GB2404826A - Packet router which re-routes packet to an alternative output port when the primary output port buffer is overloaded - Google Patents

Abstract

A method of re-routing data in a data network (500) comprises the steps of receiving one or more data packets and identifying (515) a congested route (235) from a plurality of selectable routes at a router (410) in the data network based substantially on queue length per route. The router selects (520), in response to the identification, an alternative, preferably less-congested, route (255) and re-routes (525) the one or more data packets to the alternative route. The router may additonally segment the re-routed data and direct the re-routed data to edge routers in preference to core routers. In this manner, in a multi-route data network, a data packet may be delivered to a destination node more efficiently than in known networks. Indeed, the data packet(s) will encounter fewer congested routes when the respective routers have been adapted in accordance with the aforementioned inventive concepts.

Description

2404826

- 1 -

RE-ROUTING IN A DATA COMMUNICATION NETWORK Field of the Invention

5 This invention relates to data flow in communication and/or computer networks. The invention is applicable to, but not limited to, providing an improved delivery quality of service when re-routing information through such networks.

10

Background of the Invention

The Internet is becoming more and more popular, with access to the Internet being provided via computer 15 networks and communication networks. Such networks can be vast in size with a multitude of communication units and devices re-routing communications from source nodes to destination nodes. The standard communication mechanism that communication units use to access the 20 Internet is the well-known Internet Protocol (IP) version 4 and version 6.

Recently, users have increasingly wanted to access the Internet whilst on the move, via their mobile 25 communication device(s). These devices are termed mobile nodes in IP-parlance. Different types of mobile nodes may be employed for this purpose, for example, a portable personal computer (PC), a mobile telephone or a personal digital assistant (PDA) with wireless communication 30 capability. Furthermore, mobile users are accessing the Internet via different types of fixed or wireless access networks, for example a cellular radio communication network, such as a Universal Mobile Telecommunication

- 2 -

System (UMTS) network, a HiperLAN/2 or IEEE 802.11b local area network, a Bluetooth™ local communication system, or fixed accesses such as the Ethernet, and so on.

5 Within IPv6 networks, host devices and client devices are allocated addresses comprising a hundred and twenty-eight bits to identify the device to other devices within, or external to, the network. The one hundred and twenty-eight bit addresses are known as Internet Protocol 10 version 6 Addresses (IPv6 addresses).

A method employed in some IP networks is for routers to analyse IP data packets received from client devices. Every data packet transmitted through the network 15 includes a source IP address and a destination IP

address. The source/destination addresses within these data packets are extracted by intermediate communication devices, such as routers, and used to update route information in delivering data packets to particular 20 devices.

Referring now to FIG. 1, a data communication network that employs a known IP routing mechanism is illustrated. In this regard, the data communication network comprises 25 a transmitting node 160 that is operating in a micro-

mobility (local-mobility) computer/ communication domain. The micro-mobility computer/ communication domain includes a gateway 120 between nodes in the local domain and, say, the Internet 110. For example, let us assume 30 the transmitting node is a portable computer that is able to communicate with the Internet 110 by wirelessly coupling to any of a number of access routers 140-150. Let us further assume that the destination node for a

- 3 -

communication is an application server node 115 that is connected to the Internet 110.

The access routers are typically coupled to many other 5 routers to provide a wide variety of route paths to deliver data packets to (or request data packets from) destination node 115 via the Internet 110. It is known that the router functionality is preferably placed as far as possible towards the edge of the local mobility domain 10 in order to ensure correct propagation of routes.

The micro-mobility domain includes further routers 130, 132, 134 to link the access routers 140-150 to the gateway 120. When the transmitting mobile node 160 has 15 logged on to access router 150, it is able to access applications from application server 115, via a route 180 that encompasses the Internet 110, the gateway 120, and any of intermediate routers 130-134.

