A communication system

FIELD OF THE INVENTION

The present invention relates to a security server for use in a telecommunications network, a network processing element, a telecommunications network and a method of performing a security check on a message incoming to a telecommunications network.

BACKGROUND OF THE INVENTION

It is known to provide a wireless telecommunications network across which two users of mobile equipment can communicate, or a mobile user can communicate with a fixed location user by transfer of a signal from the wireless network to a land line. One known type of wireless communications network is the 3^rd Generation Partnership Projects (3GPP) system which is currently being brought into use around the world. This network is known as the Universal Mobile Telecommunications System (UMTS) and one advantage that it has over previous wireless network standards is that it allows far faster rates of data transfer using a packet-switched (core) network (PS-CN) in addition to voice transfer over a circuit-switched (core) network (CS-CN). The PS-CN can connect to the Internet and the CS-CN can connect to the Public Switched Telephony Network (PSTN) and the Integrated Digital Services Network (ISDN).

In practice, the CS-CN functionality is achieved via a subsystem called the IP Multimedia Subsystem (IMS) in the PS-CN. The IMS can connect to an IP based network such as the Internet to provide services such as Voice over IP. The signalling protocol used between user equipment (UE) such as mobile telephones and the IMS and between components of the IMS is the Session Initiation Protocol (SIP). This protocol has user registration (e.g. location and communication capability), addressing and routing capabilities.

One important set of components within an IMS network is the Call Session Control Functions (CSCF). These perform a server service in that they process signals and control a wireless user's session, as well as performing an address translation function and handling of subscriber profiles. If a user is in the home network, the network is accessed via the Serving-CSCF (S-CSCF), and this server provides session control and other services for the user. If the user is roaming, the local network in the roaming location is accessed via a Proxy-CSCF (P-CSCF) which provides local control and services for the user as well as being in contact with the user's S-CSCF. The S-CSCF and if necessary the P-CSCF also perform a billing function. It is usual to have a number of S-CSCFs within an IMS network.

A further type of CSCF is an Interrogation CSCF (I-CSCF). The I-CSCF is the first point of contact within a home network for an access by a visiting user. It is arranged to communicate with the Home Subscriber Server (HSS), which holds subscriber account information and subscriber location information. The I-CSCF is set up to perform load balancing within the S-CSCFs using information provided by the HSS. Since it provides a single point of entry into the network for users from other networks, it is often used as a means to prevent operators of other networks from knowing the specific structure of the IMS network.

A problem that arises with the type of network described above is that messages arriving at the I-CSCF from outside the IMS network are not necessarily from a reliable source. Since the I-CSCF works in conjunction with the HSS it is a simple matter to determine whether an incoming message is from a user identified as having details held in the HSS or who is a subscriber to another network with which the IMS network has roaming agreements. If the message is not from such a user then access to the network can be restricted, for example by not providing any services which need to be paid for by the user unless payment is taken up front. However, a specific problem arises with messages that apparently do originate from a network subscriber listed in the HSS.

When a user attempts to access the network from outside the network, a message is sent to the network which includes an identification of the requesting user. This identification is checked by the I-CSCF in conjunction with the HSS as explained above. Many' user identifications are publicly-known, so that an unauthorised user can adopt a publicly-known identification when making an access request to the network. If this publicly-known identification belongs to a subscriber of the network, even though it is determined that a user of that identity is a subscriber to the network, the access is in fact not being requested by that subscriber. Consequently the unauthorised user gains access to the network and furthermore gains access to the account of the subscriber whose identification is being adopted. Such an unauthorised user could thus use the subscriber's account and run up a significant bill without the subscriber being aware of this, perhaps until the subscriber's next monthly bill is received.

In a similar manner an unauthorised user could use an identity of a subscriber to another network who would be permitted to use the network in view of a roaming agreement between the two networks.

It would be desirable to provide a telecommunications network in which the likelihood of unauthorised access using publicly-known subscriber identifications is minimised.

<patcit id="pcit0001" dnum="WO0211469A"><text>WO0211469</text></patcit> describes a technique for authenticating a user to a server using SIP messages.

