WO2016015783A1 - Service availability in mobile communications networks - Google Patents

Abstract

The invention relates to a method, by a gateway controller (100), to select a gateway (200) from a plurality of gateways (200a-i) for a data packet flow in a mobile communications network, the gateway being selected for applying a service to the data packet flow. The method comprises a step in which service data is received from the plurality of gateways, the service data including information which at least one service the corresponding gateway is able to apply to the data packet flow. The service data for the plurality of gateways (200a-i) is stored in a service list (70) where it is stored for each of the plurality of gateways, which at least one service the corresponding gateway is able to provide. Furthermore, data indicating an unavailability which indicate which at least one service is not available at the corresponding gateway is received from the plurality of gateways (200a-i). A request is received to select one gateway from the plurality of gateways for the data packet flow to which the service has to be applied. One of the plurality of gateways (200a-i) is selected for the data packet flow based on the service list and based on the data indicating an unavailability so that the service is applied to the data packet flow at the selected gateway.

Description

Service availability in mobile communications networks

Technical Field

The present invention relates to a method, by a gateway controller to select a gateway from a plurality of gateways for a data packet flow in a mobile communications network and to the corresponding gateway controller.

Background When an audio or video call is tried to be established, the call processing may require audio or video transcoding in case of incompatible codecs (Coder - Decoder). Different User Equipment (UEs) support different set of speech and video encoders and a gateway, being a node or a functional entity, is needed in between to invoke transcoding between the different encoding formats so that user plane connection for the speech and video will work properly on both directions.

At the border of IMS (IP Multimedia Subsystem) networks or at access site, there can be a node called Border Gateway Function (BGF) to ensure the needed security functions, such as Network and Port Translation (NAPT), topology hiding and firewall-like services are performed for the call, when needed.

Session Description Protocol (SDP) offer-answer model (RFC3264) is used to negotiate the media properties that the different entities in the call path support. SDP is carried over a Session Initiation Protocol (SIP) between the server / gateway controller and the UE and between the different server nodes in the network. Also, the SDP protocol is carried in H.248 protocol which can be used between the gateway node and the server node in the network.

In networks, the gateway controller and the gateway can be provided by different vendors or the gateway types may be different so the services provided by the gateways can differ. That is why the gateway needs to be carefully selected for the call, depending on the requested services. VoLTE (Voice over Long Term Evolution) is one field of application where BGF transcoding may occur. Figure 1 shows a scenario in which such a transcoding may be necessary. The SDP negotiation example shown in Fig. 1 involves a user equipment A with reference numeral 10, a session gateway controller, SGC 30, the gateway (border gateway function) 40 and the mobile entity UE-B 20. In step S1 , UE-A 10 transmits a SIP INVITE message to the SGC 30 which carries an SDP offer. In step S2, the SGC 30 selects a suitable user plane node, being a gateway and forwards the offer to the selected gateway 40 using the H.248 protocol including the SDP offer codecs and the SGC added codecs. In step S3, the possible media (SDP) restrictions are considered by the gateway 40 and the gateway returns the H.248 add reply message to the SGC 30, the add reply message including the supported codecs by BGF 40, in the example the codecs A, B, C, D. The BGF 40 is now aware of the media handling capabilities of UE-A, but is not aware of the UE-B's media handling capabilities. In step S4, the gateway controller, SGC 30, sends the SIP INVITE message towards UE-B 20 carrying the UE-A codecs plus the SGC added codecs for the SDP offer, thus in the example the codecs A, B, C, D. In step S5, the UE-B 20 checks the offer and returns the SDP answer based on the media parameters that it supports. In the example shown, the UE-B selected codec C on the SDP answer, the SGC 30 receiving the SIP message 183 (Session in Progress). In step S6 a modify request is transmitted towards the gateway which includes the UE-B remote address and the media codecs that the UE-B 20 supports. Now, the gateway 40 can make a decision whether the codecs are compatible or not and whether transcoding is needed for the call or not. There may be other services invoked at this point as well, depending on the UE-B's capabilities which are now known. In step S7, the gateway 40 transmits a reply message back to the SGC 30 carrying the BGF selected codec for both the UE- A site and the UE-B site. In the example shown, the BGF selected the codec B for UE-A and C for UE-B. In step S8, the SGC 30 sends the SIP 183 message (Session in Progress) towards the UE-A. Now, the UE-A and UE-B can establish a user plane connection via the BGF as symbolized by step S9.

The following problems exist for the above-identified scenario: A transcoding decision cannot be taken before the media properties of both sites, i.e. UE-A and UE-B are known. This is too late to make a re-selection of the gateway if needed since the IP address of the gateway has already been transmitted in the SDP offer in step S4 to UE-B. The problem is furthermore visible in the following use cases:

Transcoding require specific hardware (HW) or processing units, currently a pool of DSPs (Digital Signal Processors) and is bound to the load or performance of the processing units. The specific hardware can only process a limited amount of connections and on high load of the hardware, the pool of DSPs or the entity which is monitoring the DSP pool's load, will start rejecting the connections. As a result, the H.248 Modify Reply response will be sent with an error code 510 meaning insufficient resources and the call be rejected. As a result the user could not establish the call.

A processing hardware is not always reserved in a termination addition phase due to a capacity impact. The reservation of processing capacity / DSP would worsen the node's performance and it is thus not an option to always reserve processing resources even if transcoding will not be needed for the call. A number of calls can be connected without DSP resources. Thus, the DSP resources have to be dimensioned according to the actual usage. - When the operator has not enabled the transcoding feature on one BGF gateway node, but has enabled it on other nodes, the call is still rejected and the end-user cannot make the call when the gateway is selected where the transcoding feature was not enabled. This is because the transcoding decision can only be made at the Modify H.248 phase, when UE-B answer is known.

