TRANSIT NETWORK
This invention relates to the field of communications networks. More specifically, but not exclusively, the invention relates to a method of transferring traffic over a transit network between switch nodes in a communications network such as, for example, a mobile telecommunications network.
Transit networks for transferring traffic, for example voice or data traffic, between switch nodes in a communications network are known in the art. An example of such a communications network is a mobile telecommunications network. Figure 1 is a schematic outline of a mobile telecommunications network according to the GSM standard. The network comprises a number of transit nodes, which handle traffic which neither originates nor terminates in this node. In a network according to the prior art, traffic from a first switch node is routed to a second remote switch node via several transit nodes or a network of transit nodes, the so-called transit network. One important requirement for a transit network is resilience, i.e. to provide alternative routes between an originating and a terminating node in the case of congestion or failure of certain nodes or communications links. The alternative routes should ideally have no common infrastructure, i.e. neither common transit nodes nor common physical communications links in order to ensure a fully and effectively operational network in the case that certain links or nodes fail. In order to meet this resilience requirement usually
a plurality of links are provided between the origination switch node and the terminating switch node via different intermediate nodes and different communications links of the network, as is illustrated in Figure 1.
The number of transit nodes needed in a typical communications network like for example a mobile telecommunications network is large.
Thus, the resilience requirement results in a very complex network as the number of links required proliferates exponentially with the number of nodes of such a transit network. A consequence is a very complex routing procedure. In order to exploit the capacity of a network and to avoid congestion it is important to ensure that the traffic load between the communications links and also between the transit nodes is as much as possible evenly distributed.
Both the complex routing procedures and the difficulties of providing an evenly distributed traffic load are fundamental problems of transit networks according to the prior art.
Another problem of a transit network according to the prior art is that a large number of ports of the transit nodes are used by connections to other transit nodes because of the large number of physical links required between the transit nodes. Another problem of a transit network according to the prior art is its scalability. This is especially important because the traffic load in many communications networks like for example mobile telecommunications networks grow very quickly and it is thus important to be able to adopt a
transit network to the required enhanced capacity. Because of the large number of communications links required between the transit nodes, an expansion of the transit network is very complex.
Routing may be simplified by defining different sub-domains of a communications network. In patent application WO 01/01640 a method for creating routes is described by using a root and one or more leaf networks.
However, the method suggested does not address the other problems of transit networks as discussed above.
It is therefore an object of the present invention to overcome the disadvantages described above and to provide an efficient, resilient and scalable transit network which supports a simplified routing and load-sharing procedure.
According to an aspect of the present invention, there is provided a transit network for transferring traffic between switch nodes in a communications network, said transit network comprising a plurality of transit nodes wherein direct communications links are provided between said switch nodes and said transit nodes and wherein no communications links are provided between said transit nodes.
In this way an easily scalable and resilient transit network is provided with a relatively small number of transit nodes. Moreover, the ports of the transit nodes are not deployed with physical links to other transit nodes and can thus be used for communications links to the switch nodes of the communications network. Routing via alternative routes which have no
common infrastructure, i.e. no common transit nodes and no common communications links, is thus easily available.
According to another aspect of the present invention, there is provided a transit network comprising a plurality of transit nodes for transferring traffic in a communications network between switch nodes, wherein said transit network provides a plurality of symmetric routes via a single transit node between every pair of switch nodes served by said transit network and wherein each of said routes between a pair of switch nodes is via a different transit node. In this way a highly resilient and symmetric transit network is provided which allows for a simplified routing procedure even if a large number of communications links are needed to deploy the communications network. Generally, a transit network with the proposed architecture needs less transit nodes than a transit network according to the prior art. In addition, load sharing can be easily optimised.
According to another aspect of the present invention, there is provided a communications network comprising a plurality of transit networks, wherein each transit network serves one or more regions of said communications network, said regions each comprising a plurality of switch nodes, wherein at least two of said transit networks are networks of a first, lower level for providing communication within a region of said communications network, wherein at least one of said transit networks is a network of a second higher level for providing communication between different regions of said
communications network, and wherein at least one of said transit networks is a transit network according to the present invention.
In this way it is ensured that the transit network works efficiently in a communications network with a large number of transit nodes. If for example a large geographical area is served by the communications network the region areas are preferably geographical region.
