GB2394148A

GB2394148A - Method of routing messages to a roaming subscriber unit

Info

Publication number: GB2394148A
Application number: GB0223855A
Authority: GB
Inventors: Jin Yang; Jakub Tichy; Gerry Foster
Original assignee: Motorola Inc
Current assignee: Motorola Solutions Inc
Priority date: 2002-10-12
Filing date: 2002-10-12
Publication date: 2004-04-14
Anticipated expiration: 2022-10-12
Also published as: AU2003282113A1; WO2004036948A2; AU2003282113A8; GB2394148B; WO2004036948A3; EP1554908A2; GB0223855D0

Abstract

A method of routing information in a communication system where a serving communicating unit serves a plurality of subscriber units with a communication resource and is identified by both a public and a private address. The method includes the steps of a subscriber unit (310) roaming from a home communication network (110) to a visited communication network (150) and requesting (350), using a GPRS transport protocol message, a communication resource. A visited serving communication unit (SGSN) (152) processes the request in order to extract a private address. In response to determining that the private address identifies a serving communication unit (SGSN) of a different communication network (110), the serving communication unit (152) incorporates its public address in the request and forwards (375) the request to a home serving communication unit (112). In this manner, subsequent messages to the subscriber unit can be routed via the visited serving communication unit using its public address. A communication system and serving communication units, for example a serving GPRS service node and a gateway GPRS service node, are also provided.

Description

2394 1 48

- 1 COMMUNICATION SYSTEM, SERVING COMMUNICATION UNIT AND

METHOD OF ROUTING INFORMATION

Field of the Invention

This invention relates to the allocation of addresses in order for data to be routed to communication units. The invention is applicable to, but not limited to, addresses used by the Core Network entities when subscriber units 10 roam between public land mobile networks (PLMNs).

Background of the Invention

Present day communication systems, both wireless and 15 wire-line, have a requirement to transfer data between i communication units. Data, in this context, includes signalling messages, multimedia, speech communication, etc. Such data transfer needs to be effectively and efficiently provided for, in order to optimise use of 20 limited communication resources.

For data to be transferred across and between communication networks, a communication unit addressing protocol is required. The communication units are 25 generally allocated addresses that are read by a communication bridge, gateway and/or router, in order to determine how to transfer the data to the addressed unit.

The interconnection between networks is generally known as internetworking (or interned).

Networks are often divided into sub-networks, with protocols being set up to define a set of rules that allow the orderly exchange of information. Currently,

- 2 - the two most popular protocols used to transfer data in communication systems are: Transfer Control Protocol (TCP) and Internet Protocol (IP). In all but the simplest of communication systems, these two protocols 5 often work as a complementary pair. The IP portion corresponds to data transfer in the network layer of the well-known OSI model and the TOP portion to data transfer in the transport layer of the OSI model. Their operation is transparent to the physical and data link layers and 10 can thus be used on any of the standard cabling networks such as Ethernet, FDDI or token ring.

The Internet Protocol adds a data header on to the information passed from the transport layer. The 15 resultant data packet is known as an Internet datagram.

The header of the datagram contains information such as destination and source IP addresses, the version number of the IP protocol etc. An IP address is assigned to each node and network element. It is used to identify 20 the location of the network and any sub-networks.

Each node using TCP-IP communications requires an IP address that is then matched to its token ring or Ethernet MAC address. The MAC address allows nodes on 25 the same segment to communicate with each other. In order for nodes on a different network to communicate with one another, each node must be configured with an IP address. 30 Nodes on a TCP-IP network are either hosts or gateways.

Any nodes that run application software, or are terminals, are defined as hosts. Any node which is able to route TCP-IP packets between networks is called a TCP/IP gateway node. A TCP/IP gateway node must have the

À 3 necessary network controller boards to physically interface to other networks.

A typical IP address consists of two fields:

5 (i) The prefix field, where a network number

identifies the network associated with that particular address, and (ii) The suffix field, where a host number

identifies the particular host within that network.

The IP address is 32 bits long and can therefore theoretically address 232 (over four billion) physical networks. One problem, however, associated with using an IP address containing prefixes and suffixes lies in the 15 decision on how large to make each field. If the prefix

is too small, only a few networks will be able to be connected to the Internet. However, if the prefix is made larger, then the suffix has to be reduced, which results in a network being able to support only a few 20 hosts.