20 A known problem with routing data across such a data communication network is that a quality of service in delivering the data packet to its destination node cannot be guaranteed. Further, there is no QoS guarantee for any of the intermediate links (hops) between routes 25 across the data communication network.

However, some systems are being developed that add a quality of service (QoS) provision to the Internet by attempting to control endpoints. These provisions may be 30 based on end-to-end call/session flows or piecemeal hop-by-hop approaches. In this context, individual specific virtual links may provide a guaranteed QoS, for example by use of bandwidth reservation, amongst other known

- 4 -

techniques. However, within the Internet there are many routes that a data packet may take between a source node and a destination node. Therefore, it is typically impossible to guarantee a QoS performance across a whole 5 data network, as the QoS on each and every hop, on each and every route, would need to be guaranteed. In some cases, the optimum data packet route may not be the shortest due to congestion at specific nodes.

10 Current router QoS techniques do not address this problem at all well. Current router QoS techniques include:

(i) 'Resource reservation Protocol' (RSVP): Where RSVP sets up a hop-by-hop end-to-end (ETE) link, operated 15 between a client node and a server node. RSVP is known to be very inefficient in terms of holding links and operates in a completely contrary manner to the original principles of routing.

20 (ii) 'Multi-Protocol Label Switching' (MPLS):

Sets up links of differing quality, and then multiplexes data into the respective pipe by ETE negotiated QoS for that data stream or data packet. This is usually operated within the network and not between edge nodes 25 and the core routers or clients/servers and the edge routers. In this context, an edge router is a low capacity router at the edge of the network, usually with a large fan out and low capacity, in contrast to a core or enterprise router that is a low port count router with 30 massive capacity.

(iii) 'Differentiated Services' (DiffServ): Which forwards data packets based on a priority tag located

- 5 -

within the header of each data packet. Otherwise the process queues the data packets. If the queues overflow, the process then discards data packets. This technique is used primarily between the Client/Server and the Edge 5 router. However, it is known that DiffServ can be used within the network core as well.

Often MPLS and Diffserv are used together as the mature way to support ETE QoS over Internet Protocol Networks, 10 where Diffserv is applied at the edges and MPLS is applied in the Core network.

Current techniques for improving QoS in re-routing are therefore focussed on improving existing micro 15 techniques, as manufacturers are focused on improving the re-routing performance between the relatively few routes within their micro-domains. Micro techniques are universally used and propose that routers run standard QoS techniques under normal operation. However, they 20 introduce a new approach when the router becomes congested. Under such congested conditions it is normal to start to discard data packets at the router.

In summary, the inventors of the present invention have 25 recognised that such an approach is clearly inefficient. Thus, there exists a need to provide an improved mechanism for routing data in a communication or computing network, wherein the aforementioned problems of known mechanisms may be alleviated.

30

- 6 -

Statement of Invention

In accordance with a first aspect of the present invention, there is provided a method of re-routing data 5 in a data network, as claimed in Claim 1.

In accordance with a second aspect of the present invention, there is provided a data network router, as claimed in Claim 15.

10

In accordance with a third aspect of the present invention, there is provided a data communication network, as claimed in Claim 21.

15 In accordance with a fourth aspect of the present invention, there is provided a storage medium, as claimed in Claim 22.

In accordance with a fifth aspect of the present 20 invention, there is provided an apparatus, as claimed in Claim 23.

In accordance with an sixth aspect of the present invention, there is provided a communication unit, as 25 claimed in Claim 24.

In accordance with an seventh aspect of the present invention, there is provided a data communication network, as claimed in Claim 25.

30

Further aspects and advantageous features of the present invention are defined in the dependent Claims.

- 7 -

In summary, the preferred embodiment of the present invention describes a mechanism to improve re-routing of data packets in a multi-route data network under heavy load congestion conditions. In particular, a router is 5 configured to identify a congested route of a plurality of selectable routes and, in response to the identification, select an alternative less-congested route.