A communication between two switch nodes within such a region is then transferred by a regional transit node. This ensures that the traffic between two neighbouring switch nodes is served by a transit node within the same region rather than a transit node in a very remote area. In this way most of the physical connections can be kept short.
According to another aspect of the present invention, there is provided a method of transferring traffic between switch nodes in a communications network, whereby a route out of a plurality of alternative routes is provided between any pair of switch nodes and each of said plurality of routes between a pair of switch nodes is via a different single transit node.
In this way routing is performed in a highly resilient and symmetric manner, and thus a simplified and more efficient routing procedure is provided. According to another aspect of the present invention, there is provided a method of routing traffic from an originating switch node to a terminating switch node in a communications network according to the present invention, whereby routing is performed by transferring the traffic from said originating
switch node either to a transit node of said first level transit network or to a transit node of said second level transit network. In this way the routing is also simplified in such a hierarchically organised transit network.
Preferably, alternative routing is performed in the originating switch node by transferring the traffic to an alternative transit node. If routing is done via a transit network according to the present invention, it is ensured that no common infrastructure is used in the alternative routes. Thus alternative routing and load-sharing can easily be achieved or optimised and can easily be adapted to changed network conditions, for example in the case of link or transit node failures.
Further aspects and advantages of the invention will be appreciated from the following description and accompanying drawings, wherein:
Figure 1 is a schematic outline of a GSM mobile telecommunications network according to the prior art; Figure 2 is a schematic diagram of a communications network according to one embodiment of the present invention;
Figures 3 a and 3b are examples of a fully meshed transit network with four and six transit nodes, respectively, according to the prior art;
Figure 4 is a schematic outline of a meshed transit network connected to local exchange nodes according to the prior art; and
Figure 5 is a schematic outline of a transit network connected to local exchanged according to another embodiment of the present invention.
In Figure 1 a schematic outline of a mobile telecommunications network according to the GSM standard is shown. A Mobile Switching Centre (MSC) is connected via communication links to a number of Base Station Controllers (BSCs) 4. The BSCs are dispersed geographically across areas served by the Mobile Switching Centre 2. Each BSC 4 controls one or more Base Transceiver Stations (BTSs) 6 located remote from, and connected by further communication links to, the BSC 4. Each BTS 6 transmits radio signals to, and receives signals from, mobile stations 8 which are in an area served by that BTS 6. The area is referred to as a "cell". A GSM network is provided with a large number of such cells, which are ideally continuous to provide continuous coverage over the whole network territory.
The mobile switching centre is provided with a Home Location
Register (HLR) 12 which is a database storing subscriber data. The mobile
Switching Centre 2 is also provided with a Visitor Location Register (VLR) 14 which is a database temporarily storing subscriber data for mobile stations which are active in the area served by the Mobile Switching Centre 2.
A Mobile Switching Centre 2 is also connected via communication links to other mobile switching centres in the remainder of the mobile communications system 10, which also includes other GSM networks and to a Public Switched Telephone Network (PSTN), which is not illustrated.
A direct link 30 may be provided between two MSCs (for example MSCs 2 and 20) for economic reasons if the traffic load between the two switches is particularly high. In addition, MSC 2 is connected to transit
switching centres (TSC) 22 and 24 via communication links 32 and 34 in order to provide communication links to other MSCs. MSC 26 may for example be reached from MSC 2 via TSC 24 using the communication links 34 and 36. In addition, communication links between two transit switches are provided, for example link 38 between TSC 22 and TSC 24. The communications system schematically outlined in Figure 1 is a so-called meshed core network, wherein communication links are provided between the different MSCs, between the MSCs and the TSCs and also between the TSCs. In Figure 3 two so-called fully meshed transit networks are shown. In Figure 3a) a transit network consisting of four TSC nodes is shown. Six physical links are needed in order to provide a fully meshed network, i.e. to connect every possible pair of nodes. In Figure 3b) a transit network of six nodes is depicted. In this case already 15 links are needed in order to provide a fully meshed network. Generally, in a fully meshed network of n nodes, n(n-l)x/2 physical links are needed to provide all possible connections.
The number of TSCs needed to deploy a typical mobile telecommunications network is large. Moreover, it is clear from the two examples discussed above with reference to Figure 3 that the number of links needed to connect all nodes in such a typical meshed network proliferates very fast. A fundamental problem of meshed transit networks is therefore that a large number of ports of the TSCs are immediately deployed by connecting the TSCs to each other, thus resulting in a reduced number of ports being available to connect external nodes, such as the MSCs.
