mechanism defined for that type of network. Once this connection between the UE and the access point is complete, the UE establishes an IP Security (IPSec) tunnel with an interworking function, Non-3GPP Interworking Function (N3IWF), at the edge of the 5G core network using its SIM credentials to authenticate with the N3IWF. Control and user plane data is delivered to the UE via this IPSec tunnel. The interfaces between the various entities shown in Figure 1 are:

• N1 - the signalling interface between the UE and the Access Management Function, AMF

• N2 - the signalling interface(s) between the AMF and each access network (3GPP Access, and N3IWF)

• N3 - User plane traffic encapsulated in Generic Tunnelling Protocol (GTP) messages between each access network and the User Plane Function (UPF)

• N4 - signalling interface between the Session Management Function (SMF) and the UPF

• N6 - User plane traffic between the UPF and the data network (e g. the internet) - this is usually IP, but could be Ethernet

• N11 - Signalling interface between the AMF and SMF

• NWu - IPSec tunnel that carries N1 signalling and user plane traffic between the UE and N3IWF

• Y1 (a non-3GPP defined interface) - Radio interface between the UE and Access Point e.g. WLAN

• Y2 (a non-3GPP defined interface) - Backhaul interface between the non-3GPP access network and the N3IWF (typically IP i.e. the internet).

In Figure 1 , the dashed line separates the network components based on their ownership - such that all components above the dashed line are owned by the Home Public Land Mobile Network, HPLMN, operator whilst all components below the dashed line are owned by the non-3GPP network operator. Where interfaces pass through nodes without termination in the above diagram, they are being tunnelled through the node by other interfaces. For example, the N1 interface between the UE and the AMF via the untrusted non-3GPP access is tunnelled over NWu between the UE and the N3IWF and N2 between the N3IWF and the AMF.

Figure 2 illustrates an example of trusted access between a UE and data network via the 5G core network. The trusted non-3GPP access point, known as a Trusted Network Access Point (TNAP) is known to the 5G core network and authenticates the UE using its SIM credentials at the Trusted Network Gateway Function (TNGF). Control and user plane data is delivered to the UE through an NWt tunnel (an IPSec tunnel without encryption) between the UE and TNGF. The following interfaces that are not defined above for the untrusted access example and are shown in Figure 2 are:

• NWt - unencrypted IPSec tunnel (that is, an IPSec tunnel using NULL encryption) that carries N1 signalling and user plane traffic between the UE and TNGF

• Ta - interface that carries Extensible Authentication Protocol (EAP) 5G signalling from the TNAP to the AMF for authentication of the UE with the TNAP

• Tn - Interface between TNGFs to support UE roaming between TNAPs and TNGFs. There may be a plurality of TNGFs in a network, each one supporting an area containing multiple TNAPs

• Yt (a non-3GPP defined interface) - Radio interface between the UE and Access Point e.g. WLAN. This must support a high level of security such as Wi-Fi Protected Access version 2 - Enterprise (WPA2-Enterprise).

Figure 3 illustrates a residential gateway connected in hybrid access mode. The residential gateway is connected to the 5G core network via both a wireless connection and a wired connection. In the context of residential gateways, the wired connection to the 5G core network is known as the Wireline 5G Access Network (W-5GAN), in which a Wireline Access Gateway Function (W-AGF) terminates the wired connection and relays 3GPP signalling and data to the residential gateway over the wired connection (via a Y4 interface).

3GPP Release 16 introduces a policy driven framework called Access Traffic Steering, Switching and Splitting (ATSSS). ATSSS defines how traffic should be delivered to a UE or residential gateway, on a per user per application basis, for all different accesses available to that UE or residential gateway. Any traffic flow that is eligible to be delivered using the ATSSS framework will be mapped to a MA-PDU Session. In ATSSS, a new traffic flow for the UE or residential gateway may be steered to use a specific access for a MA-PDU Session. If network conditions change, the traffic flow may be switched to an alternative access. Furthermore, the traffic flow may be split and delivered on a packet by packet basis across the plural accesses (in which scheduling of the traffic flow will be based on measured round-trip times or congestion of each access). The ATSSS policy is implemented by the UE or residential gateway, and the UPF. The UE, residential gateway and UPF therefore contains proxies that can split and reassemble traffic flows depending on the traffic direction (that is, uplink or downlink). Figure 4 illustrates the relevant functions for implementing ATSSS, through which the UE/residential gateway supports by using MA-PDU Sessions. The following additional terminology used in Figure 4 is defined below:

