Balancing Message Replication in Opportunistic Store and Forward Networks

We propose a novel technique utilizing distributed counter values to provide for balanced data replication in "opportunistic networks" - networks of mobile nodes who are for the most part entirely disconnected from one another, only occasionally meeting to exchange data.

Our technique for balanced replication works to spread data produced by nodes evenly throughout the network.

By spreading data evenly, our system both helps to minimize the average time a data item takes to reach an information gatherer and also increase the robustness of the system to failure of a subset of its nodes.

This algorithm is (unsurprisingly) fully distributed and (more notably) automatically self-adjusts to any network size, without needing to know the number of nodes in the network or their movement patterns.

replica growth We envision this technique being of application to:
  • autonomic opportunistic communications such as those proposed by the Haggle Project
  • opportunistic surveillance mechanisms
  • PDA-based disaster scenario messaging systems

Selected Publications:

Joshua Reich, Vishal Misra and Dan Rubenstein, Roomba MADNeT: a Mobile Ad-hoc Delay Tolerant Network Testbed, MC2R: Mobile Computing and Communications Review, ACM Sigmobile, 2008. [Paper] [BibTeX]