The present version of the Internet protocol addressing scheme (IPv4) can accommodate a few very large networks or many small networks. In reality, a reasonable number 25 of networks of various sizes are required to be supported. However, most organizations tend to have their own IP addressing scheme, arranged to accommodate a larger network than they generally need, to allow for future network expansion.

As a consequence, the current version of Internet Protocol (IPv4) has scarce addressing space and future versions are currently being developed. It is envisaged that each Public Land Mobile Network (PLAN) will be

À 4 unable to allocate a unique permanent IP address to each subscriber unit using IPv4. Moreover, even in the event that IPv6 were to be deployed in the future, many networks will still consist of legacy networks 5 implementing IPv4.

An IP address can be defined in the form: aaa'.'bbb'.'cac'.'ddd'; Where: 'aaa', 'bbb', 'can' and 'ddd' are integer values in the range 0 to 255.

On the Internet the 'aaa'.'bbb'.'ccc' portions normally 15 define the subnetwork and the 'ddd' portion defines the host. Such numbering schemes are difficult to remember.

Hence, symbolic names (often termed domain names) are frequently used instead of IP addresses to identify individual communication units.

Normally, the DNS server is reachable by all the hosts on the network via the IP transport protocol. Therefore the DNS protocol for performing address lookup can be carried over IP.

The directory network services on the Internet determine the IP address of the named destination user or application program. This has the advantage that users and application programs can move around the Internet and 30 are not fixed to a particular node and/or IP address.

In systems employing a limited number of addresses by which to identify individual communication units, a technique called dynamic addressing is used. Dynamic

- 5 - addressing requires a pool of addresses to be maintained by an address allocation server, for example a Dynamic Host Configuration Protocol (DHCP) server. Whenever a host is connected to a network, a signalling process is 5 performed between the host and DHCP server to assign an available IP address to the host. In order to do so, the host needs to send the DHCP server its unique ID. When the signalling process is de-activated, the IP address will be returned to the addressing pool and will wait to 10 be assigned to other terminals.

If a packet data subscriber unit initiates an Internet connection, the DHCP server recognizes the need to identify the subscriber unit and typically informs a 15 domain name server (DNS) that a new Internet Protocol address assignment has occurred. Subsequently, the local DNS can then map the subscriber unit's domain name to an Internet Protocol address allocated by the DHCP, and pass the address information to an Internet Host.

Due to the static nature of typical devices that use IP, such as networked personal computers (PCs) and servers, DHCP has been widely used in the Intranet environment to allocate IP addresses dynamically to any hosts that are 25 connected to a network.

However, it is clear that such an arrangement is unacceptable in a wireless domain when the communicating unit requiring an IP address, is not physically connected 30 to the Internet. With such wireless technology, the subscriber unit needs to have previously established a logical connection with the Internet, in order to have been allocated an IP address and access Internet services, information and applications. This logical

- 6 - connection is generally referred to as a packet data protocol (PDP) context.

Furthermore, as wireless subscriber units will not be 5 permanently connected to the Internet, there will be many occasions when the subscriber unit will be in a mode where no PDP context with the Internet has been established. 10 Due to the recent growth in data communication, particularly in Internet and wireless communications, there exists a need to provide TCP-IP data transfer techniques in a wireless communications domain.

15 An established harmonized cellular radio communication system is GSM (Global System for Mobile Communications).

An enhancement to this cellular technology can be found in the General Packet Radio System (GPRS), which provides packet switched technology on a basic cellular platform, 20 such as GSM. A further harmonised wireless communications system currently being defined is the universal mobile telecommunication system (UMTS), which is intended to provide a harmonised standard under which cellular radio communications networks and systems will 25 provide enhanced levels of interfacing and compatibility with other types of communication systems and networks, including fixed communication systems such as the Internet. 30 Information to be transmitted across the Internet is packetised, with packet switching routes established between a source node and a destination node. Hence, GPRS and UMTS networks have been designed to accommodate packet switched data to facilitate Internet services,

- 7 - such as message service, information service, conversational service and casting service.

When a GPRS or UMTS user roams to a foreign network, in 5 many cases the user needs to use the gateway GPRS service node (GGSN) function from the user's home network to access Internet or Intranet data. The traffic is transported across a Gp interface over an inter-PLMN backbone. Although, from a roaming support viewpoint, it 10 would be better to use public IP addresses for the network elements such as the serving GPRS service node SGSN, the GGSN, and a Charging Gateway etc., notably in many cases Operators prefer to use private IP addresses.