In the following a mobile communications network with a non-meshed transit architecture according to one embodiment of the present invention is described. In Figure 2 an example of such a network is shown. As can be seen from Figure 2, the network is divided into three different regions, 110, 120 and 130. These regions are preferably geographical regions in order to ensure that the physical connections within such a region are short. However, also non-geographical regions can be chosen. In each region there are three transit switching nodes (TSC) provided, the so-called regional TSCs. For simplicity only one or two MSCs are depicted in each of the regions shown in Figure 2.
Each MSC in a particular region is connected to the three regional TSCs via physical links . Each of the MSCs 111 and 112 of region 110 is for example connected to the regional TSCs 113, 114 and 115. As can be seen from Figure 2, the architecture of the transit network is totally symmetric. Thus the MSCs in one region may use any of the three regional TSCs to transit calls to other MSCs in the same region. If for example a call requires communication between the originating MSC 111 and the terminating switch node 112 within the same region 110, then the call can be routed via any of the three transit switching nodes 113, 114 or 115. As the architecture of the network is totally symmetric, none of the TSCs or the possible paths between the two switch nodes is preferred. In the embodiment described each of the TSCs offers exactly the same capacity and allows access to the same set of destinations as the other TSCs in the same region. In order to ensure that the
network provides resilience, a minimum number of three transit nodes have been provided in each region. In this way each MSC can be reached by three different routes which do not have any physical links in common. In order to share the traffic load between all the TSCs in one region, as for example TSCs 113, 114 and 115 of region 110, a proportional bidding facility is provided in each of the MSCs. A proportional bidding facility may be used to control the amount of traffic which is directed to the various TSCs. The traffic may be distributed in predefined proportions to all outgoing routes. Thus a proportional bidding facility may be used to control the traffic load in the TSCs in a flexible way. In the embodiment described where all the TSCs and the links provide the same capacity, the load is shared substantially equally between the TSCs with help of a proportional bidding facility. However, a network according to the present invention may also be deployed without such a proportional bidding facility. But using a proportional bidding facility is a convenient way to organise load sharing of traffic in a flexible way.
If a chosen link from a MSC to a TSC is congested, the traffic is re-directed via one of the remaining routes to another TSC in the same region by removing the proportion of the first chosen link and re-defining the proportions of the remaining links such that again the load is shared equally between the non-congested links.
In some proportional bidding facilities only whole percentages can be allocated to each sub-destination. In this case the split is chosen which allows the most even distribution between the nodes. For the three sub-destinations
in the embodiment described above the most even split would thus be 33% for two TSCs and 34% for the third. A small bias towards one destination results. In order to alleviate this bias, the MSCs within one region may each select a different TSC as the one with the highest proportion. In this way it is ensured that it is not always the same TSC is favoured and the traffic load is more evenly distributed.
The other regions 120 and 130 are each deployed with three TSCs in a similar manner to region 110. In addition to the regional TSCs, four national TSCs are provided. If a call is for example routed from a first MSC 111 in a first region 110 to a second MSC 121 in a second region 120, the call is directed via one of the national TSCs 140 to 143. In the embodiment shown in Figure 2 every MSC is connected to all three regional TSCs and to all four national TSCs. Again the network architecture is totally symmetric and proportional bidding facilities are provided in each MSC to share the traffic load between the three regional and four national TSCs.
In a communications network with a non-meshed transit architecture the alternate routing, i.e. the selection of one of a plurality of alternative routes as it is for example implemented by a proportional bidding facility, occurs solely in the edge nodes. Thus, the edge connections are maximised compared to fully or partially meshed transit networks.
With the architecture described, a regional transit network will be used to route between MSC within a region and a national transit network will be used to route traffic between the regions. In the embodiment described, there
are no links between the TSCs, neither between the regional or national TSCs within a transit network nor between TSCs of different transit networks.
Routing and load sharing is much simpler in such a non-meshed transit network compared to a meshed transit network. Although the number of links needed in a non-meshed environment might be greater compared to a fully meshed network, the operation of a non-meshed transit network is much simpler. This is mostly because of the symmetric architecture of such a non- meshed network.
Reference is now made to Figure 4. A fully meshed transit network according to the prior art and four local exchanges (LE) 201 to 204 are shown.
