WO2005002267A2

WO2005002267A2 - Seamless inter-system handover using pre-authentication and session pre-activation

Info

Publication number: WO2005002267A2
Application number: PCT/IB2004/002122
Authority: WO
Inventors: Govindarajan Krishnamurthi; Hemant M. Chaskar; Dirk Trossen; Leopoldo Alarcon; Sandro Grech; Pedro Serna
Original assignee: Nokia Corporation
Priority date: 2003-06-30
Filing date: 2004-06-24
Publication date: 2005-01-06
Also published as: GB0315278D0; WO2005002267A3; EP1639853A2; US20050130659A1

Abstract

The present invention relates to a method for ensuring continuity of a communication session when a user equipment (100) hands over from a first communication network (A) to a second cellular communication network (B) comprising the steps of performing an authentication procedure for a packet data session with the second network whilst still being attached to the first network and simultaneously performing a packet data session establishment procedure with the second network whilst still being attached to the first network.

Description

A METHOD FOR OPTIMISING HANDOVER BETWEEN COMMUNICATION NETWORKS

Field of the Invention

The present invention is concerned with the optimisation of the handover process when a user equipment (UE), for example, a mobile node (MN), requires a seamless transfer during movement between, for example, the coverage area of a wireless local area network (WLAN) and the coverage area of a cellular communication network.

Background of the Invention

Communication systems providing users thereof with a possibility for wireless communication are known. A typical example of such a system is a cellular or mobile communications system. The cellular communication system is a communication system that is based on use of radio access entities and/or wireless service areas. The access entities are often referred to as cells. A characteristic feature of the cellular systems is that they provide mobility for the users of the communication system. Hence, they are often referred to as mobile communication systems. Another type of wireless communication system can be provided by way of a wireless local area network (WLAN). A WLAN is typically provided to allow access over a limited area such as within or in the close vicinity of a building. A WLAN network provides a low cost and high speed wireless access solution for localised "hotspots" e.g. a WLAN where only employees of the company are authorised to access the network without being charged a fee or a bookstore WLAN where customers would be charged a reader fee to access the network. In contrast, cellular access in a 3G network area would typically always be charged to a user's account with the cellular operator.

Non-limiting examples of cellular communications systems include standards such as the GSM (Global System for Mobile communications) or various GSM based systems (such as GPRS General Packet Radio Service), AMPS (American Mobile Phone System), DAMPS (Digital AMPS), WCDMA (Wideband Code Division Multiple Access), TDMA/CDMA (Time Division Multiple Access/Code Division Multiple Access) in UMTS (Universal Mobile Telecommunications System), IMT 2000, i- Phone and so on.

In a cellular system, a base transceiver station provides a wireless communication facility that serves mobile stations (MS) or similar wireless user equipment (UE) via an air or radio interface within the coverage area of the cell. As the approximate size and the shape of the cell is known, it is possible to associate the cell to a geographical area. The size and shape of the cells may vary from cell to cell. Several cells may also be grouped together to form a larger service area.

Each of the cells can be controlled by an appropriate controller apparatus. For example, in the WCDMA radio access network the base station (which may be referred to as a Node B) is connected to and controlled by the radio network controller (RNC). In the GSM radio network the base station may be connected to and controlled by a base station controller (BSC) of a base station subsystem (BSS). The BSC/RNC may be then connected to and controlled by a mobile switching center (MSC). Other controller nodes may also be provided, such as a serving GPRS support node (SGSN). The MSCs of a cellular network are typically interconnected and there may be one or more gateway nodes connecting the cellular network e.g. to a public switched telephone network (PSTN) and other telecommunication networks such as to the Internet and/or other packet switched networks.

Various types of user equipment (UE) such as computers (fixed or portable), mobile telephones, personal data assistants or organisers and so on are known to the skilled person and can be used to access the Internet to obtain services via a mobile communication system. Mobile user equipment is often referred to as a mobile station (MS) and can be defined as a means that is capable of communication via a wireless interface with another device such as a base station of a mobile telecommunication network or any other station. Each mobile user equipment can typically be identified based on an unique identifier, for example, based on the International Mobile Subscriber Identity (IMSI).

The 3G Partnership Project (3 GPP) defined a reference architecture for the Universal Mobile Telecommunication System (UMTS) core network which provides the users of user equipment UE with access to a wide range of services such as Internet Protocol Multimedia IM Services, conferencing, telephony, gaming, rich call, presence, e- commerce and messaging. The UMTS core network is divided into three principal domains. These are the Circuit Switched (CS) domain, the Packet Switched (PS) domain and the Internet Protocol Multimedia (IM) domain.

The core network may be based on the user of the general packet radio service (GPRS). The GPRS operation environment comprises one or more subnetwork service areas, which are interconnected by a GPRS backbone network. A subnetwork comprises a number of packet data service nodes (SN), which in this application will be referred to as serving GPRS support nodes (SGSN), each of which is connected to the mobile communication access network (typically to base station systems by way of radio network controllers (RNC)) in such a way that it can provide a packet service for mobile user equipment via several base stations, i.e. cells. The intermediate mobile communication access network provides packet-switched data transmission between a support node and mobile data terminals. Different subnetworks are in turn connected to an external data network, e.g. to a packet switched public data network (PSPDN), via GPRS gateway support nodes (GGSN). An example of an external data network is an Internet Protocol (IP) network. The GPRS service thus allows packet data transmission between mobile user equipment and external data networks when the cellular network functions as an access network.