15 A problem arises when private IP addresses are used for intra-PLMN backbone operation. In this scenario, normal IP routing between two PLMNs cannot be performed, as there is no unique address for the network elements between the two respective PLMNs.

A known solution to this problem, in general, is to deploy Network Address Translation (NAT) technology at a border gateway (BG) within the PLAN. In this manner, the source addresses of IP packets from SGSN are translated 25 at the BG to public IP addresses. The IP packets are then forwarded to a GGSN in another PLMN.

However, this solution cannot be easily used at the BG when a GPRS transport protocol (GTP) is implemented as 30 described in RFC 1631 (The IP Network Address Translator (NAT). K. Egevang, P. Francis. May 1994.). Basically, NAT technology changes only the source and/or destination IP address in the header of an IP packet. NAT may also

- 8 - be configured to change the source and/or destination port numbers in the header of an IP packet.

However, as the SGSN address of a PDP context is 5 negotiated by relevant GTP messages (i.e. "Create PDP context" and "update PDP context"), the NAT will fail to cope with the addresses when the SGSN and GGSN IP addresses are negotiated in the data packet payload using application layer protocols.

The common practice to work around this problem is to develop Application Layer Gateway (ALG) software, which interpret the relevant protocol messages. The ALG software is then able to intercept packets and modify the 15 packet addresses if necessary. The ALG is normally combined with the BG and shares a common platform.

The inventors of the present invention have recognized significant limitations and problems in the use of ALG to 20 resolve the addressing problem when a data packet is communicated between two PLMNs. In particular, a new product, i.e. an ALG for GTP operation, has to be developed. This means that standard NAT products cannot be used directly. Furthermore, the performance of the BG 25 is seriously impacted, as the ALG would need to check each GTP packet and determine if it includes a target message. Additionally, the use of the ALG would increase the system latency, as each packet will be delayed whilst being processed by the ALG. Even worse, when encryption 30 on GTP control (GTP-C) messages is performed, the ALG has no way to decode the GTP-C messages. To enable the ALG to decode such messages, extra functionality has to be incorporated into the ALG to deal with issues such as encryption key management. Thus, the impact on the

- 9 performance makes an ALG-based solution particularly unattractive. As a result, a need exists to provide a communication 5 system, a communication unit and method of routing information wherein the abovementioned disadvantages may be alleviated.

Summary of the Invention

In a first aspect of the preferred embodiment of the present invention, a method of routing information in a communication system is provided, in accordance with Claim 1.

In a second aspect of the preferred embodiment of the present invention, a communication unit is provided, in accordance with Claim 11.

20 In a third aspect of the preferred embodiment of the present invention, a communication system is provided, in accordance with Claim 12.

In a fourth aspect of the preferred embodiment of the 25 present invention, a serving communication unit is provided, in accordance with Claim 13.

In a fifth aspect of the preferred embodiment of the present invention, a gateway GPRS Service Node (GGSN) is 30 provided, in accordance with Claim 17.

In a sixth aspect of the preferred embodiment of the present invention, a communication system is provided, in accordance with Claim 18.

- 10 In a seventh aspect of the preferred embodiment of the present invention, a serving communication unit is provided, in accordance with Claim 24.

In accordance with an eighth aspect of the present invention, there is provided a storage medium, as claimed in Claim 25.

10 Further aspects of the present invention are as claimed in the dependent Claims.

In summary, the inventors of the present invention

propose that, instead of relying on ALG to intercept and 15 modify the relevant GTP messages, the functionality of the SGSN within the network is enhanced. In particular, a visited SGSN determines when a PDP context message is destined for an alternative network, and in response to such a determination replaces the home network's SGSN 20 private address with the visited SGSN's public address, so that subsequent messages can be routed to the subscriber unit when supported by the visited SGSN.

Brief Description of the Drawings

Exemplary embodiments of the present invention will now be described, with reference to the accompanying drawings, in which: 30 FIG. 1 illustrates an architecture involving intra-PLMN and inter-PLMN networks, adapted to support the preferred embodiments of the present invention;

- 11 FIG. 2 is a block diagram illustrating the address interaction between an SGSN and a BG operably coupled to a NAT adapted to support the inventive concepts of the preferred embodiments of the present invention; and FIG. 3 illustrates a block diagram and associated method to support a subscriber unit performing inter-PLMN roaming, in accordance with the inventive concepts of the preferred embodiments of the present invention.