Summary

Accordingly, there is an object to avoid at least some of the above mentioned drawbacks and to improve the selection of a gateway for the application of a service.

This need is met by the features of the independent claims. Further embodiments are described in the dependent claims. According to a first aspect of the invention, a method, by a gateway controller is provided to select a gateway from a plurality of gateways for a data packet flow in a mobile communications network wherein the gateway is selected for applying a service to the data packet flow. According to the method, service data is received from the plurality of gateways, the service data including information which at least one service the corresponding gateway is able to apply to the data packet flow. The service data is stored for the plurality of gateways in a service list where it is stored for each of the plurality of gateways, which at least one service the corresponding gateway is able to provide. Furthermore, data indicating an unavailability are received from the plurality of gateways which indicate which at least one service is not available at the corresponding gateway. A request to select one gateway from the plurality of gateways is received for the data packet flow to which the service has to be applied and one of the plurality of gateways is selected for the data packet flow based on the service list and based on the data indicating an unavailability so that the service is applied to the data packet flow at the selected gateway.

With the knowledge of the services each gateway is able to provide and with the knowledge of the data indicating an unavailability indicating which service is not available in view of an unavailability, e.g. an overload situation, the gateway controller can improve the selection of the gateway and the number of calls that are lost is minimized. Furthermore, the load on already loaded gateways is decreased as the gateway controller may avoid the selection of a service at a gateway for which an unavailability such as an overload situation exists.

The invention furthermore relates to the corresponding gateway controller comprising a receiver configured to receive the service data. The gateway controller furthermore comprises a database configured to store the service data for the plurality of gateways in the service list, a control unit being configured to select the gateway taking into account the data indicating an unavailability and taking into account the service list.

According to another aspect, a method for operating a gateway is provided in a mobile communications network, the gateway applying a service to a data packet flow transmitted through the mobile communications network. According to one step, it is determined which at least one service the gateway is able to apply to the data packet flow. Furthermore, an unavailability information is determined for each of the at least one service and it is detected when an unavailability is present for one of the at least one service. Service data is generated including the determined at least one service the gateway is able to apply to the data packet flow. Furthermore, data indicating an unavailability including the information which at least one service is not available at the gateway is generated when the unavailability has been detected. The service data and the data indicating an unavailability is transmitted to the gateway controller, which selects the gateway from the plurality of gateways for the data packet flow.

With the gateway monitoring the load situation and informing the gateway controller which services the corresponding gateway is able to provide, the gateway supplies the information needed by the gateway controller to select a gateway for the data packet flow as mentioned above.

The invention furthermore relates to the gateway itself including a service monitoring unit configured to determine which at least one service the gateway is able to apply to the data packet flow. A load monitoring unit of the gateway is configured to determine an unavailability information and a control unit generates service data including the determined at least one service the gateway is able to apply to the data packet flow and to generate the data indicating an unavailability. A plurality of processing entities provide the at least one service to a data packet flow and a transmitter transmits the service and the data indicating an unavailability to the gateway controller.

The invention furthermore relates to a method for controlling an application of a service to a data packet flow by a system of a gateway and a gateway controller in which the gateway carries out the steps mentioned above and the gateway controller carries out the corresponding steps discussed above in connection with the gateway controller. According to another aspect, a computer program comprising instructions which, when executed on at least one control unit, causes the at least one control unit to carry out the method mentioned above. Furthermore, a computer program product comprising the computer program discussed above is provided in a data carrier containing the computer program.

The invention furthermore relates to a gateway controller for selecting a gateway from a plurality of gateways for a data packet flow in a mobile communications network, the gateway controller interacting with the gateway, the gateway being selected for applying a service to the data packet flow. The gateway controller comprises a receive service data module, a receive data indicating an unavailability module, an update service list module, a receive request to select gateway module and a select gateway based on updated list module. The data indicating an unavailability can include hierarchy information which indicates at which hierarchical topology level in a topology of the mobile communications network the unavailability exists. The gateway is then selected from the plurality of gateways taking into account this hierarchy information.

The data indicating an unavailability can be data indicating an unavailability resulting from an overload of the service. Further advantages and features of the invention will now be described with reference to the drawings.

Brief description of the drawings

Figure 1 is a signaling diagram illustrating a Session Description Protocol negotiation example when a call is established between two user equipments as known in the art; Figure 2 is a block diagram illustrating an embodiment of a gateway controller ;

Figure 3 is a block diagram illustrating an embodiment of a gateway ;

Figure 4 is a block diagram illustrating an embodiment of a system including a gateway and gateway controller shown in Figs. 2 and 3;

Figure 5 is a block diagram illustrating an embodiment of a system including a gateway and a gateway controller shown in Figs. 2 and 3; Figure 6 is a block diagram illustrating an embodiment of a system including a gateway and a gateway controller as shown in Figs. 2 and 3;

Figure 7 is a table illustrating data maintained by an embodiment of the gateway controller;

Figure 8 is a table illustrating an embodiment of provisioning data ; Figure 9 is a block diagram illustrating an embodiment of a system including a gateway and a gateway controller shown in Figs. 2 and 3;

Figure 10 is a block diagram illustrating an embodiment of a system including a gateway and a gateway controller shown in Figs. 2 and 3;

Figure 1 1 is a signaling diagram illustrating an exchange of signals in an embodiment of a network; Figure 12 is a block diagram illustrating an embodiment of a system, and

Figure 13 is a flow chart illustrating embodiments of method steps comprising the steps carried by a gateway controller to select a gateway for a data packet flow. Figure 14 is a block diagram illustrating an embodiment of a system.

Detailed description In the following, concepts according to embodiments of the invention will be explained in more detail by referring to the accompanying drawings. It should be understood that each of the features discussed above in the summary part and discussed herein below in the detailed description might be used in the context in which it is described. However, it may also be used in isolation or in combination with any of the other features described above and below.