• MPTCP - MultiPath Transmission Control Protocol

• PMF - Performance Measurement Function

• ATSSS-LL - ATSSS Low Layer - a mechanism for delivering Layer 2 traffic such as Ethernet, and Layer 3 traffic such as IPv4 or IPv6

Use of the ATSSS capability is defined by operator policy. The PCF (Policy Control Function) stores a set of ATSSS rules for each UE/residential gateway that control which applications can use the ATSSS capability as well as the specific ATSSS configuration for the application. These rule sets are delivered by the PCF to the UE/residential gateway (over N1) for uplink traffic as well as a matching set of rules to the UPF (over N4) for the downlink traffic.

Figure 5 illustrates a UE that supports MA-PDU Sessions, in which a first access connection is via a cellular access connection (signalling not shown) and the second access connection is via a residential gateway (which also supports MA-PDU Sessions). The boundaries of the residential gateway’s network and of the UE network are illustrated by dotted line enclosures. In the example of Figure 5, the residential gateway is not known by the UE network and the UE therefore utilises untrusted non-3GPP access to the UE network (via an N3IWF). The residential gateway is therefore treated as a separate entity to the UE network, even if both networks are owned by the same operator. The UE network is connected to a data network.

The PCF of the residential gateway’s network defines one or more ATSSS policies for the residential gateway’s network. These policies are delivered to the residential gateway and UPF of the residential gateway’s network. Similarly, the PCF of the UE network defines one or more ATSSS policies for the UE network. These policies are delivered to the UE and UPF of the UE network.

Figure 6 also illustrates a UE that supports MA-PDU Sessions, in which a first access connection is via a cellular access connection (signalling not shown) and the second access connection is via a residential gateway. However, in this scenario, the residential gateway is known by the UE network and the UE can therefore utilise trusted non-3GPP access to the UE network (via a TNGF). Again, the PCF of the residential gateway’s network defines one or more ATSSS policies for the residential gateway’s network. These policies are delivered to the residential gateway and UPF of the residential gateway’s network. Similarly, the PCF of the UE network defines one or more ATSSS policies for the UE network. These policies are delivered to the UE and UPF of the UE network.

A first example policy that may be implemented by the UE and UPF of the UE network (in both the arrangement of Figure 5 and of Figure 6) will now be described. This first example policy is for an HD video streaming service to use the ATSSS priority mode where the preferred access is the fixed access (i.e. the residential gateway’s NWu or NWt connection). Priority mode means that the UE should use its preferred access only, unless it is congested, at which point the secondary access may also be used. This ATSSS policy is triggered in the UE or UPF by a flow’s application ID or 5-tuple (UE IP address + port, App Server IP address + port, protocol type). This is a typical bonded use case.

Furthermore, the residential gateway and UPF in the residential gateway’s network have the same priority mode policy for any traffic which includes the IPSec tunnel of the UE’s untrusted access.

In this first example, a video streaming client on the UE requests content to be delivered at a maximum data rate until its internal buffer is full. It then plays content from its buffer and, when the buffer has emptied by a certain amount, it requests the content to be delivered again at the maximum data rate (to fill its buffer as fast as possible). This burstiness in demand may cause congestion on the residential gateway’s wired access connection, and so the residential gateway will also use its cellular access connection to provide additional capacity. However, the residential gateway’s cellular access connection may also become congested. The UE will then detect its non-3GPP access (the residential gateway NWt or NWu connection) as congested and so use its cellular access connection to provide additional capacity. The UE is therefore effectively using the cellular access twice which may be unfair in areas where there is competition for radio resources.

