SVC routing in network with static routing tables

Details SVC routing in network with static routing tables SVC routing in network with static routing tables

2001-03-22

2003-05-06

Vincent, David (Department: 2661)

Multiplex communications

Pathfinding or routing

C370S395320

Reexamination Certificate

active

06560218

ABSTRACT:

This invention relates to a method of establishing a switched virtual circuit in digital networks with static routing tables. The invention is applicable to any technology in which switched virtual circuits are set up through a packet switched network, for example, ATM and Frame Relay networks.
Congestion may occur on a network link if many incoming streams of traffic all terminate on the same outbound link, or the outbound link may busy or down due to a failure. Congestion is a problem in all networks. In packet switched networks, congestion is handled by end to end applications. There are no methods for handling congestion at the packet routing level. In circuit switched networks, bandwidth is reserved for each circuit by the network operator. Once again, there are no methods for handling congestion.
In networks which use switched virtual circuits (SVCs), a method is needed to reroute around congested or failed links. In the P-NNI protocol, which applies to dynamically routed SVC networks, a crankback IE (Information Element) is used for rerouting. When a node receives a call set-up request with which it is unable to comply, it sends a crankback IE back through the network. This IE passes through several nodes until it reaches a node which is programmed to respond to a crankback IE. This node, which is equipped to respond to a crankback element, then attempts to re-route the call on the basis of its current routing table.
This method, however, cannot be used in statically routed SVC networks, where the nodes contain manually pre-configured routing tables, which predetermine the path through the network that a call between any two endpoints will take. Furthermore, there is no mechanism to prevent continuous crankback attempts. In a network with static routing tables, failure in one of the configured links will result in failure of the call set-up process.
Routing loops may occur due to configuration errors, the use of alternate routes when failures occur, or transient routing tables after a failure in the network. Routing loops are also a problem in all networks. In statically routed packet and circuit switched networks, loops are assumed to be detected by the network operator. No loop detection mechanisms are present in these networks.
In networks which use switched virtual circuits, both transient loops and permanent loops are a problem. A reliable method of loop detection is therefore needed in these networks. The P-NNI protocol uses source routing to avoid loops. This adds unnecessary complexity to SVC routing.
The article “An Alternate Path Routing Scheme Supporting QOS and Fast Connection Setup in ATM Networks”. Proceedings of the Global Conference, San Fransisco, Nov. 28-Dec. 2, 1994, vol. 2 of 3, by E. M. Spiegel et. al discloses to an alternate path routing scheme based on a combination of progressive control and source routing for use within a connection-oriented data communications network, namely an ATM network. Spiegel teaches that a source node, which is to setup a connection through the network to an endpoint, computes a primary (first) path from its routing table information. Source routing is utilized to effect this path computation which typically is a function of link cost and quality of service requirements. The source node also computes a cost threshold and crankback limit which, together with the computed primary path, are included in a connection setup message generated at the source node. Additionally, the setup message includes fields for accumulating the cost of a path fragment already established and for accumulating the number of crankbacks already completed.
The setup message is propagated along the primary path until either the network endpoint is reached or until the connection setup message encounters a blocked link at an intermediate node. The intermediate node, responsive to encountering the blocked link, then computes an alternate path as a new path tail through the network to the endpoint. Again this path computation is effected using source routing, but in this instance the node must take into consideration the re-route guidelines contained in the connection setup message. The total cost of the path, comprising the accumulated cost of the established path fragment together with the cost of the new path tail, must be less than or equal to the cost threshold specified in the setup message.
If no path tail that satisfies the cost threshold can be found by the node, and the connection setup message has not yet been cranked back the limit specified therein, then crankback proceeds to the previous node in the path fragment and that node in turn attempts to compute an alternate path using source routing and complying with the constraints specified in the setup message.
According to the present invention there is provided a method of establishing a switched virtual circuit call in a digital communications network comprising a plurality of network nodes, each network node having a local static routing tables providing next hop routing information to adjacent nodes, characterized in that said routing tables define a primary route and an alternate route to adjacent nodes; a network node receives a setup message for the call and searches its routing table for corresponding routing information; the node, based on the corresponding routing information, attempts to forward the setup message on the primary route; if the primary route is not usable due to congestion or physical failure, the node then attempts to forward the setup message on the alternate route; and if the alternate route is the same route on which the setup message is received, the node cranks the call back to a preceding node which either forwards the setup message over the alternate route defined in that node's routing table or again cranks the call back to a further preceding node.
Preferably, the crankback IE has a predetermined lifetime to prevent continuous crankback attempts.


REFERENCES:
patent: 4345116 (1982-08-01), Ash et al.
patent: 5130982 (1992-07-01), Ash et al.
patent: 5649108 (1997-07-01), Spiegel et al.
patent: 5805593 (1998-09-01), Busche
patent: 5933412 (1999-08-01), Choudhury et al.
patent: 6215765 (2001-04-01), McAllister et al.
“An alternate path routing scheme supporting QOS and fast connection setup in ATM networks”, Spiegel et al., Institut of Electrical and Electronics Engineers, vol. 2 of 3, 1994, pp. 1224-1230.
“Operation and maintenace (OAM) mechanisms for layer management in ATM networks”, G. Schroder, 13th Annual Conference on European Fibre Optics Communications and Networks, vol. 2, 1995, pp. 97-101.
“Explicit path routing for switching network”, IBM Technical Disclosure Bulletin, 1976, vol. 18, No. 9, pp. 3059-3062.

Affiliated with

Also associated with

Rating

SVC routing in network with static routing tables does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with SVC routing in network with static routing tables, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and SVC routing in network with static routing tables will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3006251