10 Preferably, if the router employs a data packet discarding process, the number of data packets or portions of data packets being discarded can be minimised as the respective load congestion levels are monitored. Furthermore, for those data packets or portions of data

15 packets that are discarded, they can be re-routed via identified less-congested routes to a destination node or an intermediate node on the preferred original route.

Brief Description of the Drawings

20

FIG. 1 illustrates a simplified block diagram of a known computer/communication mobile network and mechanism for routing data packets between a source node and a destination node.

25

Exemplary embodiments of the present invention will now be described, with reference to the accompanying drawings, in which:

30 FIG. 2 illustrates a simplified block diagram of a computer/communication data network adapted to implement a preferred embodiment of the present invention;

- 8 -

FIG. 3 illustrates a simplified block diagram of a computer/communication data network indicating the rerouting process according to a preferred embodiment of the present invention;

5

FIG. 4 illustrates a simplified block diagram of a router adapted according to a preferred embodiment of the present invention; and

10 FIG. 5 illustrates a flowchart of a process implemented by a router according to a preferred embodiment of the present invention.

Description of Preferred Embodiments

15

A preferred embodiment of the present invention proposes a mechanism that considers data packet queuing in data networks as a macro (wide-area)-technique rather than a micro (local-area)-technique, particularly within

20 Internet Protocol (IP)-based data networks. In this manner, by considering, say, the Internet/Intranet as a 'partial' queue within a wide-area data network context, a quality of service (QoS) of re-routing data packets across the whole network can be improved.

25

In particular, in accordance with a preferred embodiment of the present invention, it is proposed that the router identifies a congested route, for example based on monitoring of queue lengths per route of a number of

30 selectable routes, and does not discard data packets under extreme local congestion. Instead, the router reroutes data packets, or a portion of those data packets that were to be discarded, to the nearest hop router,

- 9 -

which itself may or may not be less congested. Advantageously, allowing for the time-to-live (TTL) parameter of the IP protocol suite, minimal extra congestion is likely to be caused at the next targeted 5 hop router as, if it is also under congestion, it may perform the same re-routing operation. In this manner, data packets are propagated through the network in the most delivery-efficient way, such that they are re-combined at (or before) the destination node.

10

Referring now to FIG. 2, a data network 200, adapted to incorporate the preferred embodiment of the present invention, is shown. The data network 200 illustrates a client terminal 205 attempting to communicate to a Server 15 260 via a number of intermediate communication devices and networks/sub-networks including through a router (herein termed 'router under consideration') 220. In reality, the data network 200 may comprise a large number of intermediate interconnected routers with one 20 intermediate router being shown for clarity purposes only. The data network includes two edge routers 210, 245, located at the extremities of the data network 200, to provide the final hop to/from the client terminal 205 and/or to/from the server 260 respectively.

25

The router under consideration 220 that employs the preferred embodiment of the present invention comprises a main input buffer 230 and three output ports 215, 235, 250 containing buffers that queue data to be output from 30 the router 220. In practice, the router may have many more output ports than those shown.

- 10 -

In order to illustrate the new router functionality, let us assume a scenario where the client 205 is sending a data packet to the server 260 and that the main input buffer 230 of the router 220 is not full. Let us further 5 assume that the preferred output port 235 of the router under consideration 220 has a congested output queue.

In accordance with the preferred embodiment of the present invention, the router 220 has been adapted by 10 introducing an output queue polling function 225 to perform a data packet re-route operation within the router 220. Notably, the output queue polling function 225 determines that the selected output port (port-2) 235 to route the data packet via sub-network 240 is 15 congested, i.e. congestion is determined based on queue length per route. In response to this determination, the output queue polling function 225 decides to re-route the data packet (and any subsequent new data packets whilst the congested state remains) to an alternative nominated 20 and (hopefully) less-congested output port.