ETSI STANDARDS, EUROPAN TELECOMMUNIATIONS STANDARDS

INSTITUTE, SOPHIA-ANTIPO, FR, <nplcit id="ncit0001" npl-type="s"><text>vol. 3-SA3, no, V520, December 2002</text></nplcit> relates to IP related security in the UMTS core work.

<patcit id="pcit0002" dnum="US20020116463A"><text>US2002/0116463</text></patcit> describes a filter mechanism for unwanted e-mail messages.

<patcit id="pcit0003" dnum="WO0165806A"><text>WO01/65806</text></patcit> describes a system for avoiding rerouting in a computer network during secure remote access.

<patcit id="pcit0004" dnum="US5884025A"><text>US5884025</text></patcit> describes a system for Screening data packets transmitted between a private network and a public network.

<patcit id="pcit0005" dnum="EP1343342A"><text>EP1343342</text></patcit> constitutes prior art according to Article 54(3) EPC and describes a method for client authentication and integrity protection of communicated data between an electronic user apparatus and server when a standardized synchronization protocol such as SyncML-D&lt;/DS is used.

<patcit id="pcit0006" dnum="WO03007489A"><text>WO03/007489</text></patcit> describes a system for transmitting CDMA call set-up parameters including authentication parameters through an IP-based infrastructure to an authenticating entity.

The present invention provides a security server for use in a telecommunications network according to claim 1, a method according to claim 15 and a computer program according to claim 25.

Embodiments of the invention are defined in claims 2 to 14 and 16 to 24.

An embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings in which: <ul id="ul0001" list-style="none"><li><figref idref="f0001">Figure 1</figref> shows schematically an IMS network and access thereto from outside the network;</li><li><figref idref="f0002">Figures 2a and 2b</figref> show flow charts in accordance with a first embodiment of the invention;</li><li><figref idref="f0003">Figures 3a and 3b</figref> show flow charts in accordance with a second embodiment of the invention;</li><li><figref idref="f0004">Figures 4a and 4b</figref> show flow charts in accordance with a third embodiment of the invention.</li><li><figref idref="f0005">Figure 5</figref> shows two security domains including Za and Zb interfaces.</li></ul>

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Before describing embodiments of the invention, an explanation will firstly be given regarding the Za and Zb interfaces that can exist between networks and within networks respectively. This explanation is taken from the 3GPP TS 33.210 V6.0.0 (2002-12) Technical Specification, Release 6. This specification covers Technical Specification Group Services and System Aspects; 3G Security; Network Domain Security and IP network layer security. <figref idref="f0005">Figure 5</figref> shows two security domains and the Za and Zb interfaces between entities of these domains.

The interfaces are defined for protection of native IP based protocols:

Za-interface (SEG-SEG)

The Za-interface covers all NDS/IP (Network Domain Security/Internet Protocol) traffic between security domains. The SEGs (Security Gateways) use IKE (Internet Key Exchange) to negotiate, establish and maintain a secure ESP (Encapsulating Security Payload) tunnel between them. Subject to roaming agreements, the inter-SEG tunnels would normally be available at all times, but they can also be established as needed. ESP shall be used with both encryption and authentication/integrity, but an authentication/integrity only mode is allowed. The tunnel is subsequently used for forwarding NDS/IP traffic between security domain A and security domain B.

One SEG can be dedicated to only serve a certain subset of all roaming partners. This will limit the number of SAs and tunnels that need to be maintained.

All security domains compliant with this specification shall operate the Za-interface.

Zb-interface (NE-SEG / NE-NE)

The Zb-interface is located between SEGs and NEs and between NEs within the same security domain. The Zb-interface is optional for implementation. If implemented, it shall implement ESP+IKE.

On the Zb-interface, ESP shall always be used with authentication/integrity protection. The use of encryption is optional. The ESP Security Association shall be used for all control plane traffic that needs security protection.

Whether the Security Association is established when needed or a priori is for the security domain operator to decide. The Security Association is subsequently used for exchange of NDS/IP traffic between the NEs.

The security policy established over the Za-interface is subject to roaming agreements. This differs from the security policy enforced over the Zb-interface, which is unilaterally decided by the security domain operator.