As will be shown in more detail below, the invention enhances the behavior of a selection of a gateway or border gateway function, especially in such a way that the re-selection of a gateway is already possible at a step when the gateway controller receives the add reply message in step S3 discussed above in connection with Fig. 1 . The following mechanisms may be applied in this context. It is to be noted that the terms "gateway", "gateway node", "gateway function" and "border gateway function" are used interchangeably throughout the following description and that these expressions relate to the same functional entity.

The gateway internally monitors the unavailability of the provided services, e.g. the load situation of the processing entities on the node to which the gateway has access. Different user plane services consume different amounts of processing power and hardware resources may be added, deleted, locked and unlocked. Furthermore, calls are established and released. All these activity needs to be tracked, to keep track of the internal resource pool load levels. The gateway should store and update this information periodically.

As will be explained in more detail with reference to Figures 9 and 10, the gateway should indicate which levels of overload problems it can detect, e.g. whether the overload situation occurs on application level, node level or site level. Furthermore, the gateway can indicate which service it can provide, using inter alia H.248 auditing mechanisms and a H.248 service identification package where this information is placed. The gateway controller can use audit mechanisms towards the gateway and updates a service list indicating services provided by the gateway and is able to understand the meaning of the topology information, such as the actual load level and overload information occurs on a node or site level, for the service availability.

When the gateway detects that the load level of the service providing processing entities, e.g. the transcoding pool, is high, e.g. higher than a certain threshold level such as 95 percent, then the gateway can reply already to the H.248 add reply message with an error code 510, but also involving a topology identifier which may be embedded in the error text. This can be a) an alternative method to the audit/notify method above, or b) an extension method, i.e. during normal operation with enough call setups in time, the data is included in the add reply, but if there are no call setups in certain intervals, the gateway may send a notify message periodically. Additionally, the gateway may include a timer value indicating how long the expected service availability may last. The topology identifier indicates at which topology level the overload situation exists, e.g. on application level, node level or site level. The gateway controller can then select another gateway with the same service capability, e.g. transcoding capability from another node. The gateway re-selection mechanism on the gateway controller is amended and takes into account the usage of the hierarchical topology level as an additional parameter when another gateway is selected in a re-selection algorithm when one of the gateways indicates an overload situation. The gateway can be implemented having a multi-blade or multi-card structure. By way of example, if all the cards providing the same service capability within the gateway are locked, then the gateway will send a notification, e.g. a new H.248 notify request to the gateway controller telling the controller that the service list in the gateway controller should be updated and not to include the specific service anymore. When the first card or blade providing said service, e.g. the transcoding capability is unlocked, the gateway will send a new H.248 notify request to the gateway controller telling the gateway controller that the gateway can provide the service provided on the unlocked card, e.g. the transcoding capability.

Figure 2 shows a schematic view of a gateway controller 100 which selects a gateway taking into account the service capabilities and unavailability data or load information, e.g. load levels, from other available gateways to the gateway controller 100. The gateway controller 100 comprises an input/output unit 1 10 with a transmitter 1 1 1 and a receiver 1 12. The transmitter 1 1 1 symbolizes the possibility of the gateway controller 100 to transmit messages or data to other nodes and to communicate with other nodes in the network in which it is incorporated, e.g. a mobile communications network. The receiver 1 12 is provided to receive messages or data from other entities. The input/output unit 1 10 is used inter alia for communication with different gateways and for controlling user plane data handled by the gateway. A controller 130 is provided comprising one or more microcontrollers or processors which are responsible for the operation of the gateway controller 100. The control unit 130 can generate inter alia the commands that are needed to carry out the improved selection of a media gateway discussed above and below. A memory 140 is provided which may be implemented as a read-only memory, a flash read-only memory, a random access memory, a mass storage, hard disk or the like. The memory includes suitable computer programs to be executed by the control unit 130 so as to implement the functionalities relating to the gateway controller described above and also described in further detail below.

A database 120 is provided in which a service list or table 70 is stored, the service list containing information for each of the gateways controlled by the gateway controller 130 which of the services the corresponding gateway is able to provide. The service list 70 furthermore comprises information about possible overload situations at the different gateways as will be explained in further detail in connection with Fig. 7 below. In connection with Figure 3 a schematic view of a gateway 200 is shown which comprises an input/output unit 210 with a transmitter 21 1 and a receiver 212. The transmitter 21 1 is used for sending data or messages to other entities such as the gateway controller, the receiver being configured to receive data/messages from other entities.

The gateway 200 furthermore comprises a service monitoring unit 220 which determines which service the gateway is able to provide, e.g. transcoding. Transcoding is used as one example for a service provided by the gateway. Tanscoding occurs when e.g. end users 10 and 20 apply different Codecs, or when different IP formats are applied such as IPv4 and IPv6. However, it should be understood that the invention is not restricted to the provision of a transcoding service, other services, such as network and port translation, topology hiding, firewall-like services, may be provided by the gateway. Other services could be: video transcoding, access transfer gateway and ICE (Interactive Connectivity Establishment) gateway. Access transfer gateway is related to the media anchoring and transfer gateway function needed when a VoLTE client performs handover from LTE radio network to 2G/3G radio network. The service monitoring unit 220 monitors which of the services the gateway is able to apply to the user data plane. This information can be transmitted to the gateway controller 100 embedded into service data transmitted by a transmitter 21 1 . An availability or load monitoring unit 230 keeps track of each service and of the load level of each service that is relevant for the service. The load monitoring service 230 furthermore is able to determine a hierarchy information which indicates at which hierarchical topology level in a topology of the mobile communications network the overload situation exists, e.g. whether the overload situation is site specific or node specific. The load monitoring unit follows up the usage of the service loads which are reachable by the gateway, i.e. the services provided by the processing entities 241 , 241 , 243, 244 which symbolize the different processing entities which provide the different services. The different processing entities can be provided on different cards or blades. For determining the load of the services, the load monitoring unit 230 may monitor the processes which impact the service load such as the number of ongoing calls, the type of services involved, the processing power and bandwidth consumption. Furthermore, maintenance actions on the hardware cards such as locking, unlocking or adding or removing hardware cards or reconfiguring the hardware cards with different hardware pool contents influence the availability of a service or the load. The load monitoring unit 230 keeps track of the load levels of the services provided by the different processing entities 241 -244 and, if a certain threshold is met, by of example, a load of 95, 90 or 80 percent, then load monitoring unit 230 and, thus, the gateway considers the corresponding hardware pool as overloaded. The load monitoring unit 230 furthermore applies detection means whether other gateway applications can possibly reuse the same pools of hardware or not and will provide this information as part of the hierarchy information to the gateway controller such that the re-selection mechanism can be triggered at the gateway controller 100. Control unit 240 controls the operation of the gateway 200. The control unit 240 can generate the commands that are needed to carry out the procedures discussed above and below in connection with the gateway 200. A memory 250 is provided which may be a read-only memory, a flash read-only memory, a random access memory, a mass storage, a hard disk or the like and which includes a suitable program code or a computer program code to be executed by the control unit 240 so as to implement the functionalities of the gateway.