A second example policy that may be implemented by the UE and UPF of the UE network (in both the arrangement of Figure 5 and of Figure 6) will now be described. This second example is similar to the first example but uses a constant rate video communication (which differs from the first example in that the traffic is substantially constant instead of bursty) and is relatively sensitive to jitter. If the required data rate for the video exceeds the capacity of the UE’s fixed access connection (i.e. both 3GPP and non-3GPP accesses of the residential gateway are fully utilised, possibly due to other traffic on the local network passing through the residential gateway), the UE will also use its own cellular access connection to provide additional capacity. In this scenario, there are 3 different paths delivering packets to the UE and most likely each will have a different latency. The UE will experience these different latencies as jitter and this will be additional to the jitter introduced on each path. There will be additional latency added over the fixed access through the process of splitting and reassembling IP flows in addition to the latency added by using IPSec which is likely to exacerbate the jitter experienced by the UE.

A first embodiment of a telecommunications network 100 of the present invention will now be described with reference to Figure 7. The telecommunications network 100 includes a UE 110, a first access network 120, a second access network 130, and a hybrid access management function 140. This first embodiment relates to the UE 110 in hybrid access mode, in which a first access connection is via a Wireless Local Area Network (WLAN) Access Point (AP) 121 of the first network and the second access connection is via a cellular access node 131 of the second network 130. In this embodiment, the first access network is operated by a first network operator and the second access network is operated by a second, different, network operator. Furthermore, the WLAN AP 121 is not known to the second network operator and the UE 110 therefore utilises untrusted non-3GPP access to the second access network. The WLAN AP 121 is connected to a core network 122 of the first network operator. The UE 110 is able to connect to the second access network via the WLAN AP 121 and core network 122 of the first access network. In this embodiment, the WLAN AP 121 operates according to any one of the Institute of Electrical and Electronics Engineers (IEEE) 802.11 family of standards (e.g. 802.11ac and/or 802.11 ax). The WLAN AP 121 may be a residential gateway, but this is non-essential.

The second access network 130 includes an AMF 132, SMF 133, N3IWF 134, UPF 135 and PCF 136, which generally operate in a manner known in the art and as described above. The UPF 135 of the UE network is connected to a second external data network.

The N3IWF 134 and PCF 136 of the second access network 130 are each connected to the hybrid access management function 140. The hybrid access management function 140 is used to manipulate an ATSSS policy of the UE 110 when it operates in hybrid access mode. In this embodiment, the hybrid access management node 140 includes a configuration function 141 , a cache 143, a metrics function 145, a policy builder 147 and a session manager 149.

The configuration function 141 includes a set of ATSSS policy modification rules indicating how an ATSSS policy should be modified in order to meet the UE network operator’s goals (such as, for example, improving user experience or achieving network fairness). These modification rules can modify the following aspects of the ATSSS policy of the UE:

• Access type preference, such as cellular access only, non-cellular access only, or hybrid access mode;

• Non-seamless offload preference, such as to prefer non-seamless offload to the residential gateway network 120,

• Other ATSSS rules, such as to use a preferred access connection only unless congested (and then use the non-preferred access connection to provide additional capacity) or fails (i.e. active standby), to use the access connection with the lowest latency, and/or to balance load between the access connections by a particular split of traffic across each access connection.

The cache 143 is a memory module for storing ATSSS policies, session identities and device identities. These policies, session identities and device identities are stored with an associated timestamp (representing the time the data was received at the cache 143). The relevance of this data will become clear upon review of the following description of embodiments of a method of the present invention.

The metrics function 145 is configured to store information on each UE and WLAN AP in the network, the information being relevant to the creation and/or modification of an ATSSS policy. This information includes, for example,

• Historical time of day-based hybrid access performance measurements for the first access network, e.g. link speed/throughput, latency jitter, cell id;

• Historical time of day-based hybrid access performance measurements for each UE, e.g. link speed/throughput, latency, jitter; and

• Cell tower/sector live and historical time of day loading.

This information may be collected by the session manager 149 and stored as both realtime information and historical information in the metrics function 149. Information for the UE can be obtained by U F 135 (with the exception of the cell id, which is obtained from the AMF 132). Furthermore, as explained below, information relating to the WLAN AP 121 in the first access network 120 is obtained from the UE 110.

The policy builder 147 is adapted to create a new ATSSS policy (or a new set of ATSSS policies) based on the information stored in the metrics function 145 and information stored in the cache 143. The relevance of the policy builder 147 will also become clear upon review of the following description of the embodiments of the method of the present invention.

The session manager 149 is configured to coordinate the ATSSS policy modification process by cooperation with the configuration function 141 , cache 143, metrics function 145 and policy builder 147. This will also become clear upon review of the following description of the embodiments of the method of the present invention.

