In this paper, we study the problem of distributed RDFS reasoning over structured overlay networks. Distributed RDFS reasoning is essential for providing the functionality that Semantic Web and Linked Data applications require. Our goal is to present various inference techniques for RDFS reasoning in a distributed environment, and analyze them both theoretically and experimentally. The reasoning methods we present are based on bottom-up and top-down techniques and have been implemented on top of the distributed hash table Bamboo. Our algorithms range from forward and backward chaining ones to rewriting algorithms based on magic sets. We formally prove the correctness of the algorithms and study the time-space trade-off they exhibit analytically and experimentally in a local cluster.

http://link.springer.com/article/10.1007%2Fs13740-013-0018-0