In a GPRS network the mobile user equipment may send a message requesting to activate a packet data protocol (PDP) context in the network. A serving GPRS support node (SGSN) authenticates the mobile user and sends a PDP context creation request to a GGSN selected according to a GGSN address stored in the subscriber data or according to the access point name given by the user equipment, or to a default GGSN known by the SGSN.

In such a network, a packet data protocol (PDP) context is established to carry traffic flows over the network, each PDP context including a radio bearer provided between the user equipment and the radio network controller, a radio access bearer provided between the user equipment, the radio network controller and the SGSN, and switched packet data channels provided between the serving GPRS service node (SGSN) and the gateway GPRS service node (GGSN). Each PDP context can carry more than one traffic flow, but all traffic flows within one particular PDP context are treated the same way as regards their transmission across the network. The PDP context treatment requirement is based on PDP context treatment attributes associated with the traffic flows, for example, quality of service and/or charging attributes. 3G technology encompasses both WCDMA (Wideband Code Division Multiple Access) and cdma2000 (Code Division Multiple Access 2000) air interfaces. 2.5G technology may employ GPRS (General Packet Radio System). At present, both 3 G and 2.5G technologies are proliferating and are likely to be required for some time. A complimentary technology has also been introduced which is known as IEEE 802.1 lb (Wi-Fi or wireless fidelity) and is used in a WLAN (Wireless Local Area Network).

Whilst UMTS networks, in particular 3G networks, are designed to support moderate bandwidth requirements under high mobility conditions, i.e. a wide coverage area, in contrast, a WLAN network is applicable to high bandwidth low mobility scenarios, i.e. a localised coverage area. With an increase in mobile terminals having mobile access interfaces, i.e. a combination of cellular and WLAN radio interfaces, end users would naturally want to be able to seamlessly transfer an ongoing Internet session between a WLAN and a UMTS network as they move between the coverage areas of these networks. The present invention is, therefore, concerned with the optimisation of the handover process in such a situation.

During a handover at IP (Internet Protocol) level between a WLAN network and a UMTS/GPRS network, the mobile terminal or MN (Mobile Node) must first achieve link layer (L2) connectivity with the UMTS RAN (Radio Access Network). In order to achieve that, the MN gets synchronisation with the RAN and establishes a L2 connection. After synchronisation, the authentication procedure is started and the MN and the UTMS network are authenticated by each other. If the procedure is successful, the MN is authorised to access the UMTS network. As a final step, the MN gets IP connectivity by performing the PDP (Packet Data Protocol) Context Activation procedure. As a result, the MN gets an IP address and also the UTMS network is configured with the negotiated Qos (Quality of Service) parameters for that IP session.

One prior art solution addresses the handover between a WLAN and a cdma2000 network and is concerned with minimising the time involved in "establishing" IP bearers in the cdma2000 network. However, there is no attempt to solve the particular problem of how network layer (L3) IP bearers are established in conjunction with link layer (L2) authentication. This prior art solution describes only how the network perfonns L2 authentication and PDP context establishment once the MN has moved into the UMTS (3G) domain. The resulting delay in handover time means that a security association has to exist between the two networks.

In a typical scenario, a MN initiates an IP session whilst roaming from a WLAN network into 3G coverage. If the MN has to perform all the protocols described earlier, the time involved will cause a disruption in the IP session. Furthermore, in certain situations, depending on the local environment, the region of overlap between the signals from WLAN and UMTS networks may not be very large. For example, when moving in and out of tunnels and when there is disruption due to certain types of building construction. In such a scenario, it has been found that when the MN moves from a WLAN network to a cellular network, the WLAN signal may fade very fast and, as a result, the time frame for carrying out handover is very small. Therefore, in such a situation, a MN must minimise the latency of IP level handovers between WLAN and UMTS networks to avoid the chance of a non-seamless handover arising. A seamless handover arises when the handover time is reduced (i.e., lack of IP connectivity is reduced) and when there is a very small, if any, loss of IP packet. The present invention aims to reduce the time for IP level handover by preparing the UTMS network for arrival of the MN both at the link layer (L2) and the IP network layer (L3) before the MN arrives at the UMTS network.

Summary of the Invention

It is an aim of the preferred embodiments of the present invention to address the problems herein described.

According to the present invention, there is provided a method for ensuring continuity of a communication session when a user equipment hands over from a first communication network to a second cellular communication network comprising the steps of performing an authentication procedure for a packet data session with the second network whilst still being attached to the first network and simultaneously performing a packet data session establishment procedure with the second network whilst still being attached to the first network.

Preferably, the authentication procedure includes authentication of the second network by the user equipment.

Preferably, the authentication procedure also includes authentication of the user equipment by the second network.

Preferably, the first communication network is a WLAN network and the second communication network is a cellular network. Preferably, the information sent by the user equipment for authentication and packet data session establishment travels either as a separate IP package or is piggybacked with existing signalling.

