Abstract | ||
---|---|---|
Routing policy, also known as selection function, is an important constituent of adaptive routing algorithm for interconnection networks. However, relatively very small research effort has been put on it until now. A generic design framework for it is absent, and the behind mechanism of how it affects the network performance needs further exploration. In this paper we study the routing policy for interconnection networks as a separate topic. The action of routing policy on network traffic is analyzed. We then clarify that routing policy actually works as a traffic scheduler. It together with some other mechanisms constitutes a traffic shaping paradigm for interconnection networks. Five generic routing policies are proposed in conformance to the analysis. They explicitly schedule the real-time network traffic based on traffic measurement. Counters are used as cost-effective measurement facilities. To standardize the design, constituents of a counter-based routing policy (CRP) are formally defined. Simulations covering extensive traffic patterns exhibit potentials of CRP in maximizing the network performance. |
Year | DOI | Venue |
---|---|---|
1999 | 10.1007/978-3-540-46642-0_56 | HiPC |
Keywords | Field | DocType |
counter-based routing policies,real time,network performance,traffic shaping,cost effectiveness,adaptive routing | Multipath routing,Link-state routing protocol,Dynamic Source Routing,Policy-based routing,Computer science,Static routing,Enhanced Interior Gateway Routing Protocol,Destination-Sequenced Distance Vector routing,Computer network,Wireless Routing Protocol,Distributed computing | Conference |
ISBN | Citations | PageRank |
3-540-66907-8 | 2 | 0.39 |
References | Authors | |
16 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xicheng Liu | 1 | 7 | 3.20 |
Yangzhao Xiang | 2 | 2 | 0.39 |
Timothy J. Li | 3 | 5 | 1.48 |