In the illustrated embodiment, the output queue polling function 225 re-routes the client's data packet through output port-3 250 and via a second sub-network 255 to the 25 server 260. Advantageously, although the finally selected (nominated) route was not a preferred route from a network routing perspective, the selected route has a better chance of being an optimal route from a data delivery perspective. In this manner, the selection of a 30 less-congested route improves the system's QoS.

A preferred operation of the output queue polling function 225 is to maintain a data 'queue' look-up table

- 11 -

containing current queue performance statistics for all of the router's output ports. The output queue data is regularly updated in the look-up table, for example following polling the .respective output ports for their 5 buffer loading status. In this manner, the output queue polling function 225 is cognisant of potential route congestion problems and is able to respond proactively to reduce such congestion problems.

10 Advantageously, as the output queue polling function 225 is aware of the buffer loading status on all of the router's output ports, it is able to select or nominate a particular route based on the destination address and the perceived congestion or loading of certain routes via 15 particular output ports.

A skilled artisan would recognise that the number of elements shown in the data network of FIG. 2 is limited for clarity purposes only. Furthermore, it is within the 20 contemplation of the invention that other network configurations and inter-coupling of elements can be used and that the illustrated configuration is merely a preferred simplified network configuration.

25 The preferred embodiment of the present invention is further explained with regard to the IP routing diagram 300 of FIG. 3. The routers in the data network of FIG. 3 employ the known technique of discarding data packets under heavy congestion. However, this known technique 30 has been improved by utilising the inventive concepts herein described.

- 12 -

In FIG. 3, two core routers '1.x' 305 and '2.x' 310 each have reporting level edge routers '1.1' 315, '1.2' 320 and '2.1' 325, '2.2' 330. Each of these edge routers 315, 320, 325, 330 has a local area network (LAN) 5 connected user terminal hosts 335, 340, 345, 350, which generate network load. Under normal conditions, no data packets received and routed by these routers are discarded. Furthermore, under normal conditions, QoS controls operate well at each edge router 315, 320, 325, 10 330 under light to medium load.

However, under a heavy traffic load per router, for example as indicated in the case of edge router '1.2' 320, the edge router 320 is congested (i.e. its status is 15 'BUSY'). At this stage, in known data packet networks, router '1.2' 320 would start discarding packets.

In accordance with a preferred embodiment of the present invention, the router 320 maintains a data queue look-up 20 table and therefore has local knowledge that its output queue towards router '1.x' 305 is full, or close to full. The router 320 also knows that its output queue towards router '1.1' 315 has spare capacity. Therefore, instead of discarding low priority datagrams for data packets 25 being routed towards router '1.x' 305, the output queue polling function of the router selectively removes less packets than normally discarded. The number of data packets removed is arranged to be just below the discard eligibility threshold. Advantageously, the output queue 30 polling function routes some or all of these removed data packets to the nearest next hop router via an alternative route, in this case edge router '1.1' 315. Edge router

- 13 -

xl.l' 315 may then route the data packets to router '1.x' 305 or other intermediate routers (not shown).

It is envisaged that process queue length is used to 5 assess whether a system is overloaded or not. Dependent upon how overloaded the queue is, above its overload threshold value, it is proposed that the percentage overload level is used to drive a level of segmentation. It is envisaged that any of the known segmentation 10 mechanisms may be used in this regard. The parameters for this algorithm would be preferably determined based on:

(i) Benchmarking;

(ii) Past history for the network type; and 15 (iii) The predominant type of data carried.

It is noted that when a router suffers congestion it will generally perform several functions. First, it will start to look at any QoS fields that identify whether 20 packets are a high or a low priority, when there is an active QoS scheme at the router. It will also examine the Type of Service (TOS) field, which is a rudimentary IP level of service indicator. Then the router will tag any packets as 'DISCARD' if they are not indicated as a 25 high priority and/or a high service level.

Alternatively, the router may unilaterally 'DISCARD' the packets directly.