Preferably, the gateway node between the first and second communication networks is able to act as an access router for the first network and is able to host the packet data session in the second network.

Preferably, there is provided a method for ensuring continuity of a communication session when a user equipment hands over from a first communication network to a second cellular communication network wherein attachment of the user equipment to the second network is maintained after the user equipment moves away from the coverage area of the second network for a predetermined time in order to allow the user equipment to return to the second network without having to repeat an authentication procedure and a packet data session establishment procedure before handing over to the second network.

Preferably, there is further provided the step of releasing the packet data session if the user equipment does not handover to the second network within a predetermined time thus requiring the user equipment to repeat the authentication procedure if moving towards the second network for a further time.

Preferably, there is further provided a method comprising the following steps: (i) the user equipment sends a handover trigger indication to a gateway node in the second network, the handover trigger indication comprising the user equipment identification parameters and the packet data protocol profile (ii) the gateway node sends the user equipment identification parameters and the packet data protocol profile to the serving node in the second network; (iii) the serving node contacts the home location register to obtain the user equipment authentication parameters; (iv) the serving node sends a packet data protocol profile request to the gateway node; (v) the gateway node responds by sending a packet data protocol profile response to the serving node; (vi) the serving node sends authentication information to the gateway node; (vii) the gateway node sends the authentication information to the user equipment; (viii) the user equipment authenticates the second network; (ix) the user equipment sends a response to the serving node and moves into the second network.

Preferably, there is further provided a method comprising the following steps:

(i) the user equipment sends a handover trigger indication to a gateway node in the second network; (ii) the gateway node sends a protocol data unit notification request to the serving node in the second network; (iii) the serving node contacts the home location register to obtain the user equipment authentication parameters; (iv) the serving node sends a proxy authentication and ciphering request to the gateway node; (v) the gateway node converts the authentication information in the request which is then sent to the user equipment; (vi) the user equipment responds with an authentication message which is sent to the gateway node; (vii) the gateway node converts the authentication message from the user equipment and sends a proxy authentication and ciphering response to the serving node; (viii) the serving node sends a protocol data unit notification response to the gateway node; (ix) the serving node sends a create packet data protocol request to the gateway node; (x) the gateway node sends a create packet data protocol response to the serving node; and (xi) the gateway node replies to the handover trigger indication sent by the user equipment in step (i) by sending a handover trigger response to the user equipment.

According to the present invention, there is also provided a communication system comprising a user equipment, a first communication network and a second cellular communication network, the system being arranged to enable continuity of a communication session when the user equipment moves from the coverage area of the first network to the coverage area of the second network, wherein means are provided to simultaneously perform an authentication procedure for a packet data session with the second network and perform a packet data session establishment procedure with the second network whilst the user equipment is still attached to the first network.

Brief Description of the Drawings Figure 1 is a simplified presentation of a mobile communication system according to a first preferred embodiment of the present invention;

Figure 2 depicts the signal flow in the first preferred embodiment of the present invention;

Figure 3 is a simplified presentation of a mobile communication system according to a second preferred embodiment of the present invention; and,

Figure 4 depicts the signal flow in the second preferred embodiment of the present invention.

As described earlier, during handover the MN has to perform a number of actions each of which contribute to the total handover time. Some of the actions are, for example, MN authentication in the UMTS/GPRS network, obtaining a new IP address in the UMTS/GPRS network and even specific L2 procedures depending on the access technology the MN will use in the UMTS/GPRS network. Clearly, all of these actions take time which would result in a non-seamless transfer if performed on entry into the second network.

With the aim of performing a seamless transfer at least some of the actions will be performed whilst the MN is in i.e. attached to the WLAN network. Preferably, all of the actions will have been completed before the movement from the WLAN to the UMTS/GPRS network takes place.

The main contributions to the handover time when moving from the WLAN to the UMTS/GPRS are currently as follows: 1. Authentication of the MN in the target network and also authentication of the target network by the MN. Link layer authentication is required because the target network has to establish whether the MN is allowed to access that network or not;

2. Activation of PDP contexts. If the target UMTS network is GPRS, the activation of PDP contexts is carried out during handover. The PDP contexts are logical connections needed inside the GPRS network for the transmission of PDUs (Packet Data Units) of upper layers (layers placed above the link layer e.g. IP) in this case IP packets between the MN and the GGSN (Gateway GPRS Support Node). The GGSN acts as an AR (Access Router) in the GPRS network from the point of view of the MN.

Reference will now be made to Figure 1 which is a simplified presentation of a first preferred embodiment of the present invention for handover between a WLAN network A and a GPRS network B.

The mobile node (MN) 100 is engaged in an IP communication session between the WLAN network A and the IP network C. The IP communication session is provided by, for example, a service provider 111. The MN 100 wirelessly receives and transmits signals from and to base station 102. There is an access router (AR) 103 for routing the signals from the base station 102 to the IP network C. If the MN now moves towards the GPRS network B and the IP communication session is to continue, the present invention proposes that handover is accomplished whilst the MN 100 is still attached to the WLAN network A. Although Figure 1 depicts the WLAN network A as completely within the GPRS network B, there could simply be an overlap between the two coverage areas. In the GPRS network B there is a gateway GPRS support node (GGSN) 104, a serving GPRS support node (SGSN) 105, the home location register (HLR) 106 and a second GGSN 108 through which the IP communication session continues with the IP network C. The SGSN 105 is connected to a radio network controller (RNC) 109 in the GPRS network B and the RNC 109 is connected to a base station (Node B) 110. Once authentication and PDP context establishment is completed, the signalling will pass from the MN 100 to and from base station 110 within the RAN of GPRS network B as the IP communication session continues with IP network C via SGSN 105 and through GGSN 104.

