An on Demand Path Marking and Capacity Reservation Method using Split Agent
Date Issued
2004
DOI
http://doi.ieeecomputersociety.org/10.1109/CLADE.2004.1309099
Abstract
Different schemes for large scale networks hosting distributed applications have been recently adopted for network path marking based on adaptive behavior of swarm-based agents. Topologically complex networks must use efficient routing methods in order to route data traffic from a source to a destination. In this paper the Split Agent-based Routing Technique (SART) is applied to a network in which a certain amount of data traffic is injected to mark different paths. Split Agent-based Routing Technique (SART) [Split agent-based routing in interconnected networks] is a variant of swarm-based routing [Ant-based Load Balancing in Telecommunications Networks, Ant based probabilistic routing with pheromone and antipheromone mechanisms] where agents are split after their departure to the next node on a hop-by-hop basis at the same time performing path marking. Packets that are delay sensitive are marked as prioritized. Agents recognize these packets as being a part of, and try to influence the two way routing tables. Thorough examination has been done for the performance and path recoverability using SART algorithm for marking any requested path, taking into account a number of metrics. It is shown that the split agent scheme offers an efficient on demand path marking and path-capacity reservation in a decentralized manner (particularly efficient for large scale networks). SART method undoubtedly represents a solution for significant network optimization particularly for the equal share of network resources. Also it is shown that this scheme offers a way to increase overall performance.