A Bloom Filter-Based Algorithm for Routing in Intermittently Connected Mobile Networks

Abstract

In this paper, we present a new protocol for routing in intermittently connected mobile networks that, by periodically exchanging constant-size Counting Bloom filters, assigns to every node in the network probabilities of reaching any destination. The gradients defined by these probabilities are further used to forward data packets towards any node in the network. The proposed protocol is based on two novel operations defined over the Bloom filters, namely, the unary degradation operation that models the loss of topological information as it gets stale or as it is propagated away from the place where it was generated; and the binary addition operation that is used to acquire topological information from other nodes. These two operations are used to implement a probabilistic form of soft state that is defined in terms of the content of the Counting Bloom filters. We present a series of experimental results based on extensive detailed simulations that show that the proposed protocol outperforms the Epidemic routing protocol by delivering more data packets with less delay, while inducing less total overhead in both MANET and VANET scenarios.