US 6069895 Distributed route server

ABSTRACT – A design for a network route server in which network routing functions are distributed throughout the processing elements that constitute a switching node, while maintaining the global identity and routing information exchange functions of a route server element (RSE). Intelligent line-cards are provided having the ability to route independently of the RSE. This removes the RSE as a bottleneck resource and ensures that the capacity of the switching node is limited only by the switching capacity of its switch fabric. The RSE serves the functions of network topology discovery and routing table construction using a network topology database and an optimal routing algorithm. Copies of the dynamically maintained routing tables are distributed to the intelligent line-cards on a periodic basis governed by predetermined criteria. Wider geographical distribution of the RSE is enabled and most efficient utilization of the switch fabric is ensured. Scaling of distributed switching architectures is also enabled. The advantage is a significant increase in switching capacity as well as an increased degree of network connectivity.

FIELD OF THE INVENTION

The present invention relates to telecommunications packet or cell switches and, in particular to a scalable architecture for a network route server which is capable of efficiently switching both connectionless and connection-oriented payload traffic.

BACKGROUND OF THE INVENTION

In the currently-evolving network environments, there is an increasing requirement for flexible connectivity between switching nodes for the carriage of payload, control and management traffic. Although switching nodes were once monolithic entities, they are more and more becoming complex distributed architectures which facilitate diversity in services and scalability in capacity. The rapid growth of packetized data traffic in LEC and IEC networks requires that switching nodes perform muli-layer switching functions and handle both connectionless and connection-oriented network traffic. Therefore, there exists a need for scalable routing and forwarding functions in a distributed switching architecture.

Both existing and evolving routing protocols such as Asynchronous Transfer Mode (ATM), Forum Private Network-Network-Interface Specification (PNNI), Internet Open Shortest Path First (OSPF), and International Standards Organization (ISO) Open Systems Interface (OSI) identify a network node by the identify of its Route Server Element (RSE) which is some form of global address. Each RSE generates a routing table based on a preferred performance metric and a most recent topological view of the network. The table thus creates optimal routes for communication throughout the network. The RSE is responsible for administration of the algorithms that enable a node to keep its view of the network topology and performance metric current, referred to as Routing Information Exchange (RIE). This combination of responsibilities often necessitates the need for the RSE to act as a central focus for the routing of traffic through the switching node. The consequence of this is that the RSE becomes a bottleneck resource limiting the capacity and scalability of a distributed switch architecture.

The impact of this bottleneck has been recognized and efforts to minimize its impact have resulted in the invention of intelligent input/output controllers which autonomously handle a portion of the packet routing function. Such controllers are described for example in U.S. Pat. No. 4,494,230 which issued on Jan. 15, 1985 to Turner, and U.S. Pat. No. 5,367,518 which issued on Nov. 22, 1994 to Newman. These patents respectively describe packet switching systems dedicated to connection-oriented traffic in which intelligent controllers request packet routing information from a central control unit only on call connection setups. The routing information supplied by the central controller is stored in a routing look-up table in controller resident memory. Thereafter, the controller uses the same routing information to route all packets associated with the call connection, thus freeing the central controller to attend to other functions. While this switching model improves the switch throughput of connection-oriented packet switching systems, it does not alleviate the central controller bottleneck for connectionless payload traffic. The exponential growth of packetized data traffic is now placing onerous switching burdens on network nodes. There therefore exists a need for a network route server having scalable routing and forwarding functions in a distributed switching architecture in order to supply the packet switching capacity now required.

SUMMARY OF THE INVENTION

It is therefore an object of the invention to provide a scalable switching architecture for a network route server in which the routing functions are not deeply embedded in the switch architecture.

It is a further object of the invention to provide a scalable architecture for a network route server in which routing functions are distributed to intelligent line-cards which autonomously handle the routing of all packet traffic.

It is a further object of the invention to provide a scalable architecture for a network route server in which the central controller is dedicated to network topology discovery and dynamic routing table maintenance based on network payload traffic to ensure improved connectivity and switching efficiency in the network.

The invention therefore provides a network route server with a scalable switching architecture comprising a route server element with which a node identity in the network is associated, the route server element including processor means for executing a topology discovery protocol and maintaining a topology database of the network, and for creating a routing table for the route server using the topology database and a routing algorithm, the route server element periodically copying the routing table to an array of packet processing elements associated with the route server element;

each packet processing element having a link-port interface to the network, memory means for storing the copy of the routing table, and processor means for using the routing table to route packets received by the processing element whereby each packet processing element is enabled to autonomously route any packet received by the packet processing element by reference to the copy of the routing table; and

a switch fabric for switching of packets between the packet processing elements in accordance with the routing information provided by the packet processing elements.

The invention therefore provides a scalable switching architecture which removes routing as a functionality that is closely embedded within the switch architecture. It therefore becomes possible to integrate switching and routing across the network over a wider geographical area. The resulting architecture is capable of significantly increasing the switching capacity as well as increasing the degree of network connectivity. The network connectivity is increased because the RSE is dedicated to performing information routing exchange functions and relieved of the load-dependent function of traffic forwarding which is handled by the packet processing elements, hereinafter referred to as “intelligent line-cards”. Consequently, the scalable architecture in accordance with the invention permits improved scaling as network size increases.

The performance requirements of the distributed route server architecture are preferably enabled with a two-level priority control mechanism. The control connectivity mechanism should also be over-provisioned with sufficient capacity to permit multi-cast and reliable end-to-end transmission capability. Overall the performance of the distributed architecture provides reduced transit delays and a capacity that scales to the capacity of the switch fabric.

The network route server in accordance with the invention consists of a central control which is responsible for attending to all admission control for connection-oriented services and Virtual Circuit Identifier (VCI) tables for connection-oriented traffic. It is also responsible for network topology discovery and executes a network topology discovery protocol in which network topology packets are exchanged with other network nodes in order to maintain a dynamic link-state topology database that reflects the current network connectivity. Using the topology database and a preferred routing algorithm, the central control builds and rebuilds routing tables used for routing packets to an adjacent node. The VCI tables and the routing tables are downloaded to the intelligent line-cards on a periodic basis dictated principally by the amount of time required by the central control to perform routing table calculations. Thereafter the updated routing tables are distributed to the intelligent line-cards as soon as possible without putting undue stress on nodal resources. Routing table updates may likewise be dictated by a change in network topology or significant traffic fluctuations reported via link-state packets.

Related Posts