Thus, a further envisaged optimisation mechanism is that 30 each router may locally assess whether re-directing its own previously discarded and tagged frames has helped to reduce network congestion. This may be achieved, for example, using Internet Control Messaging Protocol (ICMP)

- 14 -

messaging to supply the relevant information to control the modification of the router congestion scheme. In response, the router may modify the algorithm controlling the discard redirect parameter accordingly. In this 5 regard, if a router drops data packets due to ignorance of a potentially less congested and alternative route option, then it is envisaged that it may actually be introducing more congestion to the network as a whole, as it will in turn stimulate extra re-tries. Therefore, it 10 is advantageous to instead re-route, or segment and reroute, if such an alternative and less-congested route can be identified. Alternatively, those IP packets tagged as 'DISCARD ELIGIBLE' may be stored and forwarded within the Time-To-Live (TTL) timeframe of each IP 15 segment.

Referring now to FIG. 4, a router 410 is illustrated that has been adapted in accordance with the preferred embodiment of the present invention. The router 410 20 includes a signal processing function 420 comprising a router processing function 440. The signal processing function 420 is operably coupled to a receiving function 425 for receiving data packets from remote data units and a memory element 4 30. The memory element 430 comprises a 25 router table 450.

In accordance with a preferred embodiment of the present invention, the router processing function 440 has been adapted to include an output queue polling function 455. 30 In addition, router table 450 has been adapted to include a look-up table 458 that stores output buffer load information for the respective output ports of the router 410.

- 15 -

The router 410 communicates to other devices in the network via an input port 460 and an output port 470. The router 410 uses the router (address) table 450 to 5 keep track of assignments of addresses to communication nodes, as well as approving and denying requests.

In accordance with a preferred embodiment of the present invention, the receiving function receives at least one 10 first IP data packet and determines that a preferred output port is congested. In response to such a determination, the signal processing function re-routes the IP data packet, or at least a part of the IP data packet, to an alternative output port. Notably, the 15 router 410 is able to select an alternative route by accessing its continuously updated router table 450. The continuously updated router table 450 comprises information identifying the load condition of output buffers (and therefore subsequent routers) on a 20 particular path as well as the load condition of one or more alternative output buffers (router paths).

Referring now to FIG. 5, a flowchart 500 illustrates a preferred operation of the router. Following the process 25 starting in step 505, the router monitors the data load, i.e. the data queues, at the router's output buffers/ ports and the routers main input buffer, as shown in step 510. Preferably, the monitoring process includes regularly or intermittently polling the respective 30 buffers.

A determination is therefore continuously made to identify when a route becomes congested, in step 515.

- 16 -

This is preferably based substantially on queue length per route. In this regard, preferably the router identifies when a buffer of a port associated with the route becomes full. When the router identifies that a 5 particular buffer is at, or is approaching, full capacity, the router internally makes a determination as to whether one or more alternative routes could be used in preference to the congested route, in step 520. The router then re-routs new data packets to the alternative 10 (less congested) port, in step 525.

Advantageously, the router continues to monitor (poll) the load status of the output buffers and the main buffer, in step 530. As the router identifies that the 15 congestion is clearing, in step 535, then the router removes the re-route instruction from the look-up (router) table.

Preferably, the proposed mechanism is used during heavily 20 congested periods, in addition to the queuing schemes that are normally implemented at each router node. In this context, the mechanism complements the following during such congested periods:

(i) When used with MPLS: re-route the MPLS

25 link or re-route as a plain datagram with no MPLS. This reduces requested MPLS QoS, but only for a potentially limited time, which may be more acceptable than a complete discard operation.

(ii) When used with Diffserv: re-route Diffserv 30 tagged frames to a corresponding less-loaded edge router that is able to forward frames. This would be performed using the standard path assessment frames at an ICMP level between routers. This assessment is available as

- 17 -

routes change normally during re-route requests when routers are added, removed and /or enhanced in the network.

