Distributed Name-based Entity Search

Internet can be seen as a network of peers that store digital representations of entities from the real world (e.g., person, locations, events). Different peers locally represent different “versions” (i.e., different points of view) of the same real world entity. In these different versions, entities are normally identified by multiple (possibly different) names. We propose a distributed entity search based on names that aims to (i) find all the different versions of an entity starting from any name used somewhere in the network to identify such entity; and (ii) allow peers to have full control over the privacy of their local representations. We evaluate our approach by setting up a network of 150 peers on Plan- etLab. The results show that the performance of our algorithms is stable with the network growth, which is promising in terms of scalability.