It should be understood that the structural features shown in Figs. 2 and 3 are merely schematic and that the gateway controller 100 and the gateway 200 may include further components which, for the sake of clarity, have not been illustrated. Furthermore, it should be understood that different entities shown in Figs. 2 and 3 need not be incorporated in the separation of entities as shown. By way of example, two different entities may be implemented in a single entity. Furthermore, the different entities shown in Figs. 2 and 3 may be incorporated by hardware or software or a combination of hardware and software.

A gateway may have its own processing capabilities as shown in Fig. 3, however, it is also possible that different gateways share their processing capabilities. In connection with Figs. 4-6 different examples are shown of a hardware architecture how different gateways use the processing capabilities for processing data packet flows in order to apply a service to the data packet flow.

In Figure 4 a first example is shown in which different controllers 100 control different gateways, named gateway application in the example of Figs. 4-6. In the example of Fig. 4 four different gateways are shown wherein two gateways 200a and 200b are located on one site, site A, wherein the other gateways are located at site B, the gateways 200c and 200d. The fact that gateways 200a and 200b are located on one site, whereas the other gateways 200c and 200d are located at another site, is explained with reference to Fig. 8, where a hierarchical topology level in the topology of the mobile network is shown. The lowest level in a topology is the application which, in the present context, is the process that is handling the signaling and provides the requested service, the application being the process on a blade or process in a virtual machine or processing unit which processes the user plane data in order to provide the service. The next topology level is the node, e.g. a location where several blades belonging to the same network operator are located wherein one node is operated by the operator by one interface. The next topology level is the site level meaning that all components within one site are located in a single room and have a common interface to the IP backbone. In other words, all hardware elements within one site are located in the same location. The next topology level is the province meaning that it covers all components located in a certain geographical region of the mobile communications network, the highest topology level being the country in which the mobile communications network is installed.

Referring back to Figs. 4-6 the gateway is located on one site, e.g. all components in one room have a common hardware pool which provides, for example, the services A, B, C and D. This means that gateway application 1 and gateway application 2 use the hardware pool 240a which provides certain services. In the same way, the gateway applications 200c and 200d use a second hardware pool 240b which provides the services A, B, C, D.

Another possibility is shown in Figure 5 where each of the gateway applications 200a, 200b, 200c, 200d has its own hardware pool. This means that gateway application 200a uses a first pool of processing units, namely pool 240a, whereas gateway application 200b uses the hardware pool 240b. Gateway application 200c uses hardware pool 240c and gateway application 200d uses the hardware pool 240d. Another possible implementation is shown in Figure 6 where each hardware pool 250a, 250b, 250c, 250d within one site provides certain service capabilities. In the example of Fig. 6 this means that hardware pool 250a can provide the services A and B, whereas hardware pool 250b provides the services C and D. Each gateway application 200a and 200b can access each of the hardware pools 250a and 250b. In the same way, the gateways 200c and 200d on site B can each access hardware pool 250c providing services A and B and hardware pool 250d providing services C and D. Figure 7 shows a service list or table 70 as generated and maintained by the gateway controller 100 and stored in the database 120. The service list 70 indicates as a kind of matrix per site which services are available and unavailable and further indicates a possible level of unavailability. In the example shown in Fig. 7, the gateway controller 100 controls eight gateway applications on three different sites. Each gateway, indicated in the table header with "Gw1 ", "Gw2"...."Gw8", provides one or more different services, e.g. services A to E, but some services are not available on each gateway. A service "A", "B"...."E". could in this example each represent e.g. a transcoding activity for a certain set of codecs. By way of example, service B is not available at all at gateways 3-5 on site B. Furthermore, some services are not available due to an overload situation or other maintenance actions such as locking, unlocking or adding and removing of hardware cards. In the embodiment shown in Fig. 7 the services currently unavailable are the services that are indicated in the hashed boxes, e.g. services D and E on gateways 1 and 2, and services E on gateways 3 to 5. In the following example it is assumed that the gateway controller 100 wants to allocate service A. In this case, any of the gateways 1 to 8 can be selected. However, if service E should be selected, the gateways from site C have to be selected.