In order to ensure resilience in the example shown in Figure 4 each LE is connected to three different TSCs. In addition, each possible pair of TSCs is connected via communications links. In this way an alternative route may be provided if a chosen route is congested or currently not working. A connection from LE 201 to LE 204 may for example be provided via TSC 213 or TSC 212. In each of these connections only one TSC and two physical links are involved. Alternatively, the connection can be provided via two TSCs on seven different paths, i.e. via TSC 211 and 214, 213 and 214, 213 and 212, 212 and 213, 212 and 214, 211 and 213 or 211 and 212. Although there are nine possible paths from LE 201 to LE 204 via one or two TSCs, only three of these do not share any common infrastructure. Thus it is difficult in a network according to the prior art to provide resilience. Moreover, from this simplified example it is clear that routing may become
very complicated, especially if many more than four LEs and accordingly also more transit nodes are involved. Even more importantly, it is very difficult in a network of such a complex structure to ensure that the traffic load is evenly distributed amongst the TSCs and also amongst the physical connections. In Figure 5 a non-meshed transit network comprising four TSCs, 311 to 314, is shown which connects four LEs, 301 to 304. In this case each LE is connected to every TSC. In order to transfer traffic from LE 301 to LE 304 four different paths via the four different TSCs 311 to 314 are available. In all four possible paths only one TSC is needed and the number of physical links involved is two. Because the architecture of this network is totally symmetric all four paths are equally preferable. All four possible paths share no common infrastructure. It is thus much easier to provide a resilient network with the proposed non-meshed architecture according to the present invention. It is now easy to share the traffic load between the four TSC and thus also between the different physical links between the LEs and the TSCs by using proportional bidding facilities in the LEs.
As there are no links between the TSCs in a non-meshed transit network, there can only be one possible destination for routing in the TSC and there is no possibility for a transit node to use alternative routing. The onus of providing alternative routing in order to meet the resilience requirements is thus placed on the originating switch node (i.e. a MSC). If the link from the TSC to the terminating switch node is congested or fails, the TSC will hand back the call control to the originating node for alternative routing. The
originating switch node will then route the call to one of the remaining TSCs instead. Thus the alternative routing can be implemented easily at the MSC via the proportional bidding facility by adjusting the routing proportions to the different TSCs. Thus the network with the non-meshed transit network architecture can be made resilient easily and can also be made more resilient than a network according to the prior art. In addition, only few transit nodes have to be deployed compared to the fully meshed transit network architecture.
In a more complex scenario where a network is deployed with a larger number of local exchanges spread over a large geographical area a network architecture as in the embodiment shown in Figure 2 is advantageous. When a call is to be routed from an originating MSC, for example MSC 111, the MSC 111 distinguishes three possibilities: in the first case the terminating MSC is the same as originating MSC 111, and thus no routing to another MSC is needed. If the destination of the call is in the same region as the originating MSC 111, for example if the terminating switch node is MSC 112, the MSC 111 distributes the call via the proportional bidding facility to one of the regional TSCs 113, 114 or 115. If, on the other hand, the terminating MSC is in another region, for example region 120 or 130, the MSC 111 distributes the call to one of the national TSCs 140 to 143. The routing is again performed via the proportional bidding facility to ensure that the traffic load is evenly distributed. The selected TSC then routes the call to the terminating switch node in region 120 or 130. In this way the routing is very
much simplified compared to a network with a meshed architecture. The MSC in the embodiment described only needs to distinguish the case that no routing is needed and between the two different routing possibilities, i.e. via a regional or a national TSC. According to the embodiments described above, the ingress to a communications network with a transit network according to the present invention from another communications network (as for example a PSTN or other GSM networks) is provided in the MSC layer, i.e. in one or more of the MSCs. This allows to route communications originating from another network in the same manner and according to the same principles as described above. In this way the routing and load sharing of traffic originating from other networks may be implemented easily and in a harmonic way.
However, ingress from other networks may alternatively be, at least partly, in the transit layer, for example if the ingress has been in the transit layer before and the ingress gateways are used to form part of the new transit network according to the present invention.
