PACKET RING PROTECTION

Field of the Invention

The present invention relates to the field of telecommunications and more particularly to a method and apparatus for providing ring protection in a packet transport network.

Background of the Invention

In telecommunication networks, protection refers to the ability to recover protected traffic in the case of a failure by switching over automatically to a redundant resource. For TDM-based transport networks, sophisticated protection mechanisms have been described and defined. The basic standard in SDH (Synchronous Digital Hierarchy) that describes types and characteristics of protection architectures is ITU-T G.841 (10/98), which is incorporated by reference herein.

A basic topology in protected networks is a ring network. In order to avoid misconnections in a ring protection scheme in the case of a double failure, a mechanism known as "squelching" has been developed for SDH rings. Squelching implies that the traffic addressed to nodes that are isolated due to a double failure occurring in the ring is 'squelched' at switching nodes, i.e. replaced or overwritten with AIS (Alarm Indication Signal). A switching node is

any node that performs a bridge or switch function for a protection event. The generalized squelching logic is defined in G.841 Appendix II.

In order to make squelching work properly in a TDM ring it is required to configure each node in the ring with a ring map. Each of the ring maps must contain information regarding the order in which the nodes appear on the ring. In addition each node requires a connection map that contains the AU-4 time-slot assignments for traffic that is both terminated at that node and passed-through that node and a squelch table that identifies for each of these AU-4 time slots the node addresses at which the traffic enters and exits the ring.

Today, packet transport networks as opposed to classical TDM networks are emerging. Protocols for packet transport are currently under development, one of which is referred to as T-MPLS (Transport Multiprotocol Label Switching). Transport MPLS is a purely connection-oriented packet transport network based on MPLS that provides managed connections to different packet client layer networks.

Hence, protection mechanism and strategies for packet transport networks need to be defined. Packet ring technologies could possibly reuse the above mentioned protection techniques known from TDM networks. However, a direct adoption of TDM methods to packet transport network would not provide efficient bandwidth usage.

In a TDM network, a connection (e.g. a VC-4 timeslot) actually determines the bandwidth this connection occupies, regardless of the amount of user traffic currently transmitted over this connection. This means that since a TDM connection needs to be pre-provisioned, its bandwidth cannot be effectively used by other connections. Conversely, in packet technology, a connection (e.g. a T-MPLS label) specifies a route through the network but it does not determine the bandwidth which is actually occupied by user traffic transmitted on this connection. The total amount of bandwidth in a packet network can be effectively shared between different connections not only at the provisioning

stage but also during network operation in both normal and protection situations.

Moreover, it is expected that in T-MPLS rings the number of connections could reach more that ten thousands. This would increase the complexity and require a long time for the configuration (e.g., via network management or local crafts personnel) of all nodes in the ring through which these connections are passing and may make the target protection time of 50ms difficult to achieve.

The squelching mechanism in TDM developed for misconnection avoidance does not address the variable bandwidth nature of packet traffic and does not address the scalability problem, which arises due to the high number of connections in the packet ring. The inherent constrains of TDM technology when applied to packet rings would lead to inefficient bandwidth resource usage; configuration scalability issue; and increased protection switching time.

In view of the above discussed constraints, it is a problem of the present invention to provide a method and corresponding apparatus for providing packet ring protection.

Summary of the Invention

These and other objects that appear below are achieved by a ring protection method in a packet transport network with the following steps:

- A failure is detected at a first network node in the ring.

- The failure is communicated along the ring using an automatic protection switching protocol.

- At the first node or at a second node adjacent to the failure, a protection switch-over for at least one protected traffic stream is activated by redirecting the traffic stream into the reverse direction of the ring. Only selected network nodes along the ring squelch the traffic stream according to pre-configured squelching tables.

According to the invention, the step of squelching is performed only by networks nodes along the ring which perform either an add function or a drop&continue function for the traffic stream to be squelched.

This increases the efficiency of packet transmission compared to the SDH rings squelching mechanism since traffic streams that will be lately discarded anyway are replace at an early stage with an AIS signal that does not require much bandwidth.