Referring firstly to <figref idref="f0001">figure 1</figref>, there is shown an IMS network 1 of a UMTS system. Within the IMS network 1 there is an I-CSCF 2, connected to each of a P-CSCF 4, an S-CSCF 6 and an HSS 8. The HSS 8 and the S-CSCF 6 are also connected to one another. In practice there would be more than one S-CSCF but just one is shown for convenience. These components are all provided with suitable software for performing various functions and can be set up for the particular needs of a network by suitable programming. Some functions may be hardware-based.

In between each of the connections, is shown represented by a dotted oval, a Zb interface 10. This interface is shown dotted because it is optional. This is because network providers may not want to use the Zb interface within their network.

The Zb interface, when provided, is a security interface which exists between each pair of connected entities within the IMS network 1. The Zb interface is only present within the IMS network because it is only applied to entities within the same security domain i.e. the IMS network 1. The purpose of the Zb interface is to provide security for messages being sent within the IMS network. Since the IMS network 1 is within a UMTS network the security protocol is provided by an Encapsulating Security Payload (ESP) protocol. This provides a check for the data integrity of messages travelling over it and a data origin authenticity check. The Zb interface also makes use of security keys using the Internet Key Exchange (IKE) protocol.

As a result of the presence of the Zb interface, any entity within the IMS network 1 that receives a message from another entity within the IMS network 1 knows whether that message has come over the Zb interface (and hence is security cleared) or not.

It can be seen from <figref idref="f0001">figure 1</figref> that the I-CSCF 2 is the entry point into the IMS network 1 for messages originating from outside the network. The ICSCF 2 is connected to a Security Gateway (SEG) 12 outside the network through which messages can enter the IMS network 1 over a security checked path. This path includes a Za interface 14, represented by a solid oval. The Za interface is similar in operation to the Zb interface but is compulsory in the current specification of the UMTS network. The SEG 12 provides a gateway function between the IMS network 1 and the Za interface 14. Any messages coming over the Za interface through the SEG 12 are known by the I-CSCF 2 to have been security checked.

A second network 16 is also shown in <figref idref="f0001">figure 1</figref> and is shown to have its own SEG 18. The network 16 is a network with which the IMS network 1 has a roaming agreement. Consequently, any traffic between the network 16 and the IMS network 1 is sent over the Za interface so that the receiving network knows that messages received in this way have been security cleared and are from authorised and authenticated users. A roaming user having a mobile telephone 20 and who is a subscriber to the IMS network 1 is shown as being in a suitable location for accessing the network 16.

A second mobile telephone 22 is also shown in <figref idref="f0001">figure 1</figref>. This mobile 22 is attempting to access the IMS network 1 directly i.e. not through another network such as the network 16.

It should be appreciated that the relative sizes of the networks and their relative locations and those of the mobile telephones are only shown as example representations in <figref idref="f0001">figure 1</figref>. In practice, the IMS network 1 would have Za interface connections with a number of networks, but just one network (16) is shown for convenience. Furthermore, the invention could apply to other types of network than a UMTS network.

In operation the user having a mobile telephone 20 can access the network 16 by virtue of the roaming agreement. The mobile telephone 20 will request a handover to the network 16 when it roams into the area covered by the network and will be provided with a connection, in the manner known in the art. Before being provided with a connection, authorisation and authentication procedures will be performed. This procedure includes a password-based check or some other suitable verification that the person purporting to be the user is in fact the user and not someone else using their identity. If the authorisation and authentication check is successful, a secure channel is set up for use. Thus, if the user of the mobile telephone 20 then wishes to use his home network services, any requests are sent over the Za interface as SIP messages so that the IMS network 1 knows that they are genuine requests.

Such a message contains a header identifying the sender, called a P-Asserted-Identity header. The format of this header if the sender is a user with a publicly-known user identification is : &lt;sip:user1_public1 @home1 .net&gt;

If the mobile telephone 22 sends a request directly to the IMS network 1, the I-CSCF 2 knows that the request has not come over the Za interface and that consequently no security check has been carried out on the request. The P-Asserted-Identity header of such a system can not be trusted as genuine because a false user may have adopted another user's public identity. In prior art systems, such a message would nevertheless be passed directly to the S-CSCF 6 for processing and the S-CSCF would not know that the message was potentially a problem. In a prior art system with a Zb interface, such a message would be passed over the Zb interface, so, again the S-CSCF would proceed to process it. Consequently, if the user of the mobile telephone 22 is in fact simply using the identity of a subscriber to the IMS network 1, that user would have been allowed access to the subscriber's account.