A further advantage of a communications network of a non-meshed architecture is that circular routing is inherently avoided. In a meshed network architecture it is difficult to maximise the chance of routing a call under congestion conditions. Referring again to Figure 4, this is now illustrated in the example that a connection is to be provided between LE 201 and LE 204. In the list of possible connections between those two local exchange nodes as given above, the connection may occur over TSCs 212 and
213 or over 213 and 212. In a situation where all other connections are congested or fail, it may thus happen that the call is routed continuously forward and backward between TSCs 212 and 213. Such circular routing can cause major problems and can cause further congestion of the network. There are of course mechanisms known in the art to protect against this circular routing but they are complicated and often, especially in larger implementations, they are not sophisticated enough to allow exploitation of the largest number of possible routes. In a network with a non-meshed transit network circular routing is inherently avoided and routing, even in the case of congestion, is well defined and simple.
Another major advantage of a non-meshed network is its scalability. This is especially important because of traffic load in a mobile communications system grows quickly and often even optimistic forecasts for growth of mobile telephony have been outstripped by the actual demand. As already described above with reference to Figure 3, in a meshed network the total number of connections needed proliferate exponentially. Therefore any expansion of the transit network is very complex. In contrast, in a non- meshed transit network the capacity can easily be enhanced by simply adding one or more additional TSCs. The proportions of the proportional bidding facilities in the MSCs can be adjusted easily to route the selected amount of traffic to the new node and to adjust the proportions of the other nodes. In addition, in the proposed symmetric architecture a transit node can be easily replaced by a new node, which may have a larger capacity than the old node.
Again, the proportional bidding facilities can be adjusted easily to any demanded changes in the network architecture such that the traffic load is adjusted to the capacities of the individual transit nodes and communications links. Another advantage of a transit network with a non-meshed architecture is that the required overall capacity of all links is smaller than it would be for a meshed network handling the same amount of traffic.
Generally, the total number of links required in a non-meshed transit network according to the present invention is greater than for a meshed network according to the prior art. It is noted that in the example described above with reference to Figures 4 and 5, the total number of links is less for the non-meshed network (Figure 5) than in the case of the meshed network (Figure 4). However, this is generally not true for larger networks and/or communications networks using multiple transit networks. Although the number of links in a non-meshed network is greater, the required combined capacity of the links is lower than for a meshed network. This is due to the fact that no links are required between the transit nodes. In this sense a lower overall capacity is required in the case of a network with non-meshed transit network compared to a meshed network in order to handle the same amount of traffic.
The proposed architecture is generally flexible and changes can easily be incorporated and adjusted to changed demands. In the embodiments described above, no direct communication links are provided between the
MSCs. Thus any call has to be routed via one of the transit nodes. However, additional direct routes between MSC may be implemented for economical reasons between some pairs of MSCs where sufficient traffic exists between two MSCs. If the routes are individually identified in the according MSCs, they may be provided without impact to the overall architecture and routing procedure. Such direct routes may help to reduce the traffic load in the transit network.
If direct routes between MSCs are provided, the routes via the transit nodes may be used as alternative routes to the direct routes in the case of congestion. In this way, direct routes can be designed such that they are highly utilised and thus very effective.
Whilst in the aforementioned embodiments a communications network with one or two hierarchical levels (i.e. regional and national transit nodes) are described, it is to be appreciated that also more than two levels may be provided.
Whilst in the aforementioned embodiments it is described that the traffic load is shared evenly between the TSCs of one region, it is appreciated that alternatively any other proportion may be chosen, for example if the capacity of the TSCs are not the same. Whilst in the aforementioned embodiments a communications network with a combination of non-meshed regional and national transit networks is described, it is appreciated that also a combination of meshed and non- meshed transit networks may be used. A communications network according
to the present invention may for example consist of meshed regional networks and a non-meshed national transit network. Moreover, in a network according to the present invention different transit networks may be combined on the same level. A network may, for example, comprise a fully or partially meshed transit network in a first region and a non-meshed transit network in a second region.
Whilst in the aforementioned embodiments mobile telecommunications networks according to the GSM standard are described, it is appreciated that also other mobile communications networks, for example third generation telecommunications networks or fixed line communications networks may be deployed according to the present invention.
Whilst in the aforementioned embodiments voice call communications are described, it is appreciated that also other communications like for example data communications may be transferred between switch nodes according to the present invention.
The communications network according to the present invention may for example be a circuit switched network or a packet switched network.
It is to be understood that the embodiments described above are preferred embodiments only. Namely, various features may be omitted, modified or substituted by equivalents without departing from the scope of the present invention, which is defined in the accompanying claims.