(iii) When used with RSVP: it is envisaged that 5 the proposed mechanism can be used if the path set up avoids congested router nodes at the time of the RSVP set up. In this case, the solution is limited to use only during the period of RSVP set up, and would only provide congestion avoidance at the time of set-up.

10

It is to be appreciated that the arrangement and specific details of interfaces, edge routers and core routers, etc. in the above embodiments are merely examples, and the invention should not be construed as being limited to

15 these examples. The invention is viewed as capable of being applied to other types of data

(computer/communication) networks, sub-networks and/or protocols thereof. Although the inventive concepts of the present invention find particular applicability in IP

20 data networks, it is within the contemplation of the invention that the inventive concepts can be applied in any traffic network where signals can be routed via a variety of different routes between a source node and a destination node.

25

The invention, or at least embodiments thereof, tend to provide the following advantages, either singly or in combination:

30 (i) In a multi-route data network, a data packet may be delivered to a destination node much more efficiently than in known data networks. Indeed, the data packet will be routed via fewer congested routes (or

- 18 -

routers) when the respective routers have been adapted in accordance with the aforementioned inventive concepts.

(ii) In a multi-route data network, where data packets are subject to discarding when a route is

5 congested, a reduced level of data packet discarding is performed when compared to known networks.

(iii) In a multi-route data network, where data packets are subject to discarding when a route is congested, the fewer data packets that are discarded are

10 routed to the destination node (or an intermediate node on the preferred route) via an alternative less-congested route.

Whilst the specific and preferred implementations of the

15 embodiments of the present invention are described above, it is clear that one skilled in the art could readily apply variations and modifications to the preferred embodiments that fall within the inventive concepts.

20 Thus, a method and re-routing apparatus to support an improved re-routing mechanism, for example in an IP-based network, has been described, whereby the disadvantages associated with known mechanisms, apparatus and methods have been substantially alleviated.

25

- 19 -

Claims (28)

Claims

1. A method of re-routing data in a data network

(500), the method comprising the step of: 5 receiving one or more data packets at a router in the data network;

wherein the method is characterised by the steps of:

identifying (515), at the router (410), a congested route from a plurality of selectable routes; 10 selecting (520), in response to the identification, an alternative route; and re-routing (525) the one or more data packets to the alternative route.

15

2. The method of re-routing data in a data network

(500) according to Claim 1, wherein the step of selecting (520) an alternative route comprises selecting an alternative less-congested route.

20

3. The method of re-routing data in a data network

(500) according to Claim 1 or Claim 2, wherein the step of re-routing (525) the one or more data packets on that alternative route comprises the step of adding a level of segmentation to the one or more data packets.

25

4. The method of re-routing data in a data network

(500) according to any preceding Claim, the method further characterised by the step of:

monitoring (510), at the router, a port load 30 status at a plurality of output ports of the router, such that the steps of identifying a congested route and/or selecting an alternative route are based on the monitored port load status.

- 20 -

5. The method of re-routing data in a data network (500) according to Claim 4, wherein the step of monitoring (510) includes polling the plurality of output

5 ports of the router to obtain port load status information.

6. The method of re-routing data in a data network (500) according to any preceding Claim, wherein the step

10 of re-routing comprises re-routing a portion of one or more data packets via the alternative route.

7. The method of re-routing data in a data network (500) according to Claim 6, wherein the router

15 additionally employs the step of discarding one or more data packets when a route is identified as being congested,

the method further characterised by the step of:

re-routing a portion or all of the one or more 20 discarded data packets via the alternative route.

8. The method of re-routing data in a data network (500) according to Claim 6 or Claim 7, wherein the step of re-routing a portion of one or more data packets via

25 the alternative route comprises re-routing the one or more data packets to an intermediate router on an originally preferred route.

9. The method of re-routing data in a data network 30 (500) according to any preceding Claim, wherein the router is a core router within the data network and the step of re-routing via the alternative route comprises

- 21 -

re-routing the one or more data packets to one or more edge routers.