According to a further aspect of the invention a network node for a packet ring network is provided which has first type interfaces for connection to the ring network, second type interfaces for adding or dropping traffic signals to or from the ring network, respectively, and a switch matrix arranged between said first type and second type interfaces. The switch matrix is controlled by a controller to redirect a protected traffic stream in the event of a failure. The controller is adapted to determine in accordance with a set of squelching rules and based on a pre-configured squelching table traffic streams to be squelched. In accordance with the invention, squelching table contains only traffic streams for which the network node performs either an add or a drop&continue function.

Brief Description of the Drawings

Preferred embodiments of the present invention will now be described with reference to the accompanying drawings in which figures 1 to 4 show a packet transport network with ring topology at different failure scenarios and figure 5 shows a block diagram of a network node for use in a packet transport network.

Detailed Description of the Invention

An ring network for packet transport is shown by way of example in figure 1 . The example network contains six network nodes Nl -N6 connected to form a two-fiber ring. Fiber 1 is operated in clockwise direction while fiber 2 is operated in counter-clockwise direction. Accordingly, signals on fiber are also designated as "eastbound" and signals on fiber 2 as "westbound".

Both fibers 1 , 2 can carry traffic and offer additional protection resources. Each node can be reached over both fibers. Nodes Nl , N2, N5, and N6 perform add or drop functions on the traffic in the ring. Nodes N3 and N4 only pass the ring traffic through. In particular, node Nl adds a traffic stream 4 to the ring. Node N2 performs a drop & continue function on the traffic stream 4. "Drop & Continue" function is a function that is implemented within a node wherein the traffic is dropped from the ring and, at the same time, continued over the ring. Similarly, node N5 performs a drop & continue function on the same traffic stream 4 and node N6 finally drops the traffic stream 4 from the ring. Hence a point-to-multipoint connection from node Nl to nodes N2, N5, and N6 exists.

When a failure occurs on the ring, such as shown between nodes N4 and N5 in figure 1 , this failure is communicated using an APS protocol (APS: automatic protection switching) to signal a failure and initiate a switch-over of traffic in the reverse direction. A well known APS protocol is the Kl /K2 protocol in SDH. Since in the present case, the ring is broken between nodes N4 and N5, both nodes detect a server signal failure (SSF) and communicate this failure over the remaining ring segment. As indicated by arrows 3, APS signaling hence occurs between nodes N4 and N5. Intermediate nodes N3, N2, Nl , N6 will analyze the APS messages. Due to this, all nodes Nl -N6 along the ring 1 , 2 will know about the failure.

As a reaction to the failure, node 4 performs a loop-back of the protected traffic stream 4 destined for nodes 5 and 6 from fiber 2 to fiber 1 . In the same

way node 5 would loops back protected traffic designated for node 4 from fiber 1 to fiber 2.

In the undisturbed ring, traffic stream 4 would be routed over ring 2 along the path Nl , N2 (drop&continue), N3, N4, N5 (drop&continue), and N6 (drop). After the ring failure between nodes N4 and N5, traffic stream 4 is re-routed along the path Nl , N2 (drop&continue), N3 to N4 over ring 2 and back from N4 over ring 1 along the path N4, N3, N2, Nl , N6 to N5 and back from N5 (drop&continue) to N6 (drop).

In this situation, no squelching is required, since no node is isolated through the failure. There is hence no substantial difference in the protection mechanism as compared to known TDM mechanisms.

In figure 2, the same ring network is shown but with a dual ring failure disconnecting the ring between nodes N4 and N5 and between nodes N6 and Nl . Now nodes N6 and N5 are isolated. In such situation APS signaling 3 will occur between nodes N4 and Nl , so that the all nodes Nl -N4 on the left ring segment are aware of the unavailability of nodes N5 and N6. In addition APS signaling 3' will occur between nodes N5 and N6.