A first embodiment of the method of the present invention will now be described with reference to the network of Figure 7 and the flow diagrams of Figures 8 to 10. A first process of this first embodiment relates to the UE 110 associating with the WLAN AP 121 and is illustrated by the flow diagram of Figure 8. In a first step S1101, the UE 110 sends a Generic Advertisement Service (GAS) request to the WLAN AP 121. GAS is defined in the IEEE 802.11u-2011 standard and improves interworking of WLAN networks operating according to the 802.11 standards. This GAS request includes an Access Network Query Protocol (ANQP) query to the WLAN AP 121 to discover the following data:

• Hybrid Access Indicator (HAI), indicating that the WLAN AP 121 uses one of fixed, fixed wireless access or hybrid access;

• Mobile Country Code (MCC) and Mobile Network Code (MNC) - together uniquely identifying the cellular network operator of the cellular connection of either the fixed wireless access or hybrid access;

• Wide Area Network (WAN) metrics, such as uplink/downlink speeds and uplink/downlink loads (as a percentage of capacity).

In response, the WLAN AP 121 sends the requested data to the UE 110.

In step S1102, the UE 110 associates with the WLAN AP 121 and obtains an IP address on the first access network 120.

Subsequently, in step S1103, the UE 110 initiates a request to the N3IWF 134 (which may be via a data network associated with the first access network) to establish a connection to the second access network 130. This request includes the data requested in step S1101 , such as one or more of the HAI, MCC-MNC, and WAN metrics. The request also includes the WLAN AP’s Basic Service Set I Dentifier (BSSID), which is a globally unique Media Access Control (MAC) address of the WLAN AP 121 . This data may be included in the request as part of an Extensible Authentication Messaging (EAP)- 5G / Extended Type signalling message. Following acceptance of this request, an NWu connection (as noted above, an IPSec tunnel) is set up between the UE 110 and the N3IWF 134 and the UE 110 is connected to the second access network 130.

On successful setup of the NWu connection between the UE 110 and the N3IWF 134, in step S1105, the N3IWF 134 sends an update message to the HAMF 140 to inform the HAMF 40 of the new NWu connection between the UE 110 and N3IWF 134. This update message includes the UE’s identity, such as an IMSI, TIMSI or GUTI, and the data received in step S1103 (such as one or more of the HAI, MCC-MNC, WAN metrics, and BSSID). The HAMF 140 receives this update message at the session manager 149. In response, in step S1107, the session manager 149 retrieves an ATSSS policy (or policies) for the UE 110. These may be retrieved from cache 143 based on a stored mapping between the UE’s identity and the ATSSS policy or policies. If there is no stored mapping in cache 143, or again if the timestamp associated with the stored mapping indicates that the data is too old, then the session manager 149 sends a request message to the PCF 136 of the second access network 130 for the UE’s ATSSS policy or policies, the request message including the UE’s identity. The response message from the PCF 136 includes the ATSSS policy or policies. The session manager 149 may then store the ATSSS policy or policies in cache 143 with a new timestamp and a flag indicating that they are the original ATSSS policies.

In step S1109, the session manager 149 stores the WAN metrics in cache 149, together with a mapping to the BSSID. In an alternative example in which the data received in step S1103 does not include WAN metric data, the session manager 149 may retrieve the last known WAN metric data for the WLAN AP 121 based on the BSSID mapping.

In step S1111 , the policy builder 147 reviews each UE ATSSS policy and determines whether a modification should be made. These determinations are based on the ATSSS policy re-write rules stored in the configuration function 141 and may also be based on any WAN metric data (as received in step S1103 or retrieved in step S1109). If so, then the corresponding modification prepares a new UE ATSSS policy, in which the new ATSSS policy is a modification of the input UE ATSSS policy. In this example, these policy re-write rules include:

• If the first network operator (for the first access network) and the second network operator (for the second access network) use the same cellular network, then the UE 110 is prevented from using cellular network (e.g. via fixed wireless access or via hybrid access) so that only a single cellular access connection is used between the UE 110 and WLAN AP 121 ;