10. The method of re-routing data in a data network 5 (500) according to any of preceding Claims 1 to 8,

wherein the router is an edge router within the data network and the step of re-routing via the alternative route comprises re-routing the one or more data packets to one or more further edge routers.

10

11. The method of re-routing data in a data network (500) according to any preceding Claim, wherein the data network (500) router employs a Multi-Protocol Label Switching process, the method further characterised by

15 the step of:

re-routing a Multi-Protocol Label Switching link;

or re-routing, as a plain datagram with no Multi-Protocol Label Switching.

20

12. The method of re-routing data in a data network (500) according to any preceding Claim, wherein the data network (500) router employs a differentiated services process, the method further characterised by the step of:

25 re-routing differentiated service tagged frames to a less-loaded edge router able to forward the frames.

13. The method of re-routing data in a data network (500) according to any preceding Claim, wherein the data

30 network (500) router employs a resource reservation protocol, wherein the steps of identifying, selecting and re-routing are applied when setting up a resource reservation protocol link.

- 22 -

14. The method of re-routing data in a data network (500) according to any preceding Claim, wherein the data network (500) is an Internet Protocol (IP)-based computer

5 or communication data network.

15. A data network router (410) for routing data packets in a data network, the router comprising:

a receiving function (425) receiving one or more

10 data packets; and a processor (420), operably coupled to the receiving function, analysing the received one or more data packets to determine a destination node address and/or a route to deliver the one or more data packets;

15 wherein the data network router (410) is characterised in that the processor is configured to identify a congested route from a plurality of selectable routes and, in response to the identification, select an alternative route for routing the one or more data packets.

20

16. The data network router (410) according to Claim 15, wherein the processor (420) selects an alternative less-congested route.

25

17. The data network router (410) according to Claim

15 or Claim 16, wherein the processor (420) adds a level of segmentation to the one or more data packets to be sent on the alternative route.

30

18. The data network router (410) according to any of preceding Claims 15 to 17, further characterised by a plurality of output ports, operably coupled to the processor (420), such that the processor (420) re-routes

- 23 -

(525) the one or more data packets to the alternative route via at least one of the plurality of output ports.

19. The data network router (410) according to Claim 5 18, further characterised by the processor comprising a monitoring function for monitoring a port load status at a number of the output ports, such that the identification of a congested route and the selection of an alternative less-congested route are based on the 10 monitored port load status.

20. The data network router (410) according to Claim 19, wherein the monitoring function comprises an output queue polling function (225) to poll the plurality of

15 output ports of the router (410) to obtain port load status data.

21. A data communication network (200, 300) adapted to incorporate the data network router (410) of any of

20 Claims 15 to 20 or adapted to facilitate the method steps of any of Claims 1 to 14.

22. A storage medium storing processor-implementable instructions for controlling a processor (420) to carry

25 out the method according to any of Claims 1 to 15.

23. An apparatus adapted to perform the method steps according to any of Claims 1 to 15.

30

24. A communication unit comprising apparatus according to Claim 23.

- 24 -

25. A data communication network (200, 300)

comprising a plurality of routers for re-routing one or more data packets over a plurality of data communication paths, the data communication network (200, 300)

5 characterised by a data network router (410) receiving one or more data packets and configured to identify a congested route for delivering the one or more data packets from a plurality of selectable routes based substantially on queue length per route and, in response 10 to the identification, the data network router (410)

selects an alternative route for delivering the one or more data packets.

26. A data communication network (200, 300)

15 substantially as hereinbefore described with reference to, and/or as illustrated by, FIG. 2 or FIG. 3 of the accompanying drawings.

27. A data network router (410) substantially as 20 hereinbefore described with reference to, and/or as illustrated by, FIG. 4 of the accompanying drawings.

28. A method of re-routing data in a data communication network (500) substantially as hereinbefore

25 described with reference to, and/or as illustrated by, FIG. 5 of the accompanying drawings.