The following embodiments of the invention describe solutions to this problem.

In the first embodiment of the invention, the Zb interface is not in use within the IMS network 1 so this can not be used as a means of security clearing a message.

The first embodiment is represented by the steps in <figref idref="f0002">figures 2a and 2b</figref>. Turning firstly to <figref idref="f0002">figure 2a</figref>, at the start of the process (30) a SIP message is received at the I-CSCF. This message could either have been received over the Za interface or directly from outside the network. In step 32, the I-CSCF 2 determines which of these two alternatives is the case.

If the answer is no (i.e. message not received via Za interface), the I-CSCF 2 proceeds to step 34 at which a modification is made to the P-Asserted-Identity header of the message. In this embodiment a parameter is added to the header to indicate that the message has not been through security clearance. The example header shown above is therefore modified to have the following format: &lt;sip:user1_public1@home1.net&gt;;screening=no

The I-CSCF then proceeds to step 36 in which the message, with the modified header is forwarded to the S-CSCF 6.

If the answer to step 32 is yes, the message has come over the Za interface, and the I-CSCF proceeds directly to step 36 and forwards the message to the S-CSCF 6 without making any modifications to the message.

At step 38 the message arrives at the S-CSCF bearing an indication of its authenticity. In other words, if the message arrives with a normal P-Asserted-Identity header, the S-CSCF 6 knows that it has been through a security check. If the message arrives with a modified P-Asserted-Identity header, the S-CSCF 6 knows that it has not been through a security check.

The subsequent functions of the S-CSCF 6 are represented in <figref idref="f0002">figure 2a</figref>. The S-CSCF 6 reads the message in step 40. In step 42 it determines whether or not the P-Asserted-Identity header of the message includes a parameter added by the I-CSCF 2. If the answer is no, the S-CSCF 6 proceeds to step 44 in which the message is processed. If the answer is yes, the S-CSCF 6 proceeds to step 46 in which it performs security checks such as authorisation and authentication checks on the message. If these checks show that the message is from a genuine subscriber, the message can subsequently be processed as normal. If the message turns out not to be from a genuine subscriber, the S-CSCF can decide not to process the message as normal but instead to only partially process it, for example by not allowing use of services which must be paid for. The S-CSCF could decide not to process the message at all. Hence unauthorised access to the account of the subscriber having the P-Asserted-Identity carried in the message is avoided.

In a second embodiment of the invention, the Zb interface is again not in use within the IMS network 1 so this can again not be used as a means of security clearing a message.

The second embodiment is represented by the steps in <figref idref="f0003">figures 3a and 3b</figref>. Turning firstly to <figref idref="f0003">figure 3a</figref>, at the start of the process (50) a SIP message is received at the I-CSCF. This message could either have been received over the Za interface or directly from outside the network. In step 52, the I-CSCF 2 determines which of these two alternatives is the case.

If the answer is no (i.e. message not received via Za interface), the I-CSCF 2 proceeds to step 54 at which it checks to see whether a P-Asserted-Identity header is present and if so, a modification is made to the P-Asserted-Identity header of the message. In this embodiment the P-Asserted-Identity header is removed to indicate that the message has not been through security clearance. It would be possible to remove just a part of the header if desired.

The I-CSCF then proceeds to step 56 in which the message, without the header is forwarded to the S-CSCF 6.

If the answer to step 52 is yes, the message has come over the Za interface, the I-CSCF proceeds directly to step 56 and forwards the message to the S-CSCF 6 without making any modifications to the message.

At step 58 the message arrives at the S-CSCF bearing an indication of its authenticity. In other words, if the message arrives with a normal P-Asserted-Identity header, the S-CSCF 6 knows that it has been through a security check-If the message arrives without a P-Asserted-Identity header, the S-CSCF 6 knows that it does not have a reliable originator.