In order to access the PS (Packet Switched) service in a UMTS/GPRS network, the MN must first make its presence known to the network by performing UMTS/GPRS attach. Reference should now be made to Figure 2 for the signal flow in the first preferred embodiment.

In the attach request, the SGSN 105 needs the MN's identity (IMSI - International Mobile Subscriber Identity) and an indication of which type of attach is to be executed. The SGSN 105 will then forward this information to the HLR 106 of the MN to authenticate the MN. Once authenticated at the link layer, the MN then proceeds to establish its IP bearers, also known as PDP contexts, at the GGSN 108. This process includes obtaining temporary IP addresses and establishing the QoS profile needed for its packet sessions. The GGSN 108 is chosen based on the PDP profile that the MN schedules along with the attach message.

In the present invention, the information needed to authenticate the MN at the link layer and establish the PDP contexts is sent to a GGSN 104 of the target UMTS network from the MN via the access router AR 103 of the WLAN network whilst the MN is still connected to the AR 103. The AR 103 is located between the MN 100 and the GGSN 104 in the WLAN network and simply forwards the messages between the MN and the GGSN. It is important to note that this can be implemented even when the degree of overlap between the GPRS and WLAN coverage areas is negligible, albeit with less efficiency. This is possible with help from the current AR 103 and to enable this support the AR 103 can use protocols such as CAR (Candidate Access Router) discovery. The MN is able to send the information required for link level authentication and PDP context activation to the GGSN 108 either as a separate IP packet or piggybacks the information with existing signalling for fast handover or context transfer. If the infoπnation is sent by using the fast handover procedure (i.e. the procedure used to perform a fast IP handover as described in <draft-ietf-mobileip- fast-miρv6-06.txt>), the message carrying that information would be the HI message. The context transfer procedure is another method that could be used to carry that information used to transport user's context in the IP handover (defined in <draft-ietf- seamoby-ctp-01.txt>).

The criteria that indicates to the MN that link level authentication and PDP context activation is to commence is, for example, decreasing signal strength or some added information provided by the WLAN network which indicates that the MN may be about to leave the WLAN network.

The information sent in the packet from the MN to the SGSN 105 would include, the IMSI of the MN, the Node B (base station 110) identifier, the QoS profile for PDP context activation and an indication that an IP address will be needed at the target UMTS network.

The exact information contained in the PDP profile would include, for example, PDP Type, PDP Address, Access Point Name, QoS Negotiated, TEID (Tunnel Endpoint Identifier), NSAPI (Network Layer Service Access Point Identifier), MSISDN (Mobile Subscriber International ISDN Number), Selection Mode, Charging Characteristics, Trace Reference, Trace Type, Trigger ID, OMC Identity and PDP Configuration Options.

When the GGSN 104 receives this information from the MN 100 (step 1), it forwards the IMSI to the appropriate SGSN 105 (step 2) in its domain through the Iu interface. The correct SGSN 105 in its domain is chosen based on the Node B 110 identifier. The GGSN 104 has to maintain a mapping of SGSN 105 to Node B 110 identifiers which it consults in order to choose the correct SGSN 105. Previously, the GGSN 104 has not maintained such information which clearly would aid in reducing the time taken by link layer attach procedures. The GGSN 104 also sends the Activate PDP context message which contains the PDP profile information to the SGSN 105. Once the SGSN 105 receives the IMSI and PDP profile information, it begins to authenticate the MN at the link layer (L2) and also establishes the PDP contexts, in parallel as depicted in Figure 2 (steps 5 and 6).

The SGSN 105 sends an Authentication Data Request (IMSI) to the HLR 106 (step 3). The HLR 106 then answers with an Authentication Data Response (AVI, AV2...A ) (step 4). Step 4 also involves the sending of a session key which is derived from a secret key shared between the HLR 106 and the MN 100. The SGSN 105 then sends a User Authentication Request (RAND(i)||AUTN (i)) to the GGSN (step 7). The method for calculating the authentication request is prior art. The SGSN 105 also calculates the Expected Response (ERES (i)) and stores it along with the IMSI of the MN. As stated earlier, the SGSN 105 establishes the link layer authentication in parallel with the requisite PDP contexts for the MN based on the information received by the GGSN 108 from the MN (step 5). This process also allows the SGSN 105 to choose the GGSN 108 in the target UTMS network which can satisfy the MN's IP required PDP profile. The GGSN 108 which is chosen to host the MN then informs the SGSN 105 that sends in the request about the successful establishment of PDP context (step 6). The SGSN 105 then informs the GGSN 108 in the target UTMS network that it is in communication with the WLAN network A. The AR 103 of the WLAN network A is then informed about the GGSN 108 in the target UMTS network which will host the MN. An IP address for the MN is allocated using either a stateful or a stateless means. This information is also passed on to the GGSN 104 in contact with the AR 103 of the WLAN network A to be forwarded to the MN.

