Universal scheme improving probabilistic routing in delay-tolerant networks

Many application protocols designed for the Internet architecture cannot operate well in challenged network environments. The communication in a delay-tolerant networking architecture, which is characterized by very long delay paths and frequent network partitions, basically relies on asynchronous, store-carry-forward message delivery. In most of the existing probabilistic routing approaches, messages are promptly forwarded to nodes that have higher delivery predictability values. We find that it is wise to wait till much better opportunities arise to minimize the communication cost without degrading the delivery ratio and latency. In this paper, a universal scheme, named E-Scheme, is proposed to improve most of the existing routing approaches based on the "probability to deliver" metrics. Providing that there are n opportunities, E-Scheme lets the first k opportunities go by (except when the node encounters the destination of the messages) and accepts the opportunities which are better than all of the k ones. We demonstrate that when k equals (n - 1)/e or root(n - 1)(n - 2)/e, where e is the mathematical constant, E-Scheme improves performance of existing probabilistic routing approaches in single-copy and 2-copy scenarios, respectively. The performance of E-Scheme has been evaluated in personal delay-tolerant networks in both single-copy and 2-copy situation. (C) 2013 Elsevier B.V. All rights reserved.