• If the UE 110 and WLAN AP 121 are both able to use hybrid access (a “nested” hybrid access scenario), then the ATSSS policy may be modified to address any latency/jitter problems that may occur. For example, an application of the UE (having its own unique ATSSS policy) may be for the UE to use non-cellular access only so as to meet certain latency/jitter requirements. However, if this would be split between cellular and non-cellular access at the WLAN AP 121 and the performance metrics for one or both of these connections are below the application requirements, the ATSSS policy may be modified so that the UE uses its cellular access connection instead. Furthermore, the UE may run a plurality of applications (each having their own ATSSS policy), which may have a unique solution so as to select one or more access connections (that is, the UE’s cellular and non-cellular access connection and/or the WLAN AP’s cellular and non- cellular access connection) based on the application’s requirements, the network operator’s goals, and the performance metrics of those access connections.

Once complete, the policy builder 147 sends these one or more new UE ATSSS policies to the session manager 149, in which each new UE ATSSS policy either matches the original UE ATSSS policy or is a modification of the original UE ATSSS policy.

In step S1115, the session manager 149 stores the new ATSSS policy or policies in cache 143 together with a stored mapping to the BSSID, the UE’s identity, and a timestamp. This is marked as an active session entry. The original ATSSS policy or policies remain stored in the cache 143.

In step S1117, the session manager 149 sends an update message to the PCF 136 of the second access network 130, the update message including the new ATSSS policy or policies. The second access network 130 then pushes these new ATSSS policy or policies to the UE 110 and UPF 135. This may result in session modification procedures being initiated by the SMF 133 in the second access network 130. Once the PCF 136, UE 110 and UPF 135 have been updated according to the new ATSSS policy or policies, then any future traffic for the UE 110 is routed according to this new ATSSS policy or policies.

A second process of this first embodiment will now be described with reference to Figure 9. This second process is performed periodically whilst there is an active connection between the UE 110 and the WLAN AP 121. In a first step S1201 , the session manager 149 sends a request message to the UPF 135 of the second access network 130. This request message requests data on the performance of the connection for UE traffic over the first access network 120. These statistics may include, for example, uplink average throughput, downlink average throughput, maximum speed, latency and/or jitter. In step S1203, on receipt of the requested performance metric data from the UPF 135, the session manager 149 sends the performance metric data to the metric function 145 together with the BSSID. In step S1205, the metric function 145 stores the performance metric data together with a mapping to the BSSID. In step S1207, the metric function 145 updates its statistics model for the WLAN AP 121 . As noted above, this performance metric data and/or statistical model for the WLAN AP 121 may be used by the policy builder 147 in modifying a UE ATSSS policy.

A third process of this first embodiment will now be described with reference to Figure 10. This third process relates to the UE 110 disassociating with the WLAN AP 121. In a first step S1301, the UE 110 disassociates from the WLAN-AP 121. In step S303, the UE network 30 detects the disassociation, for example, by a) a “dead peer detection” mechanism in Internet Key Exchange (IKE) version 2 (as defined in Internet Engineering Task Force (IETF) Request For Comments (RFC) 7296), or b) an ATSSS Performance Management Function (PMF) in the UE 110 signalling to UPF 135 that an access has stopped. Following detection, in step S1305, the N3IWF 134 (in the case of the dead peer detection) or UPF 135 (in the case of the PMF notification) sends an update message to the session manager 149 of the disassociation, wherein the update message includes the UE identifier.

On receipt of this update message, in step S1307, the session manager 149 inspects cache 143 to determine if there’s a stored active session entry for the UE (based on the UE’s identity). If not, then the third process ends. If this determination is positive, in step S1309, the session manager 149 sends a request message to UPF 135 of the second access network 130. This request message requests data on the performance of the connection for UE traffic over the first access network 120. These statistics may include, for example, uplink average throughput, downlink average throughput, maximum speed, latency and/or jitter. In step S1311 , on receipt of the requested performance metric data from the UPF 135, the session manager 149 sends the performance metric data to the metric function 145 together with the BSSID. In step S1313, the metric function 145 stores the performance metric data together with a mapping to the BSSID, and updates its statistics model for the WLAN AP 121. The purpose of updating the statistics model at this time is such that it is based on the last known performance data (and also in case the session duration ended before the second process was implemented). In step S1315, the session manager 149 deletes the active session entry from cache 143. In step S1317, the session manager 149 retrieves the original (flagged) ATSSS policy or policies from cache 143. In step S1319, the session manager 149 sends an update message to the PCF 136 of the second access network 130, the update message including the original ATSSS policy or policies. The second access network 130 pushes the original ATSSS policy or policies to the UE 110 and UPF 135. This may result in session modification procedures being initiated by the SMF 133 in the first access network 130. Once the PCF 136, UE 110 and UPF 135 have been updated according to the original ATSSS policy or policies, then any future traffic for the UE 110 is routed according to this original ATSSS policy/policies.