The subsequent functions of the S-CSCF 6 are represented in <figref idref="f0003">figure 3b</figref>. The S-CSCF 6 reads the message in step 60. In step 62 it determines whether or not the P-Asserted-Identity header of the message has been removed by the I-CSCF 2. If the answer is no, the S-CSCF 6 proceeds to step 64 in which the message Is processed. If the answer is yes, the S-CSCF 6 proceeds to step 66 in which it performs security checks such as authorisation and authentication checks on the message. If these checks show that the message is from a genuine subscriber, the message can subsequently be processed. If the message turns out not to be from a genuine subscriber, the S-CSCF can decide not to process the message as normal but instead to only partially process it, for example by not allowing use of services which must be paid for. The S-CSCF could decide not to process the message at all. Hence unauthorised access to the account of the subscriber having the P-Asserted-Identity carried in the message is avoided.

In a third embodiment of the invention, the Zb interface is in use within the IMS network 1 and can be used for the purposes of security clearance.

The third embodiment is represented by the steps in <figref idref="f0004">figures 4a and 4b</figref>. Turning firstly to <figref idref="f0004">figure 4a</figref>, at the start of the process (70) a SIP message is received at the I-CSCF. This message could either have been received over the Za interface or dirfectly from outside the network, In step 72, the I-CSCF 2 determines which of these two alternatives is the case. The answer to step 72 determines the manner in which the message is forwarded to the S-CSCF 6.

If the answer is no (i.e. message not received via Za interface), the I-CSCF 2 proceeds to step 74 at which the message is forwarded to the S-CSCF 6. The message is forwarded directly i.e. not over the Zb interface.

If the answer to step 72 is yes, the message has come over the Za interface, the I-CSCF proceeds to step 76 and forwards the message to the S-CSCF 6 over the Zb interface. The Zb interface may carry out further internal security checks on the message, even though it has already been checked by the network 16 from where the message was sent.

At step 78 the message arrives at the S-CSCF in a manner which indicates its authenticity. In other words, if the message arrives over the Zb interface, the S-CSCF 6 knows that It has been through a security check. If the message arrives directly i.e. not over the Zb interface, the S-CSCF 6 knows that it has not been through a security check.

The subsequent functions of the S-CSCF 6 are represented in <figref idref="f0004">figure 4b</figref>. The S-CSCF 6 reads the message in step 80. In step 82 it determines whether or not the message has been sent over the Zb interface by the I-CSCF 2. If the answer is yes, the S-CSCF 6 proceeds to step 84 in which the message is processed. If the answer is no, the S-CSCF 6 proceeds to step 86 in which it performs security checks such as authorisation and authentication checks on the message. If these checks show that the message is from a genuine subscriber, the message can subsequently be processed. If the message turns out not to be from a genuine subscriber, the S-CSCF can decide not to process the message as normal but instead to only partially process it, for example by not allowing use of services which must be paid for. The S-CSCF could decide not to process the message at all. Hence unauthorised access to the account of the subscriber having the P-Asserted-Identity carried in the message is avoided.

Thus it can be understood that the invention provides several solutions useable alone or in combination to the problem of false users attempting to access the IMS network 1 using a publicly-known identification of a genuine subscriber to the IMS network. The examples of the invention would work equally well if the user of the mobile phone 20 was a subscriber to the network 16 and was genuinely allowed to use the IMS network 1 by virtue of a roaming agreement. In other words, the invention provides a way of distinguishing between genuine and non-genuine users who are using publicly-available identifications.

The invention is not limited to the particular network components described above, nor to the SIP protocol. The mobile telephone 22 may not in reality be attempting to access the IMS network 1 directly but may do so via other telecommunications entities. Other examples falling within the scope of the invention can be envisaged, for example, it is not necessary to modify the header of a non-secured SIP message, but a different part of the message could be modified. A non-secured message could be tagged in some other way than by addition of a parameter. Although possibly less convenient, it would be possible to modify messages that have come over the Za interface rather than those that have not.

The invention could also be applied to messages having headers other than P-Asserted-Identity headers that could have been adopted by a false user.

Examples of the invention could therefore incorporate modification of such a different type of header, or deletion or partial deletion of such a header.