The service list as indicated in Fig. 7 is generated and maintained up to date by the gateway controller 100. The gateway controller should be informed about the possible services and about the load level of the different processing entities providing the corresponding service. In the H.248 protocol there exists an audit procedure which can be used by the gateway controller to request the gateway to report its capabilities to the controller. Auditing can be ordered for H.248 packages and for the properties within the packages so the gateway controller node 100 detects which packages and properties the respective gateway supports. In addition to this mechanism, there could be another package, illustrated by Figure 8, by way of example, a gateway service information package which contains the existing levels of hierarchy such as "Country", "Province", "Site", "Node", Application, as explained above in connection with Figure 8. The actual service could also be reported within this package or, as an alternative, the existing H.248 packages could be used to convey the service information. The gateway can also inform the gateway controller 100 about its services and the corresponding load in a response message sent in response to a call setup message transmitted by the gateway controller 100 at call setup to reserve resources at the gateway for the call. In the following, Figs. 9 and 10 illustrate cases where the bottleneck or overload situation occurs either on the gateway application itself (Fig. 9) or where the bottleneck is occurring in the pool of processing units (Fig. 10). The problem of Fig. 9 can be solved within the site, whereas the problem of Fig. 10 cannot be solved within the same site.

In the example of Figure 9 , based on the system configuration of gateway applications 200a-200c and hardware pools 250a-250c of figure 6, the gateway application 200a is overloaded so that the services A, B, C, D cannot be established via this gateway. The gateway application can inform the gateway controller by answering a H.248 add reply with an error code 510, wherein the error code does not include an identifier, identifying the service cluster. Based on the information transmitted from the gateway application 200a to the gateway controller, the gateway controller can select gateway 200b within the same site as the bottleneck is not within the service pool itself.

In the example of Figure 10, based on the system configuration of gateway applications 200a-200c and hardware pools 250a-250c of figure 6, the hardware pool providing services A and B is overloaded so that the services A, B cannot be established via this hardware pool. The gateway 200a answers to the H.248 add reply message with an error code 510 including the service cluster identifier, the service cluster identifier informing the gateway controller that a hardware pool is overloaded. The gateway controller, here the upper gateway controller (controller 1 ) in the figure, can react to select gateway 200c on the other site as the bottleneck is common for the whole site A. However, the hardware 250c or 250d can still provide intermediate services.

In Figure 11 , an example of a message flow is discussed in more detail in which the selection of an overloaded gateway is avoided. Steps S1 1 and S12 relating to the INVITE message and the selection of one gateway 200a, having the requested service available, to which the SDP offer included in the SIP INVITE message is forwarded correspond to steps S1 and S2 of Fig. 1 . As can be seen from Fig. 1 1 , the gateway 200a transmits a response message with an error code 510 back to the gateway controller including the service cluster identifier which indicates the hierarchical topology level where the overload situation occurs, indicated bystep S13. Based on the received response of the step S13, the gateway controller selects another gateway.

Using the table 70, the gateway controller 100 selects the gateway having the requested service, by way of example, gateway 200c which is located on a different site. In this example it is assumed that the response message in step S13 indicated that the entire site where the gateway 200a is located is overloaded. Thus, the add request message corresponding to step S2 of Fig. 1 is sent in step S14 to the gateway on the other site wherein the response message is received in step S15. Steps S16 and S17 then correspond to steps S4 and S5 of Fig. 1 and steps S18 and S19 correspond to steps S6 and S7 of Fig. 1 with step S20 corresponding to step S8. As shown in Fig. 1 1 , the other gateway 200c is selected before the SIP INVITE message is sent to the user equipment UE-B.

The solution shown in Fig. 1 1 also saves calls for which additional services are triggered when the H.248 termination is already created since the selection mechanism would follow the above described features.

In connection with Figure 12 an example is shown in which services provided by a site are not available due to an overload. In the example shown a single controller has access to gateways 200a-200i on three different sites. In some call cases, by way of example, in emergency or priority calls, it may be considered important that the end user gets service as soon as possible. In the example shown in Fig. 12 the applications 200a-200f cannot provide the emergency service even that extra buffer is used from the resource pool which is specifically reserved for emergency calls. However, the gateway applications on site C, 200g-200i, can provide the needed hardware resources.

In one solution of the prior art the gateway controller 100 would first try to select gateway 200a, then gateway 200b, then gateway 200c, and so on, until finally it would reach gateway 200g which can provide the required service. Each of the service requests makes the node which is not available performing even more congested. Furthermore, the call setup delay would be multiplied because of the unnecessary re- selection trials in the network. In the present example each of the gateways at sites A and B can indicate in the hierarchy information that the entire site is unavailable. Thus, it would only require three H.248 add request messages to establish the emergency call. If there was a large amount of small gateways on each site, e.g. ten gateway applications per site, then this factor would even be higher.

In connection with Figure 14 another example is shown in which services provided by one site are unavailable due to an overload. Four different gateway controllers (controller 1 to 4) are provided and within site A the hardware pool, here a conference device pool 240a, can be accessed by the two media servers in site A. In site B a second hardware pool 240b is provided. In the figure it is assumed that the hardware pool 240a, the conference device pool 1 is overloaded. By way of example, the media server 1 could handle five conference participants, but not six as was requested in this example. Thus, a six-party conference service cannot be established via media server 1 . As a consequence, the conference service can also not be established via media server 2. The media server 1 answers an H.248 add reply with an error code 510 including the hierarchy information or service cluster identifier. Now, controller 1 can try to select the media server 3 on the other site as the bottleneck is common for the whole site A.

As discussed in connection with Figs. 9, 10, 1 1 , 12 and 14, the overload situation or the unavailability of the requested service can occur on different topology levels. In the examples above the service unavailability occurred within a node or a site, but the level of service unavailability can go higher in the hierarchy. If a certain service becomes unavailable on the application level only, such as an application overload, as indicated in Fig. 9, then the node level is still available in case the same service is reachable using the same node. However, if, for example, an Ethernet backbone of the node is unreachable or overloaded, then the whole node is unavailable for providing the required service and the required service would still be available on site level. If, however, the whole IP routing is unusable on the site level as the IP network may not be working, then a different site has to be selected. The gateway node has to detect the type of service that is unavailable and to indicate the topology level to the controller. Typically, gateway has supervision functions to supervise the node connectivity. It can detect that Ethernet connections are physically down, or it may have specific router supervision functions, which are for example polling periodically the router, e.g. by sending ARP (Address Resolution Protocol) messages or icmp (Internet Control Message Protocol) ping messages to the router. If no answer is received, via all redundant interfaces, the node can detect that it has lost connectivity to the site. The reporting of higher topology level may require a connectivity monitoring unit on the site, which would tell the status of connectivity to all gateways on the site.

