On Independently Finding Converging Paths in Internet

Two shortest paths P (u, v) and P ′(v, u) are called converging paths if path P (u, v) is exactly the reversal of path P ′(v, u). Independently finding two converging paths from node u and node v, so that a message from u to v is bound to meet a message from v to u, is not trivial even if both nodes have the correct and complete network link state information. This paper presents three different converging path algorithms. The first one is a simple extension to Dijkstra’s shortest path algorithm. The rest two make use of the routing information existing in each node. Both theoretical complexity and simulated performance results for these algorithms are given and discussed.