Description of Preferred '!rnhodiments

Referring now to FIG. 1, an architecture involving intra PLMN and interPLMN networks is illustrated, where the 15 architecture is adapted to support the preferred embodiments of the present invention. The preferred embodiment of the present invention is described with reference to communication between two PLMNs (PLAN A 110 and PLAN B 150) via an interPLMN backbone 140 and a 20 packet data network 130. However, it is within the contemplation of the invention that the inventive concepts described herein are equally applicable to interaction and address manipulation between other network types.

Every intra-PLMN backbone network 120, 160, is a private IP network intended for packet domain data and signalling only. A private IP network is an IP network to which an access control mechanism is applied in order to achieve a 30 required level of security. As shown, the two intra-PLMN backbone networks 120, 160 are connected via the Gp interface 124 using Border Gateways (BGs) 118, 158 and the inter-PLMN backbone network 140. The particular inter-PLMN backbone network 140 functions under a roaming

- 12 agreement that includes the security functionality of the respective BGs 118, 158. The BGs 118, 158 are not defined within the scope of the packet domain. The inter-PLMN backbone 140 may be a Packet Data Network such 5 as PDN 130. An example of the PDN 130 would be the public Internet or a leased line.

SGSNs 112, 114, 152 are operably coupled to respective GGSNs 116, 156 and BGs 118, 158 via the respective intra 10 PLMN backbones 120, 160, as known in the art.

In accordance with the preferred embodiments of the present invention, one or more SGSN 112, 114, 152 are adapted to provide enhanced features. Let us assume that 15 a subscriber unit is registered with PLMN A 110, but has roamed into PLMN B 150. Furthermore, let us assume that the subscriber unit wishes to communicate and transmits a create PDP context' message to its currently serving SGSN 152. The SGSN 152 in PLMN A 150 processes the PDP 20 context to determine if the target GGSN 116 belongs to another PLMN A 110. Preferably, checking the Access Point Name (APN) within the PDP context makes this determination. Notably, the SGSN (source node) is addressed using a private IP address, where each SGSN is 25 aware of a public IP address associated with it.

Therefore, if the SGSN 152 determines that the target GGSN 116 does belong to another PLMN, (PLMN A 110), the SGSN 152 incorporates the public IP address for the "SGSN 30 address" field within the "Create PDP context" message.

In this manner, the public IP address will be used by the NAT function at BG 158.

- 13 In a similar manner, during inter-SGSN handover of a data communication unit such as a GPRS unit, if the GGSN 116 associated with a PDP context belongs to another PLMN (PLMN A 110), the SGSN 152 again uses its public IP 5 address for the "SGSN address" field in the "Update PDP

context" message. In this way, subsequent data packets may be routed to the subscriber unit supported by SGSN 152. 10 As known, the NAT is configured with a static mapping facility to map between the public IP address and the private address for the respective SGSNs, as illustrated in FIG. 2. Referring now to FIG. 2, the mapping arrangement 200 is illustrated in more detail, but with 15 regard to PLMN A 110.

An SGSN 112 within PLMN A 110 includes a private IP address (10.1.1.1) 212 and an associated public address (195.1.1.1) 214. The SGSN 112 communicates PDP context 20 messages to its respective BG 118, including the private IP address (10.1.1.1) 212 and an associated public address (195.1.1.1) 214. The NAT 220, operating with the BG 118, performs standard network address translations using these private and associated public IP addresses 25 212, 214. In this manner, the BG is able to route messages to/from the respective SGSN.

Referring now to FIG. 3, a system architecture diagram 300 illustrates the particular process messages/steps 30 used in accordance with the preferred embodiment of the present invention. In particular, FIG. 3 illustrates a preferred example of how inter-PLMN roaming, between PLMN A 110 and PLMN B 150, is supported.

- 14 The relevant PLMN configurations are: PLMN A 110:

The home GGSN 116 has a private IP address (10.1.1.1) 212, and is associated with a public IP 5 address (195.1.1.1) 214.

The BG/NAT 118 of PLMN A 110 has a (permanent) static mapping from the private IP address (10.1.1.1) 212 to the associated public IP address (195.1.1.1) 214.

