Fully Distributed Indexing over a Distributed Hash Table
Real-time communication, as well as simple data and file sharing motivates relevant research in network design nowadays. While centralized structures are generally favored over distributed ones for sake of simplicity, a considerable amount of literature has been devoted to the latter. In particular, the problem of distributed indexing is not trivial, and has been addressed by extending classical data structures such as tries, kd-trees. However, all proposed solutions seem to assume that there is a central and unique entity handling the indexing. In this paper, we propose a fully distributed indexing strategy by extending a data structure called prefix hash tree (PHT). More precisely, in this strategy, each node is part of the distributed network and participates in maintaining the distributed index. Our ideas have been implemented in a freely available framework called OpenDHT.