Resource finding in store-and-forward networks

The process of searching for a resource in a distributed system whose nodes are connected through a store-and-forward network is modeled. Based on this model, a lower bound on the number of messages needed for finding a resource when nothing is known about its location is shown. The model also helps to establish some results about the complexity of finding optimal algorithms to locate a resource when the probability distribution for the location of the resource is known. It is shown that the optimization problem is NP-hard for general networks. An algorithm is developed for tree networks which can be specialized to polynomial algorithms for a class of trees. (The polynomial algorithms can be used as the basis of heuristic algorithms for general networks.) An application of this algorithm for path networks can be adapted to find optimal search algorithms for bidirectional ring networks.<<ETX>>