Given the above described examples, some general findings can be identified for the gateway controller 100 and the gateway 200. As far as the gateway controller is concerned, the overload data received from a gateway can include hierarchy information according to the description with figure 8, which indicates at which hierarchical topology level in a topology of the mobile communications network the overload situations exist. The gateway selection is then carried out taking into account this hierarchy information. Furthermore, it is proposed that the different gateways provide the information about the requested services and the overload situation upon request. By way of example, a gateway controller 100 can transmit service requests to the different gateways 200 and each service request requesting from the gateway to which the request is transmitted to inform the gateway controller 100 which at least one service the gateway is able to apply to the data packet flow.

The service data, the data indicating an unavailability, the hierarchy information and/or the service request can be transmitted using the H.248 protocol. By way of example, the service request can be part of an audit procedure of the H.248 protocol. In another embodiment, the gateway controller 100 is informed during the call setup procedure, by way of example, when the gateway controller sends a H.248 add request command to reserve resources at the gateway, the gateway returning a negative response to the request as some resources may be overloaded. The response then also contains the service data, the data indicating an unavailability and the hierarchy information. Alternatively, the service data can be transmitted using a Media Gateway Control Protocol (MGCP) or a DIAMETER Protocol.

Furthermore, it is proposed that the received data indicating an unavailability include time information indicating how long the overload situation is expected to last. The time information can then be stored in the service list 70 in connection with the gateway for which the data indicating an unavailability is received, and when the timer expires, the service information of service list 70 for the corresponding gateway is updated to indicate that the overload situation is not present anymore.

Furthermore, the data indicating an unavailability can be stored in a database in the service list 70 in connection with the gateway for which the data indicating an unavailability is received.

As far as the gateway 200 itself is concerned, the service data informing the gateway controller 100 about the provided services can be transmitted to the gateway controller in response to the service request received from the gateway controller.

Furthermore, a hierarchical topology level in a topology of the mobile communications network can be determined at which the unavailability occurs and the information about the hierarchical topology level may be included into the data indicating an unavailability when the data indicating an unavailability is transmitted to the gateway controller.

If one of the services provided by a gateway 200 is provided by a plurality of different processing entities 241 -244 and if it is detected that an overload situation exists at each of the processing entities providing said one service, the service controller is informed that the gateway does not support this service.

Furthermore, it is possible that a processing load of each of the services provided by the gateway 200 is monitored and an unavailability or overload situation is detected when the processing load of the service is higher than a predefined load threshold.

The gateway 200 can be a border gateway provided at the border of an IP Multimedia Subsystem , IMS, and the data packet flow can contain voice data of a voice call transmitted via the IP Multimedia Subsystem .

The different services provided by a gateway 200 could also be implemented using a virtual machine on a cloud system. Using the invention as discussed above, the gateway application can detect the level service unavailability and inform the controller accordingly. On a virtual machine a number of processing entities are pre- reserved and the services will be provided on those processing entities. This can result in a situation that, depending on the requested services, the types of calls and their roles regarding the provided services, there may be problems on certain calls which are limiting the entire virtual machine. With the hierarchy information the gateway controller is informed which levels of the virtual machine are overloaded and can react accordingly. Furthermore, it may happen that in the virtualized example above the different nodes have a different capacity level. It may be difficult to make maintenance operations like shutting down virtual machines on live networks since the calls are spread on different hardware entities of the gateways which are defined for connecting to the same gateway controller. Furthermore, the gateways or virtual machines can have different capacities, capabilities and service profiles.

The above discussed solution can also apply to the H.248-based graceful locking procedure for a virtual machine. In the graceful locking method, the connections are closed down gracefully when the connection ends and, thus, there is no service downtime for the maintenance operation. No new connections are setup to the virtual machine so that it is possible to make the needed maintenance operations when there are no further present calls.

There exists the possibility to expand and decrease the pool of virtual machines based on the configured policy. The number of virtual machines could be monitored by a cloud orchestration system so that the virtual machines would report the load reports to an orchestrator periodically. The orchestration system could then allocate new virtual machines or reduce the number of virtual machines based on the current load of the system.

Furthermore, it is possible that new services are added to a gateway or existing services are removed. The service list 70 of supported services within the gateway can change due to several reasons. With the gateway periodically informing the gateway controller about the provided service list of services, the gateway controller can update the service list 70 accordingly. A supported service may be provided by one gateway, if the user can upload new licensed services or services are removed when the existing services which are bound to licensing constraints run out or expire. Furthermore, the service list 70 of available services within the gateway node can change as a result of other maintenance operations.

Furthermore, the locking and unlocking of cards or blades and the adding or removing of cards or blades can influence the services provided by a gateway. When a new card providing hardware capabilities is inserted or a locked card is unlocked, then there could be new services exposed to the network from the gateway. Then, the notify request message contains the new service list of services provided by the gateway. The gateway can send a notify request message to the server indicating that a certain service is taken out of service or brought back to service and the gateway controller updates the service list 70.