10 PLMN-B 150:

The visiting SGSN 152 also has a private IP address (10.1.1.1) 312, and is associated with a public IP address (196.1.1.1) 314.

The BG/NAT of the visited PLMN B 150 has a 15 (permanent) static mapping from (10.1.1.1) to (196.1.1.1).

A subscriber unit 310 associated with PLMN A 110 roams into PLMN B 150. The subscriber unit 310 requests, in 20 step 350, a PDP context indicating an APN in its home PLMN A 110. Within PLMN B 150, the Visiting SGSN (VSGSN) 152 attempts to resolve the APN within the PDP context message to the IP address of the GGSN 116 to be used. In this regard, the VSGSN 152 checks, in step 355, with the 25 local DNS server 330 associated with PLMN B 150.

If the local DNS server 320 does not include the required mapping, the local DNS server 330 sends a request to the DNS server 340 in PLMN-A 110. The request is, for 30 example, based on the "Operator-ID" part of the APN, or "root" of the ".gprs" domain. Such requests can be supported by, for example, GSM Association, as known to those skilled in the art.

- 15 The local DNS server 320 eventually resolves the mapping from the APN to the IP address of the DNS server 330 of the home GGSN (HGGSN) 116 in PLMN A 110. The local DNS server 320 and the home DNS server 330 preferably use the 5 standard address resolution protocol (ARP) to inform the VSGSN 152 of the APN.

The VSGSN 152 then sends a "Create PDP Context" request to the HGGSN 116. Notably, the VSGSN 152 processes the 10 "Create PDP Context" request received from the subscriber unit and determines that the identified GGSN 116 belongs to another PLMN (PLMN A 110). In this regard, in implementing the preferred embodiment of the present invention, the SGSN includes a receiver portion (not 15 shown) and a transmitter portion (not shown) for receiving and transmitting messages from/to other network elements or subscriber units. Furthermore, the SGSN includes one or more processors, for example digital signal processors or processing boards, to process and 20 interpret signals/messages. The SGSN processor(s) is also operably coupled to a memory element (not shown) to store address data.

More generally, the adaptation of one or more SGSN to 25 implement the aforementioned inventive concepts may be effected in any suitable manner. For example, new apparatus may be added to a conventional SGSN or alternatively existing parts of a conventional SGSN may be adapted, for example by reprogramming one or more 30 processors therein. As such, the required adaptation may be implemented in the form of processorimplementable instructions stored on a storage medium, such as a floppy disk, hard disk, PROM, RAM or any combination of these or other storage multimedia.

- 16 Thus, VSGSN 152 incorporates its public IP address (196.1.1.1) 312 into the value of "SGSN Address" field.

This message, as a GTP message, is then sent to the 5 address (195.1.1.1) 214 of the GGSN 116, in step 375.

The message is routed via BG B 158, in step 365, that translates the source address. It is also routed via BG A, in step 370, which translates the destination address (from SGSN's public address (195.1.1.1) 214) to the home 10 SGSN 112 private address (10.1.1.1) 212 of PLMN A 110.

The GGSN 116 records the VSGSN address (196.1.1.1) 312 as part of the POP context. After the PUP context has been set up, the GGSN 116 is now able to forward data packets 15 using the GPRS transport protocol (GTP) to the subscriber unit 310, in step 390. For example, let us consider downlink GTP packets addressed to the subscriber unit 310. The GTP packet is sent to VSGSN 152 using the public address (196.1.1.1) 314 of VSGSN 152. BG A 118 is 20 able to replace the source address of SGSN 112, in step 380. BG B 158 then translates the destination address from the public address (196.1.1.1) 314 of VSGSN 152 to PLAN B's 150 internal (private) address (10.1.1.1) 312, in step 385.

In this manner, GTP data packets can be routed between PLMNs, for example for a roaming subscriber unit, without incurring the addressing problems that currently require development of specific ALGs.

Advantageously, the above-mentioned inventive concepts can be incorporated as enhancements on the SGSN using a software upgrade, by re- programming one or more processors as described above.

- 17 A key benefit of the above-mentioned addressing methodology is that it allows the use of private address space for most addressing needs within a PLMN's network 5 infrastructure. This minimises the use of public IP addresses, as only a few network components that are directly involved in inter-PLMN communication (including SGSN, GGSN and DNS server) are allocated with public IP addresses. Although the invention has been described 10 with reference to inter-PLMN communication using GTP messages, with the address translation performed by the SGSN instead of the NAT, it is envisaged that the inventive concepts are equally applicable to any other wireless communication system supporting roaming of data 15 communication units.

