Performance Evaluation of Content-Based Information Retrieval Schemes for DTNs

Abstract

Mobile nodes in some challenging network scenarios suffer from intermittent connectivity and frequent partitions e.g. battlefield and disaster recovery scenarios. Disruption Tolerant Network (DTN) technologies are designed to enable nodes in such environments to communicate with one another. Several DTN routing schemes have been proposed. However, not much work has been done on providing information access in such challenging network scenarios. Existing client/server paradigm for information access will not be feasible in such scenarios since end-to-end path does not exist. Thus, in this paper, we explore how a content-based information retrieval system can be designed for DTNs. There are three important design issues, namely (a) how should data be replicated and stored at multiple nodes, (b) how should a query be disseminated in sparsely connected networks, (c) how should a query response be routed back to the querying node. We first describe two data caching schemes: (a) K-copy random caching, (b) K-copy intelligent caching. Then, we describe an L-hop Neighborhood Spraying (LNS) scheme for query dissemination. For message routing, we either use Prophet routing scheme or Highest Encounter First Routing (HEFR) scheme. We conduct extensive simulation studies to evaluate different combinations of these algorithms. Our results reveal that the scheme that performs the best is the one that uses the K-copy intelligent caching combined with the LNS query dissemination and HEFR scheme.