This embodiment of the present invention provides a mechanism for managing how application traffic is delivered to the UE when the UE is connected to a WLAN AP and both the UE and WLAN AP utilise respective cellular access connections. This involves a process for detecting that this scenario has occurred and, in response, retrieving and modifying the relevant traffic policies to avoid undesirable behaviour. The mechanism also involves a process for detecting when the UE has disconnected from the WLAN AP and, in response, restoring the original traffic policy. Furthermore, this solution is relevant where the WLAN AP is used in either trusted or untrusted mode, and where the first and second networks are either owned by the same operator or different operators. That is, this solution does not require any interworking between the first and second networks and so provides a general mechanism for adapting a traffic policy of the UE regardless of its connection.

In the embodiments above, the HAMF manipulates an ATSSS policy of the UE. However, the skilled person will understand that this is non-essential and that an ATSSS policy of the WLAN AP may be manipulated instead (for example, if the WLAN AP is a residential gateway and both the first and second access networks are owned by the same network operator). In this case, the most likely modification is to ensure that traffic for the UE is not split in the first access network. Furthermore, it is non-essential that the traffic policy being modified by the HAMF is an ATSSS policy. Additionally or alternatively, the traffic policy may be a UE Route Selection Policy (USRP). The USRP may be stored in the PCF and maps application traffic to one of: • Single Access PDU Session, where the traffic is only ever delivered over one access connection. This rule will specify whether 3GPP or non-3GPP is preferred. If the access changes, the UE IP address may also change (depending on the Service and Session Continuity (SSC) mode); • Multi Access PDU Session, where traffic is controlled by ATSSS policy. IP address always the same regardless of access; and

• Non seamless offload: traffic is routed over the non-3GPP access connection directly to the data network and thus not passing through the 5G core network. The skilled person will understand that any combination of features is possible within the scope of the invention, as claimed.

Claims

1. A method in a telecommunications network having a first access network, a second access network, and a User Equipment, UE, wherein the second access network includes an access point, the UE has a UE cellular access connection to the first access network and a UE non-cellular access connection to the access point of the second access network, and the access point has an access point cellular access connection to the first access network, the method comprising the steps of: determining a property of the second access network; modifying a traffic policy for the UE and/or a traffic policy for the access point relating to at least one of the UE cellular access connection and the access point cellular access connection, wherein the modification is to meet a network operator’s goal and is based on the determined property of the second access network; and initiating enforcement of the modified traffic policy for the UE and/or the modified traffic policy for the access point.

2. A method as claimed in Claim 1 , wherein modification of the traffic policy for the UE and/or modification of the traffic policy for the access point includes modification of at least one of a group comprising: access type preference, non-seamless offload indication, and an Access Traffic Steering, Switching and Splitting, ATSSS, policy.

3. A method as claimed in either Claim 1 or Claim 2, wherein the property is of the access point of the second access network and includes one or more of the following: access type, operator identity, and network performance metrics.

4. A method as claimed in Claim 3, further comprising the steps of: determining an identity of the access point; monitoring network performance to determine the network performance metrics for the identified access point.

5. A method as claimed in any one of the preceding claims, further comprising the steps of: determining that the UE and access point disconnect; and in response to said determination, reverting from the modified traffic policy for the UE to the traffic policy for the UE and/or reverting from the modified traffic policy for the access point to the traffic policy for the access point.

6. A method as claimed in any one of the preceding claims, wherein the network operator’s goal is one of a network fairness goal and a performance goal.

7. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of any one of Claims 1 to 6.

8. A computer readable carrier medium comprising the computer program of Claim 7.

9. A device for a telecommunications network, the device including a processor adapted to carry out the steps of any one of Claims 1 to 6.