When the GGSN 104 receives the authentication information, i.e. the ID of the GGSN 108 in the target network and the IP address of the MN (step 7), it packages this request and sends it to the MN (step 8) via the Internet and the AR 103 of the WLAN. This message is optionally encrypted using the session key shared between the MN and its HLR.

When the MN receives the infoπnation provided in step 8, it decrypts the message and authenticates the network calculating the Response (RES (i)). The MN also configures its 3G interface for packet sessions with the new IP information.

When the MN moves into the UMTS domain (step 9) (or when the MN chooses to prepare for handover), it sends the RES (i) along with its IMSI information, as part of the UMTS attach, to the SGSN 105 via the associated Node B 110 which then authenticates the MN. The MN can then immediately engage in packet sessions using the configured PDP context.

When the request from the MN is received by the GGSN 108 in the target UMTS network, it is necessary to associate the Node B information with a SGSN 105 in the system. Therefore, each GGSN should store a mapping of Node Bs to SGSNs. This is centrally controlled by the operator. Furthermore, this association mapping will generally last for a long time and sometimes will be relevant for the lifetime of the network, in which case update algorithms may not be needed to check the consistency of the mapping.

The GGSN 104 in some cases, does not know which SGSN 105 to contact such as when the MN sends all the information for the L2 and L3 procedures except the Node B information to the WLAN AR 103. In this scenario, the AR 103 will then identify the GGSNs (3G/GPRS networks) in its neighbourhood (with the help of protocols such as CAR discovery) that the MN is authorised to roam in. This embodiment, however, assumes that CAR discovery is implemented in the AR. The AR 103 then forwards the information that the MN has sent to all the GGSNs. The GGSNs receiving the information then initiate the same procedure for authenticating the MN at the L2 layer as described previously but store the expected response from the MN at all the SGSNs in the 3G network and also establish GTP tunnels to all the SGSNs. These tunnels have a limited lifetime or, once the MN attaches to a particular Node B and SGSN, the other tunnels will be removed. After establishing the PDP context and generating the authentication challenge as described earlier, each GGSN will send a challenge to the MN. The MN has to send in turn responses to each GGSN. Once the responses are verified, separate tickets are generated with a given lifetime for each of the networks. The associated GGSNs will send back the tickets, possibly encrypted, to the MN. When the MN hears a Node B signal, it sends the appropriate ticket to that Node B and rejects the other tickets. In most practical cases, the AR will find at least one 3G/GPRS network in its neighbourhood that the MN is allowed to roam in.

In Figure 2, two GGSNs 104 and 108 are present, the first 104 is in contact with the AR 103 of the WLAN network and the second 108 will host the PDP context of the MN. However, if the first GGSN 104 which is in contact with the WLAN network is capable of hosting the PDP context then there would be a need for only a single GGSN (as in Figures 3 and 4 described below).

A stateful means of providing the MN with an IP address has been described which involves a DHCP (Dynamic Host Configuration Protocol) server providing an IP address for the MN (this is a standard way of obtaining an IP address). However, IPv6 nodes are capable of autoconfiguring their addresses as described in RFC 2462 (see S. Thomson et al IPv6 Stateless Autoconfiguration RFC 2462 December 1998). For this purpose, the GGSN automatically and periodically sends Router Advertisement messages towards the MN after a PDP context of the type IPv6 is activated. Since in the present invention the Ipv6 prefix of this GGSN may be different to that of the GGSN known to the MN, prefix of this GGSN is also packaged in the information sent back to the MN in order to help the MN autoconfigure its IP address whilst still connected to the WLAN AR.

Although the MN is described as sending in a response to the challenge issued by the SGSN after moving into the UMTS (step 9), the response should preferably be sent via the AR of the WLAN to the GGSN before the MN decides to connect to the Node B, i.e. the network authentication by the MN and the MN authentication by the network is also performed before connecting to the Node B. In order to complete the authentication, the GGSN would then send a "ticket" after making sure that the response is correct. The MN would then send the "ticket" to the Node B along with its IMSI. This "ticket" may be encrypted using the key shared by the MN and the HLR. The "ticket" is simply a notification from the UMTS that everything is ready and set up for the MN. The "ticket" can be encrypted to ensure that no one else can see it. Preferably, this should be the default means of operation of the present invention. Partial authentication by using step 9 should only be used when the MN is unable to send a response via the WLAN AR due to being cut off prematurely before sending a response to the challenge or being cut off before getting a "ticket".

In the method described with reference to Figures 1 and 2, only part of the authentication procedure (i.e. network authentication by the MN) need be performed before the movement of the MN into the UMTS network. Preferably, the complete authentication procedure is performed before movement occurs, i.e. network authentication by the MN and MN authentication by the network.

Reference should now be made to Figures 3 and 4 which depict a second preferred embodiment of the present invention.

In this situation, the MN will be moving into the PS (packet switched) core network rather than being supposed to be attached to the PS core network (as in Figures 1 and 2).