It will be understood that the mechanism for resolving non-unique addresses between two networks, as described above, additionally provides at least the following 20 advantages: (i) No new (ALG) product needs to be developed.

Only currently available technology is required, such as an off-the-shelf NAT product used in combination with a 25 BG and enhanced SGSN functionality.

(ii) There is no impact on the BG performance, as the BG does not need to perform any additional functions such as provision of ALG.

(iii) Encryption, for example on GTP-C, can be used 30 without any limitation.

(iv) There is no impact on the standardization programs. (v) The enhancement to the SGSN functionality can be performed using a software upgrade.

- 18 The present invention finds particular application in wireless communication systems such as the UNITS or GPRS systems, employing GTP for packet data communication.

5 However, a skilled person would readily recognise that the inventive concepts contained herein are equally applicable to alternative fixed and wireless communications systems.

10 Whilst the specific, and preferred, implementations of the present invention are described above, it is clear that one skilled in the art could readily apply variations and modifications of such inventive concepts.

15 Thus, a communication system, serving communication units and a method of routing information between communication units, has been provided that alleviates some of the abovementioned disadvantages.

Claims

- 19 Claims 1 A method of routing information in a communication system

that includes a serving 5 communicating unit serving a plurality of subscriber units with a communication resource, wherein said serving communication unit is identified within the communication system by both a public address and a private address, the method comprising the steps of: 10 roaming from a home communication network (110) to a visited communication network (150) by a subscriber unit (310); requesting (350) a communication resource from a serving communication unit in said visited communication 15 network (150), by said subscriber unit (310) using a general packet radio system (GPRS) transport protocol (GTP) formatted message; the method characterized by the steps of: processing said request by said visited serving 20 communication unit (152) in order to extract a serving communication unit private address; determining that said serving communication unit private address identifies a serving communication unit of a different communication network (110); 25 incorporating, in response to such a determination, a public address (312) of said visited serving communication unit (152) in said request; and forwarding (375), by said visited serving communication unit (152), said request for a 30 communication resource to a serving communication unit (112) in said home communication network (110) of said subscriber unit.

- 20
2. The method of routing information in a communication system according to Claim 1, wherein said step of roaming, includes roaming, by said subscriber unit (310) from a home public land mobile network (110) 5 to a visited public land mobile network (150) that identifies a number of communication elements using private addresses.
3. The method of routing information in a 10 communication system according to Claim 2, wherein said step of requesting (350) a communication resource includes sending a 'create packet data protocol (PDP) context' GTP message to said home public land mobile network (110).
4. The method of routing information in a communication system according to Claim 3, wherein said step of sending a 'create PDP context' message is sent to an access point name (APN) in the home public land mobile 20 network (110).
5. The method of routing information in a communication system according to Claim 4, the method further characterized by the step of: 25 attempting to resolve said APN within said PDP context message to a public address of a home serving communication unit (156).
6. The method of routing information in a 30 communication system according to Claim 5, wherein said step of attempting includes sending a request for address information from a visited domain name server (320) to a home domain name server (340) located in said home public land mobile network (110).

- 21
7. The method of routing information in a communication system according to any preceding Claim, wherein said step of processing said request includes 5 processing said request to determine a private address of a home serving general packet radio system (GPRS) service node.
8. The method of routing information in a 10 communication system according to any preceding Claims, wherein said step of forwarding includes sending the request message, as a GPRS Transport Protocol message to a home gateway GPRS Service Node (HGGSN) (116) to record said public address of said visited serving communication 15 unit (152) as part of said request message.
9. The method of routing information in a communication system according to any preceding Claims, the method further characterized by the steps of: 20 routing (370) a message from said home communication network (110) to said subscriber unit via a border gateway (BG) in said visited communication network (150); and translating (385), by said BG, a destination 25 address of said message from said public address of said visited serving communication unit (152) to its private address.
10. A method of routing information in a 30 communication system according to any preceding Claim, wherein said communication system supports the universal mobile telecommunication standard (UMTS) or general packet radio system (GPRS) communication and said public address is an Internet Protocol address.

- 22
11. A communication unit adapted to perform the steps of any of method claims 1 to 10.

