A Heuristic Address Assignment Algorithm Based on Probability in a Network

With the fast growth of the routing table sizes in backbone routers, reducing the size of routing tables on network becomes to be a basic problem in networking. After optimizing the content of a routing table, further reduction can be achieved by renumbering the address assigned to networks. In this paper, we propose a heuristic address assignment algorithm based on probability in order to reduce the routing table size. Our algorithm can guarantee the high probability that each node uses the same next hop to the nodes assigned with consecutive addresses, and then these consecutive addresses can be compressed to compact the routing tables. To evaluate this algorithm in practice, we implement it on real Internet inter-domain topology. The experimental results have shown that our method could reduce the size of the table to 24% of the complete routing table while keeping the routes shortest paths. Moreover, we extend the routing method to allow the suboptimal routes, for each node only keeping a simpler routing table storing the routing state: O(d), where d is the node degree. By comparison with the algorithmic routing which uses a spanning tree to compute next hop for any source and destination pair, our routing method using this simpler routing table can provide the lower stretch.