In a similar way the gateway can indicate when a server is no longer available, which can be caused by card removal or locking. The H.248 notify request may be used to convey this information to the gateway controller. The service list 70 may also be amended when new licenses are uploaded or when a license expires. Typically, there are three types of licenses which are applicable to the services provided by the gateway - a try-and-buy license, in this license type the license is given for a trial period and the license has to be extended after the trial period when the service should be provided. In a feature license it is indicated on the enabled/disabled level whether a certain feature is enabled or not. This feature is not limiting the capacity, but it can be regarded as a simple switch. A third type of license is a capacity license which can be given per node typically and it may be distributed to different gateway applications within the gateway node. Such a capacity license has a certain limit. There can be threshold values, sent over the operations and maintenance interface, and these threshold values are indicated when the capacity limit is close to the maximum usage level. The above scenario is described in different examples when a change in the gateway service profile occurs. The information about the updated service profile can be conveyed to the gateway controller using the above discussed H.248 notify request message carrying the service list 70 and the impacted hierarchy information. Figure 13 summarizes the steps carried out at the gateway controller to improve the selection of a gateway for user data plane of a data packet flow. In step S21 the gateway controller 100 receives service data from the gateways in which each gateway indicates which kind of service the corresponding gateway is able to provide. Furthermore, in step S22 the gateway controller receives the data indicating an unavailability, here the overload data. The overload data indicate which service is not available due to an overload situation or any other reason, the overload data furthermore indicating at which hierarchical topology level the overload occurs. As discussed above, the messages of steps S21 and S22 may not be received in the indicated order, but may be received periodically as part of an audit procedure of the H.248 protocol or as part of a H.248 add command in which the gateway controller tries to reserve resources at the gateway for a call setup. In step S23 the service list 70 stored at the gateway controller is updated with the received information so that the service controller has up-to-date service information which service is available at which gateway. When a request is received to select a gateway, such as the request indicated by step S1 1 of Fig. 1 1 (step S24), the gateway is selected taking into account the current service list, according to step S25.

As indicated above, the steps S21 to S25, especially S21 to S24, need not be carried out in the indicated order as the service data and the overload data of step S21 and S22 may also be received after step S24 as indicated in the signaling diagram of Fig. 1 1 .

The above described solution provides the advantage that it enhances the selection of a service that has to be applied to a user data plane of a data packet flow. It enhances the service application solution and less calls are lost due to the fact that the intended service is invoked and the DSP resource is only available on another node or site. The addition of new services is detected with the service data so that the network maintenance costs are lower as the network can self-adapt the service list of supported services. This service list can be used in a maintenance situation or when new licenses are installed. Although most of the examples were made in connection with a transcoding service, the invention is not limited to a certain user plane service. The user plane node/gateway detects and registers its internal architecture and can choose to trigger the site level re-selection in case the service cannot be selected on the site due to the internal architecture. Advantageously, the load on already loaded nodes which can no longer provide the wanted services is decreased as the re-selection takes into account the load level and avoids already loaded nodes.

Multiple gateway controllers 100 may control the same gateway. As these gateway controllers apply the method elucidated, each gateway controller has an overview of the current status of the gateway. As each gateway controller in general has one or more other gateways in its service list, an imbalance of load levels among several gateways is prevented.

Application of the method presented prevents imbalance of the load levels that would have been caused due by prior art methods wherein the SGC 100 tries to select all gateway resources from the same site in order to save transmission resources in the IP backbone.

It is regarded another advantage that the call setup time can be much lower when there is load on a network since the gateway selection is more intelligent. This can be important prioritized services, such as priority calls or emergency calls.

Claims

Claims

1 . A method, by a gateway controller (100), to select a gateway (200) from a plurality of gateways (200a-i) for a data packet flow in a mobile communications network, the gateway being selected for applying a service to the data packet flow, the method comprising the steps of:

- receiving, from the plurality of gateways, service data, the service data including information which at least one service the corresponding gateway is able to apply to the data packet flow,

- storing the service data for the plurality of gateways (200a-i) in a service list (70) where it is stored for each of the plurality of gateways, which at least one service the corresponding gateway is able to provide,

- receiving, from the plurality of gateways (200a-i), data indicating an unavailability which indicate which at least one service is not available at the corresponding gateway,

- receiving a request to select one gateway from the plurality of gateways for the data packet flow to which the service has to be applied,

- selecting one of the plurality of gateways (200a-i) for the data packet flow based on the service list and based on the data indicating an unavailability so that the service is applied to the data packet flow at the selected gateway.

2. The method according to claim 1 , wherein the data indicating an unavailability includes hierarchy information which indicates at which hierarchical topology level in a topology of the mobile communications network an unavailability exists, wherein said one of the plurality of gateways is selected taking into account the hierarchy information.

3. The method according to claim 1 or 2, wherein the gateway controller (100) transmits service requests to the plurality of gateways (200a-i), each service request requesting from the gateway to which the request is transmitted to inform the gateway controller (100) which at least one service the gateway is able to apply to the data packet flow.

4. The method according to any of the preceding claims, wherein at least one of the service data, the data indicating an unavailability, hierarchy information, the service request is transmitted in accordance with a H.248 protocol.

5. The method according to claim 4, wherein the service request is part of an audit procedure of the H.248 protocol.

6. The method according to claim 4 or 5, further comprising the step of transmitting a H.248 add command to at least one of the gateways in which a reservation of resources for the data packet flow at the at least one gateway is requested, wherein at least one of the data indicating an unavailability, the service data, and the hierarchy information is received from the at least one gateway in response to the transmitted H.248 add command.

7. The method according to any of the preceding claims, wherein the received data indicating an unavailability includes timer information indicating how long the unavailability situation is expected to last, the timer information being stored in the service list in connection with the gateway for which the data indicating an unavailability is received.

8. The method according to any of the preceding claims wherein the data indicating an unavailability is stored in the database in connection with the gateway for which the data indicating an unavailability is received.

