A fault-tolerant Manhattan routing method for networks with monotonically changing routing directions

Abstract

The invention discloses a fault-tolerant Manhattan routing method for a network with monotonously changing routing directions, which comprises the steps of: judging whether a source node and a target node are error nodes, and ending the process if at least one of them is an error node; judging the minimum value of each intermediate node The next hop node allowed by the routing policy, and record the last hop node allowed by the minimum routing policy of each intermediate node; set the path count value of the source node to non-zero, and calculate the path count value of the intermediate node and the target node ;Judge whether the path count value of the target node is 0, if it is 0, end the process; take the target node as the starting point, start to search for the last hop node whose path count value is not 0 and is allowed by the minimum routing policy hop by hop until the source is found node, get the fault-tolerant Manhattan path. The invention has low complexity, high universality and does not sacrifice the available fault-tolerant Manhattan path.

Description

technical field [0001] The invention relates to the technical field of reliability calculation. More specifically, it relates to a fault-tolerant Manhattan routing method for networks with monotonically changing routing directions. Background technique [0002] In many types of computer networks and on-chip networks, the Manhattan routing method (also known as the minimum routing method) has been widely used for its characteristics of low overhead and low complexity. Due to the inevitability of wrong nodes in the network, it is very meaningful to design the Manhattan routing algorithm (also called the minimum routing algorithm) that can avoid wrong nodes. [0003] There are a lot of work on fault-tolerant Manhattan routing algorithms at home and abroad. A typical approach is to design fault-tolerant routing algorithms based on the error block model, which can be divided into three stages: the first stage is to propose an error block model and the corresponding error block c...

AI Technical Summary

Problems solved by technology

Under the existing error block model, the construction process of the error block itself is relatively complex, such as the MCC error block model, and the algorithm for judging whether there is a fault-tolerant path is also relatively complicated. Some error module models such as fault-tolerant routes that can only tolerate one error Algorithms, convex error block models, and defense zone models will even sacrifice many available non-error nodes and available fault-tolerant paths, which ultimately makes the fault-tolerant routing algorithm very complex and low in availability

These error block models also have a common problem, that is, they can only be applied to specific topology networks and specific routing algorithms, and their universality is low.

Images