5
12. A communication system, for example one supporting a universal mobile telecommunication standard (UMTS) or a general packet radio system (GPRS), adapted to facilitate the operation of the steps of any of Claims 1 to 10.
13. A serving communication unit (152) for serving a plurality of subscriber units with one or more communication resources, wherein said serving communicating unit is identified by both a public address 15 and a private address, said serving communication unit (152) comprising: a receiver for receiving a general packet radio system (GPRS) transport protocol (GTP) formatted message request (350) for a communication resource from a 20 subscriber unit (310); the serving communication unit (152) characterized by: a processor that performs at least the following functions: processing said request in order to extract a 25 serving communication unit private address; determining whether said serving communication unit private address identifies a different serving communication unit (112) of a different communication network (110); 30 incorporating, in response to such a determination, its public address (312) in said request; and

À 23 a transmitter to forward (375) said request for a communication resource to said different serving communication unit (112).

5
14. The serving communication unit (152) according to Claim 13, wherein said serving communication unit (152) operates in a public land mobile network (150).
15. The serving communication unit (152) according to 10 Claim 13 or Claim 14, wherein said communication resource request is a 'create packet data protocol (PDP) context' message.
16. The serving communication unit (152) according to 15 any of preceding Claims 13 to 15, wherein said serving communication unit (152) is a serving general packet radio system (GPRS) service node (SGSN), for example supporting universal mobile telecommunication standard (UMTS) or general packet radio system (GPRS) 20 communication.
17. A gateway general packet radio system (GPRS) Service Node (GGSN) (116) adapted to receive a message, for example a GPRS Transport Protocol message, having a 25 public address of said serving communication unit (152) and configured to record said public address of said serving communication unit (152) as part of said message.
18. A communication system comprising a plurality of 30 networks having respective serving communication units that serve a plurality of subscriber units with one or more communication resources, wherein a number of said plurality of serving communication units are identified by both a public address and a private address and said

- 24 communication system supports roaming of subscriber units between different networks, wherein a subscriber unit (310) roams from a home communication network (110) to a visited communication network (150) and requests (350) a 5 communication resource from a visited serving communication unit using a general packet radio system (GPRS) transport protocol (GTP) formatted message, said communication system characterized in that: said visited serving communication unit ( 152) 10 processes said request in order to extract a serving communication unit private address and determines that said serving communication unit private address identifies a serving communication unit ( 112) of a different communication network (110), and in response to 15 such a determination, said visited serving communication unit (152) incorporates a public address (312) of said visited serving communication unit (152) in said request; in order that data can be routed to said subscriber unit via said visited serving communication unit ( 152).
19. The communication system according to Claim 18, said communication system further characterized by said subscriber unit (310) roaming from a home public land mobile network (110) of said subscriber unit (310) to a 25 visited public land mobile network (150).
20. The communication system according to Claim 18 or Claim 19, said communication system further characterized by said message from said subscriber unit (310) being a 30 'create packet data protocol (PDP) context' message transmit to a home public land mobile network (110).
21. The communication system according to any of preceding Claims 18 to 20, said communication system

- 25 further characterized by said serving communication unit being a serving general packet radio system (GPRS) service node.

5
22. The communication system according to any of preceding Claims 18 to 21, said communication system further characterized by said serving communication unit being operably coupled to a border gateway employing a network address translation function such that said 10 border gateway (BG) in said visited communication network (150) routes a subsequent message transmit from said home communication network (110) to said subscriber unit (130) and translates (385) a destination address of said subsequent message from a public address of said visited 15 serving communication unit ( 152) to a private address.
23. The communication system according to any of preceding Claims 18 to 22, wherein said communication system supports the universal mobile telecommunication 20 standard (UNITS) or general packet radio system (GPRS) communications.
24. A serving communication unit ( 152, 156) for operating in the communications system of any one of 25 claims 18 to 23.
25. A storage medium storing processor-implementable instructions for controlling a processor to carry out the method steps of any of Claims 1 to 10, or facilitate an 30 operation of the serving communication unit (152, 156) of Claims 13 to 17 or Claim 24.
26. A communication system substantially as hereinbefore described with reference to and/or as

- 26 illustrated by FIG. 1 or FIG. 2 or FIG. 3 of the accompanying drawings.
27. A method of routing information in a 5 communication system substantially as hereinbefore described with reference to and/or as illustrated by FIG. 3 of the accompanying drawings.