In a traditional TDM protection scheme, Node Nl will continue sending user traffic 4, though it will be discarded at node N4, which means that part of the ring carries traffic that has no destination anymore.

In a traditional TDM ring protection scheme, node N4 would activate a protection switch from fiber 2 to fiber 1 but would squelch the traffic destined for nodes N5 and N6 in order to avoid misconnections. In the TDM protection scheme, each node which is not directly involved in connection add, drop, or drop&continue, i.e. nodes N3 and N4 must be aware of each passing connection to support squelching.

The number of connections passing through each node can be very high in a packet ring network. Moreover, configuring squelching table at every intermediate node for each connection addresses scalability issues.

According to a basic idea of the invention squelching tables are configured only for certain connection at nodes performing add, drop or drop&continue operations for this connection

In addition, the standard TDM ring protection scheme specifies that squelching is performed by switching nodes (i.e. node N4 in figure 2). This means that traffic is transmitted from an ingress node at least up to the switching node. In a packet network transmission of traffic to a node where it will be lately discarded means inefficient usage of ring bandwidth along the connection. Therefore, another optimization provided by the invention is to squelch the traffic as soon as possible. For point-to-point connections this means to squelch the connection at the add node. For point-to-multipoint connections this means to squelch at the last drop&continue node that is not isolated by the ring failure from the entry node.

According to the invention, only an AIS message (AIS: Alarm Indication Signal) is sent, which does require much bandwidth. The ring bandwidth is not occupied with user traffic that later will be discarded anyway. As an alternative, the failed traffic stream could also simply be terminated. So, the term squelching in the context of the present invention should be understood to include AIS insertion as well as termination of the traffic stream.

In other words, the proposed solution modifies the set of squelching rules known in TDM, so that squelching is performed only by 'add' or 'drop&continue' (d&c) nodes instead of switching nodes (node N4 in figure 2). The connection map for protected connections in such case is provisioned only to add/drop/d&c nodes involved in the corresponding connection (nodes Nl , N2, N5, N6 in figure 2), and not provisioned to all intermediate nodes along the connection path (nodes N3, N4 in figure 2).

A ring node map is, however, provisioned to all nodes in the ring as it is in the standard TDM scheme. Nodes will discover the ring failures either locally or remotely by an APS protocol supported through APS messages exchange among the ring nodes. So ring failure are detected in the same way as in the known TDM solution.

Below is a list of the modified squelching rules used in preferred embodiments.

a) for point-to-point connections:

Exit node failure:

Assumed, with respect to the add node, that the failure is in the direction of the unidirectional circuit. Squelch the circuit (replace the working signal with the AIS at add node in the downstream direction) if and only if the node failure scenario includes the exit node for the unidirectional circuit.

Entry node failure:

Assumed, with respect to the drop node, that the failure is in the opposite direction from the direction of the unidirectional circuit. Squelch the circuit (replace the drop signal with the AIS at drop node) if and only if the node failure scenario includes the entry node for the unidirectional circuit. - Add and drop nodes (not switching nodes):

Analyze APS information transmitted by switching nodes and perform squelching according to the rules above

b) for p-t-mp connections:

Next and exit node failure

Assumed, with respect to add or d&c node, that the failure is in the direction of the multiple dropped unidirectional circuit. Squelch the circuit

(replace the working signal with the AIS at add or d&c node in downstream direction) if and only if the node failure scenario includes both the next d&c node and the exit node for the multiple dropped unidirectional circuit. Entry node failure

Assumed, with respect to the d&c or drop node, that the failure is in the opposite direction from the direction from the repeatedly dropped unidirectional circuit. Squelch the circuit (replace the working signal with the AIS at d&c or drop node) if and only if the node failure scenario includes the entry node for the multiple dropped unidirectional circuit.