In Figure 3, a simplified presentation of the second preferred embodiment of the present invention is shown for handover between a WLAN network A as a GPRS network B. This Figure is substantially the same as Figure 1 except that there is only a single GGSN 104 which is able to act as the AR 103 for the WLAN network A and can host the PDP contexts of the MN 100.

In Figure 4, the SGSN 105 starts the authentication of the MN 100 by first obtaining the authentication parameters from the HLR 106 and then sending a Proxy Authentication and Ciphering Request message to the GGSN via the WLAN network. In Figures 3 and 4 the GGSN 104 acts as an AR 103 in the GPRS network B from the point of view of the MN and is capable of receiving a handover trigger indication from the WLAN network A. As mentioned earlier, there is a need for only one GGSN 104 in this preferred embodiment since it is capable of hosting the PDP contexts of the MN as well as acting as an access router 103 for the WLAN network A and the GPRS network B.

The following information should be carried by the handover trigger indication (Step 1 in Figure 4):

• MN' s identifier i.e. MN's IMSI • MN's IP address • QoS contexts of the IP sessions already running by the MN which are to be moved from the WLAN to the GPRS network • Authentication Information, i.e. if an EAP-SIM procedure is used for authentication then the information could be the ERs / SIM / START message.

After having received the handover trigger indication, the GGSN (nAR) will send a notification to the SGSN (PDU Notification Request Message) in order to indicate that the PDP contexts for the PDP addresses should be activated. The method by which the GGSN discovers the target SGSN has been described in connection with Figure 2 and consists of maintaining a mapping table between the possible target SGSNs and the Node Bs. Thus when the GGSN receives the handover trigger indication where there is information about the target cell where the MN is going to be located in the GPRS network, the GGSN can easily identify which is the target SGSN which will support the MN.

The following information should be carried by the PDU Notification Request message (Step 2(i) in Figure 4):-

• MN's identifier, i.e. MN's IMSI • The "Cause" of sending the "PDU Notification Request" message from GGSN to SGSN • QoS requirements for activation of the necessary PDP contexts in the GPRS network - The GGSN should convert the QoS contexts in the handover trigger indication into the QoS requirements to activate the PDP contexts • Authentication information if it was carried by the handover trigger indication.

The PDU Notification Request message is sent to the SGSN when the GGSN receives an external PDU (in this case, an IP message) which is targeted at a PDP address which is not yet associated to any PDP context. The purpose is to activate a PDP context for that PDP address. In this case, the transmission of that notification is also triggered when a specific external indication for handover is received at the GGSN (i.e. it is not a PDU targeted at a PDP address). The purpose is, however, the same, i.e. to create a PDP address as well as the associated PDP contexts and to perform MN authentication if the MN is not yet authenticated by the target network. The reasons for including the aforementioned parameters into the PDU Notification Request message are as follows :-

1. The "Cause" should be established so that it is clear whether the MN is supposed to be joining the PS core network or whether the MN is entering the PS core network, i.e. the values for "Cause" could be either: a) MN entering PS core network (or incoming PDU due to MN's movement into PS core network), or b) MN is already joining PS core network (or incoming PDU not due to MN's movement into PS core network).

If the MN is supposed to be already attached to the PS core network ("Cause" (b) above) then the SGSN performs as in Figure 2, i.e. MN is already authenticated by the target UMTS network. If the MN is not authenticated ("Cause" (a) above) then the SGSN should start authentication as depicted in Figure 4.

2. QoS requirements. This parameter is needed to create a PDP context with these QoS requirements. This parameter is also needed if the "Cause" parameter is set to (a) MN entering PS core network.

3. Authentication parameters are needed to carry authentication information to the SGSN. These parameters are also needed if the "Cause" parameter is set to (a) MN entering PS core network.

The authentication information received in the handover trigger indication should be converted to specific GPRS authentication parameters. This could be carried out by the GGSN directly or possibly by means of the help of an AAA (Authentication Authorisation Accounting) server inside the GPRS network domain.

When the PDU Notification Request message has been received by the SGSN (having a "Cause" value set to (a) MN entering PS core network and the MN has not been authenticated), then the SGSN should start performing the MN authentication by the target network, i.e., steps 3 to 6 in Figure 4 (using a proxy server which is not depicted in the drawings). If the MN is already authenticated by the target network then steps 2(i), (ii) and (iii) should be followed by step 7 and the method will then be substantially the same as that described with reference to Figure 2.

The SGSN will contact the HLR (steps 2(ii) and 2(iii) in Figure 4) in order to obtain the MN authentication parameters. The SGSN will then send a Proxy Authentication and Ciphering Request message to the GGSN (Step 3 in Figure 4). In this situation, the SGSN contacts the MN which is in the WLAN network through the GGSN (acting as the network Access Router) so that the authentication message is transmitted to the MN through the GGSN via the WLAN network.

When the GGSN receives the "Proxy Authentication and Ciphering Request" message, it is converted into a specific authentication protocol used by the MN (e.g. EAP-SIM) (Step 4 in Figure 4) which is then sent to the MN.

