Performance evaluation of information retrieval schemes for multi-attribute queries in DTNs

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. In an earlier work, we studied information retrieval schemes for single-attribute queries in DTNs. Our schemes disseminate replicated data copies and queries to local-neighborhood. However, data items often have multiple attributes and not every node can be trusted to store replicated data or queries. Thus, in this paper, we study the scenario where the queries have multiple attributes. In addition, we compare the effectiveness of using opportunistically encountered nodes or specially deployed index and storage points (ISPs) for storing replicated data items or indices of the data items. Specifically, we conduct extensive simulation studies to evaluate three information retrieval schemes namely (a) the pre-determined ISP advertisement (PISA) scheme where ISPs advertise indices of data items with attribute values that fall within some pre-determined ranges, (b) the opportunistic regular node advertisement with index duplication (ORNA-ID) scheme which uses opportunistically encountered nodes for advertising replicated indices, and (c) the opportunistic regular node advertisement with data duplication (ORNA-DD) scheme which uses opportunistically encountered nodes for storing replicated data items. Our results indicate that the ORNA-ID scheme and the PISA scheme achieve similar performance. This shows that either architecture can be used. In addition, the ORNA-DD scheme provides 91% more query success rate and 128% more overall success rate in the sparsest network scenario (40 nodes distributed over 4000times4000 m²).