9. A gateway controller (100) configured to select a gateway from a plurality of gateways for a data packet flow in a mobile communications network, the gateway being selected for applying a service to the data packet flow, the gateway controller comprising:

- a receiver (1 12) configured to receive, from the plurality of gateways (200a-i), service data, the service data including information which at least one service the corresponding gateway is able to apply to the data packet flow, and data indicating an unavailability which indicate which at least one service is not available at the corresponding gateway,

- a database (120) configured to store the service data for the plurality of gateways in a service list (70) where it is stored for each of the plurality of gateways (200a-i), which at least one service the corresponding gateway is able to provide,

- a control unit (130) configured to select the gateway from the plurality of gateways, wherein, when the receiver (1 12) receives a request to select one of the gateways from the plurality of gateways for the data packet flow to which the service has to be applied, the control unit (130) is configured to select one of the plurality of gateways for the data packet flow based on the service list and based on the data indicating an unavailability so that the service is applied to the data packet flow at the selected gateway.

10. The gateway controller (100) according to claim 9, further comprising a transmitter (1 1 1 ) configured to transmit service requests to the plurality of gateways (200a-i), each service request requesting from the gateway to which the request is transmitted to inform the gateway controller (100) which at least one service the gateway is able to apply to the data packet flow.

1 1 . A method, for operating a gateway (200) in a mobile communications network, the gateway applying a service to the data packet flow transmitted through the mobile communications network, the method comprising the steps of:

- determining which at least one service the gateway (200) is able to apply to the data packet flow,

- determining an unavailability information for each of the at least one service and detecting when an unavailability is present for one of the at least one service,

- generating service data including the determined at least one service the gateway is able to apply to the data packet flow

- generating data indicating an unavailability including the information which at least one service is not available at the gateway, when the unavailability has been detected, - transmitting the service data and the data indicating an unavailability to a gateway controller (100) which selects the gateway from a plurality of gateways for a data packet flow.

12. The method according to claim 1 1 , wherein the service data are transmitted to the gateway controller (100) in response to a service request received from the gateway controller.

13. The method according to claim 1 1 or 12, further comprising the steps of

- determining at which hierarchical topology level in a topology of the mobile communications network the unavailability occurs,

- including information about the hierarchical topology level into the data indicating an unavailability when the data indicating an unavailability are transmitted to the gateway controller.

14. The method according to any of claims 1 1 to 13, wherein one of the at least one service is provided in the gateway by a plurality of different processing entities (241 - 244), wherein when it is detected that an unavailability exists at each of the different processing entities (241 -244) providing said one service, the service controller is informed that the gateway does not support said one service.

15. The method according to any of claims 1 1 to 14, wherein a processing load of each of the services provided by the gateway is monitored, wherein an unavailability is detected when the processing load of a service is higher than a predefined load threshold.

16. The method according to any of claims 1 1 to 15, further comprising the step of determining gateway data indicating which other gateway accesses processing entities used by the gateway to provide said at least one service, wherein the gateway data is added to the data indicating an unavailability transmitted to the gateway controller.

17. A gateway (200) configured to apply a service to a data packet flow transmitted through a mobile communications network, the gateway comprising:

- a service monitoring unit (220) configured to determine which at least one service the gateway is able to apply to the data packet flow,

- an unavailability monitoring unit (230) configured to determine an unavailability information for each of the at least one service and determining when an unavailability is present for one of the at least one service,

- a control unit (240) configured to generate service data including the determined at least one service the gateway is able to apply to the data packet flow and to generate data indicating an unavailability including the information which at least one service is not available at the gateway,

- a plurality of processing entities (241 -244) which provide the at least one service to the data packet flow,

- a transmitter (21 1 ) configured to transmit the service data and the data indicating an unavailability to a gateway controller which selects the gateway from a plurality of gateways for a data packet flow.

18. The gateway (200) according to claim 17, wherein the gateway is a border gateway at a border of an IP Multimedia Subsystem, IMS.

19. The gateway according to claim 17 or 18, wherein the data packet flow contains voice data of a voice call transmitted via an IP Multimedia Subsystem.

20. A method for controlling an application of a service to a data packet flow by a gateway (200) in a mobile communications network, the gateway being selected by a gateway controller (100) among a plurality of gateways, the comprising the steps of at the gateway (200):

- determining which at least one service the gateway is able to apply to the data packet flow,

- determining an unavailability information for each of the at least one service and detecting when an unavailability is present for one of the at least one service, - generating service data including the determined at least one service the gateway is able to apply to the data packet flow

- generating data indicating an unavailability including the information which at least one service is not available at the gateway when the unavailability has been detected,

- transmitting the service data and the data indicating an unavailability to a gateway controller which selects the gateway from a plurality of gateways for a data packet flow,

at the gateway controller (100):

- receiving the service data from the plurality of gateways (200a-i),

- storing the service data for the plurality of gateways (200a-i) in a service list (70) where it is stored for each of the plurality of gateways, which at least one service the corresponding gateway is able to provide,

- receiving, from the plurality of gateways (200a-i), the data indicating an unavailability, and

- selecting one of the plurality of gateways for the data packet flow based on the service list (70) and based on the data indicating an unavailability so that the service is applied to the data packet flow at the selected gateway.

21 . A system configured to control an application of a service to a data packet flow by a gateway (200) in a mobile communications network, the gateway (200) being selected by a gateway controller (100) among a plurality of gateways, wherein the gateway is a gateway as claimed in claim 17 and the gateway controller is a gateway controller as claimed in claim 9 or 10.

22. A gateway controller (100) for selecting a gateway (200) from a plurality of gateways (200a-i) for a data packet flow in a mobile communications network, the gateway controller interacting with the gateway, the gateway being selected for applying a service to the data packet flow, the gateway controller comprising:

- a receive service data module (S21 );

- a receive data indicating an unavailability module (S22);

- an update service list module (S23);

- a receive request to select gateway module (S24), and

- a select gateway based on updated list module S25).