When the MN receives the authentication message, it then replies with a further authentication message (Step 5 in Figure 4). In this example, the messages shown are "ERq /SIM/Challenge" (Step 4) and "ERs/SIM/Challenge" (Step 5). The GGSN will then convert the authentication message received in Step 5 into a "Proxy Authentication and Ciphering Response" message which is sent to the SGSN (Step 6 in Figure 4). The receipt of this message by the SGSN completes the MN authentication procedure.

If the MN's authentication by the target network is successful and the SGSN can support the PDP contexts with the QoS requirement, then the SGSN replies to the PDU Notification Request message in Step 2 with a PDU Notification Response message (Step 7 in Figure 4). This message indicates "Request Accepted". The GGSN will then understand that the MN has been successfully authenticated and that PDP context activation will follow (Step 8 in Figure 4).

Alternatively, if the MN's authentication procedure was successful but the SGSN cannot support the requirements of the MN, then the SGSN replies with a PDU Notification Response message indicating the cause of rejection (such causes are already defined in the standard e.g. "no resources available", "service not supported" etc). The GGSN then understands that the MN is successfully authenticated but the PDP context will not be activated (Step 10 in Figure 4).

Furthermore, if the MN authentication procedure is not successful, the SGSN will reply with a PDU Notification Response message which indicates the cause of the rejection. In this case, the cause of rejection would be "MN not authenticated successfully" and step 10 would follow.

If the SGSN is able to support the PDP context required by the MN, then it sends a "Create PDP Context Request" message to the GGSN (Step 8 in Figure 4). The GGSN will then reply with a "Create PDP Context Response" message to the SGSN (Step 9 in Figure 4).

Since the SGSN is aware that this procedure was initiated for a MN entering the GPRS PS core network, it should finish at this point the PDP Context Activation procedure.

Finally, the GGSN replies to the message received in step 1 ("handover trigger indication") by sending a "handover trigger response" which indicates whether the authentication procedure was successful or not. For example, in the case where EAP- SIM authentication is used then a "EAP success" message would be carried in the response and also information regarding whether the PDP context has been activated successfully or not. In addition, the attach and PDP context related parameters (e.g. P- TMSI) should be carried by this message. The WLAN network will forward these parameters to the MN. Although Figure 4 suggests fast handover signalling is to be used, other types of signalling could be used with the same purpose.

After finishing step 10 the MN is successfully authenticated in the target GPRS network with the PDP contexts already actuated. When the WLAN network receives the "handover trigger response" from the GPRS network, the MN can be moved from the WLAN to the GPRS network.

Since the MN is the only MN which knows the key for the GPRS session (calculated within the authentication procedure), there is no possibility of a different MN supplanting the legitimate MN.

During movement the MN will only have to obtain L2 connectivity to the GPRS network (and also Iu connection the case of UTRAN/GPRS in order to establish the RABs (Radio Access Bearers). These steps are carried out by the "Service Request" procedure in the GPRS specification (defined in 3GPP TS 23.060).

Clearly, the fact that the authentication and PDP context activation procedures are not perfonned during handover but prior to movement from the WLAN to the GPRS network will considerably reduce handover delay times. Although Figures 1 to 4 relate to the handover between a first WLAN network and a second cellular network, it is clear that the present invention could also be utilised in various handover scenarios where the first communication network is, for example, a different high-speed wireless technology based network. Clearly, there are many alternatives for the second cellular network rather than a GPRS, i.e. networks which employ packet switching and hence require the establishment of PDP contexts.

The third preferred embodiment of present invention provides a method whereby the PDP contexts can be maintained when the MN moves out of the GPRS network to another communication network and subsequently returns to the GPRS network.

When a MN moves from a GPRS network to any other access network, e.g., a WLAN network, the MN is noπnally detached and the PDP contexts associated with that MN are deactivated. Accordingly, when the MN decides to return to the GPRS network, it will have to perform the attach and authentication procedures as well as the activation of the necessary PDP contexts once again.

The attach, authentication and PDP context activation procedures are time consuming. Therefore, the handover performance in an intersystem handover situation is very inefficient, particularly when the target network is GPRS. The first and second embodiments of the present invention try to optimise this handover performance during an intersystem handover when the MN is detached and the PDP context deactivated in the GPRS network.

According to the third preferred embodiment of the present invention the MN remains attached to the GPRS network, i.e. the PDP contexts are maintained when the MN moves from the GPRS network to any other access network. Consequently, when the MN moves back to the GPRS network for a second time and subsequent times, it will not have to waste time performing attach, authentication and PDP context procedures so that the handover delay time can be reduced considerably.

The main disadvantage in maintaining the PDP contexts is that the PDP contexts could be considered to be invalid. This could occur if the ongoing applications running on the MN are completely different to those which the PDP contexts were originally activated for, i.e. the MN has moved from the GPRS network to another access network and has started to use different applications with other requirements before returning to the GPRS network. This could imply either a modification in the QoS requirements for the maintained PDP contexts or more drastically, the release of the maintained PDP contexts and the later activation of new PDP contexts. In both cases, the signalling generated is practically the same as the signalling generated when the maintenance of PDP contexts is not utilised.

The third preferred embodiment of the present invention can be achieved by modifying the value of a timer which already exists in the SGSN in the GPRS network. The modification will depend on the MN's multi-access capabilities.