- Add, drop and d&c nodes (not switching nodes)

Analyze APS information transmitted by switching nodes and perform squelching according to the rules above

As a consequence of the dual ring failure in figure 2, node N2 performs squelching of the 'continue' traffic forwarded on the ring since both the next node N5 and the exit node N6 are isolated. Node N5 performs squelching because entry node Nl is isolated. The arrows indicating the path of traffic stream 4 on rings 1 and 2 are shown with broken lines for squelched connections in figure 2. While node N2 squelches only the 'continue' traffic on the ring, node N5 squelches both the 'drop' traffic exiting the ring and the 'continue' traffic looped back onto ring 1 towards node N6.

Figure 3 shows the same ring network but with a node failure of node N5. Due to this failure, the connection between nodes N4 and N6 via node N5 is broken and node N4 and N6 hence detect an SSF event. APS signaling occurs between nodes N4 and N6. All intermediate nodes Nl -N3 analyze the APS messages and hence know about the failure. Node N4 performs a loop-back of traffic destined for node N6. Due to the activation of the protection switching in node N4, node N6 receives the traffic stream 4 from ring 1 . Node N6 would also loop back traffic destined for node N4 if there was any.

According to the above rules, node N2 does not perform squelching since only node N5 is isolated but traffic stream 4 is still required for nodes N2 and N6.

In figure 4, a node failure affects node N6 rather than node N5. APS signaling is now between nodes Nl and N5. All nodes analyze the APS messages and hence know about the failure. Node N2 does not perform squelching, since only node N6 is isolated but traffic stream 4 is required for nodes N2 and N5.

As α reaction to the failure, node N5 activates a protection switching of traffic stream 4 from ring 2 back to ring 1 . However, since node N6, which is both next and exit node for traffic stream 4, is isolated, node N5 squelches traffic stream 4 looped back to ring 1 . The squelched signal is indicated by a broken line in figure 4.

A network node N that can be used for the implementation of the invention is shown schematically in figure 5. Network node N has west and east interfaces for connection to the ring. In particular a west input IW, a west output OW, an east input El and an east output EO. In addition, node N has a number of tributary interfaces Tl -T3 for adding and/or dropping tributary traffic streams to or from the ring, respectively. Between the interfaces, there is a switch matrix SW, which connects east and west interfaces and tributary interfaces.

A controller CT controls the switch matrix SW. Controller CT receives error messages such as SSF from the input interfaces El, Wl and reconfigures the switch matrix SW for protection switching in the event of a failure. Moreover, APS signaling is extracted at the input ports El, Wl and forwarded to controller CT.

Figure 5 shows a simplified failure scenario. Traffic at east input El is lost and a loss of signal (LOS) detected. As a consequence action, protected traffic from west input Wl is looped back by switch matrix SW to each output EO, so that it can reach its destination over the reverse ring direction. In addition the APS signal containing the signal failure indication is transmitted to the node adjacent to the node N over the short path (i.e. through interface EO) and over the long path (i.e. through interface WO).

According to the invention, controller CT stores a connection map and a squelching table. Based on this information, controller CT determines in accordance with predetermined squelching rules traffic signals to be squelched and instructs the output interfaces EO, WO to squelch such output signals. Thus controller CT implements the squelching rules discussed before.

It should be noted that east and west interface are distinguished only for the purpose of their function in the ring but can be of the same physical structure. The same comment applies also for the tributary interfaces.

After having read and understood the above, non-limiting embodiments of the invention, it will be appreciated that the proposed modifications to the squelching rules allow implementation of an efficient and scalable ring protection scheme in a packet ring network. Those skilled in the art will also appreciate that various modifications and changes are possible to the abode described implementations. It will be appreciated for example, that even though the invention has been explained for a 2-fiber ring, the proposed protection scheme can equally be applied to a bidirectional 4-fiber ring.

According to a further improvement, it is proposed to reconfigure certain squelching tables at the ring nodes after detection of a failure. In case of point-to-multipoint connections the solution implies that the d&c nodes are aware of the status of next and the last node participating in this connection. This means that when the connection is modified so that the drop point added/removed at certain node, the previous node participating in this connection (relative to the connection flow) should reconfigure its squelching table to indicate the changes.