A stateless QoS routing algorithm subject to multiple constraints

One of the key issues in QoS provisioning in high-speed networks is how to determine a feasible route that satisfies the given QoS requirements while efficiently utilizing network resources. In this paper, we study the NP-Complete problem of path selection subject to multiple constraints and propose a heuristic solution, which essentially divides an entire QoS-path into at most two "superedges" that is connected by a "relay node". A superedge is defined as a connected segment of the path on which all routers use the same routing metric for packet forwarding. The node connecting the two superedges is called relay node. This property makes the heuristic be able to support stateless forwarding, i.e., no flow-specific state information is required to be maintained at intermediate nodes on a QoS-path, which can significantly improve the scalability of a QoS routing protocol. Its computational complexity is deduced to be O(m\V\(2)), where m, a very small integer, is the number of the concerned QoS metrics and \V\ is the number of nodes in a network. Simulation results show that the heuristic can achieve near optimal performance.