The timer concerned is the RAU timer (Routing Area Update timer), e.g. T3312 specified by the standard 3GPP TS 24.008. The RAU timer performs the RAU procedure which is used by a roaming MN to inform the PS domain about its location in a certain area. The RAU timer is triggered when the MN goes to "PMM-IDLE" state from "PMM- CONNECTED" state (for Iu mode) or to "STANDBY" state from "READY" state (for Gb mode). Every time the timer expires, the MN should initiate the RAU procedure and the timer is reset. If the MN does not initiate the RAU procedure (this will occur when the MN abandons the GPRS network on moving to another access network), the network automatically performs detach and consequent resource release, i.e. PDP context release for that MN.

The value of the RAU timer is given to the MN by the SGSN in the GPRS network during the attach procedure (i.e. "Attach Accept" message) and it is assumed that the value of the timer is preconfigured in the GPRS network by the operator and that the value is the same for all of the MN's being attached to the GPRS network.

In accordance with the present invention, the SGSN will allocate different values for the RAU timer depending on the multi-access capabilities supported by the MN (the SGSN is aware of the MN's capabilities as a result of the "Attach Request" message sent by the MN). If the MN is multi-access capable, then the value for the timer should be longer than the value given to a MN which is not multi-access capable. In this way, the initiation of the RAU procedure (which the MN cannot perfoπn whilst using the WLAN network) will be delayed until the MN is supposed to be back in the GPRS network where the MN can perfonn the RAU procedure. As a result, multiaccess capable MNs are able to move to any other access technology and afterwards move back to the GPRS network having maintained the attach, authentication and PDP context activation procedures. This method is particularly pertinent to an MN which is only capable of using one radio at a time. Clearly, an MN with two radios would be able to maintain PDP contexts whilst simultaneously using a WLAN network. This prefeπed embodiment of the present invention would be particularly useful in a scenario where there is temporary missing network coverage or where there are multiple GPRS networks and roaming is heavily utilised. In the case of multiple GPRS networks, one could envisage the situation where a car in which the MN is being used travels between networks having different operators requiring constant switching between the operators.

It should be noted that whilst the aforementioned embodiments are exemplifying embodiments of the invention, there are several variations and modifications which may be made to the disclosed solution without departing from the scope of the present invention as defined herein.

Claims

CLAIMS:

1. A method for ensuring continuity of a communication session when a user equipment hands over from a first communication network to a second cellular communication network comprising the steps of performing an authentication procedure for a packet data session with the second network whilst still being attached to the first network and simultaneously performing a packet data session establishment procedure with the second network whilst still being attached to the first network.

2. A method as claimed in Claim 1, wherein the authentication procedure includes authentication of the second network by the user equipment.

3. A method as claimed in Claim 2, wherein the authentication procedure also includes authentication of the user equipment by the second network.

4. A method as claimed in any preceding claim, wherein the first communication network is a WLAN network and the second communication network is a cellular network.

5. A method as claimed in any preceding claim, wherein the information sent by the user equipment for authentication and packet data session establishment travels either as a separate IP package or is piggybacked with existing signalling.

6. A method as claimed in any preceding claim, wherein the gateway node between the first and second communication networks is able to act as an access router for the first network and is able to host the packet data session in the second network.

7. A method for ensuring continuity of a communication session when a user equipment hands over from a first communication network to a second cellular communication network wherein attachment of the user equipment to the second network is maintained after the user equipment moves away from the coverage area of the second network for a predetermined time in order to allow the user equipment to return to the second network without having to repeat an authentication procedure and a packet data session establishment procedure before handing over to the second network.

8. A method as claimed in any preceding claim, further comprising the step of releasing the packet data session if the user equipment does not handover to the second network within a predetennined time thus requiring the user equipment to repeat the authentication procedure if moving towards the second network for a further time.

A method as claimed in any preceding claim, comprising the following steps: (i) the user equipment sends a handover trigger indication to a gateway node in the second network, the handover trigger indication comprising the user equipment identification parameters and the packet data protocol profile (ii) the gateway node sends the user equipment identification parameters and the packet data protocol profile to the serving node in the second network; (iii) the serving node contacts the home location register to obtain the user equipment authentication parameters; (iv) the serving node sends a packet data protocol profile request to the gateway node; (v) the gateway node responds by sending a packet data protocol profile response to the serving node; (vi) the serving node sends authentication information to the gateway node; (vii) the gateway node sends the authentication information to the user equipment; (viii) the user equipment authenticates the second network; (ix) the user equipment sends a response to the serving node and moves into the second network.

10. A method as claimed in any preceding claim, comprising the following steps:

11. A communication system comprising a user equipment, a first communication network and a second cellular communication network, the system being arranged to enable continuity of a communication session when the user equipment moves from the coverage area of the first network to the coverage area of the second network, wherein means are provided to simultaneously perform an authentication procedure for a packet data session with the second network and perform a packet data session establishment procedure with the second network whilst the user equipment is still attached to the first network.

12. A method and communication system for ensuring continuity of a communication session when a user equipment hands over from a first communication network to a second cellular communication network substantially as herein described with reference to Figures 1 to 4.