Querying a PDMS means either flooding the network with messages to all peers or taking advantage of a routing mechanism to reformulate a query only on the best peers selected according to some given criteria. As reformulations may lead to semantic approximations, we deem that such approximations can be exploited for locating the emph{semantically best directions} to forward a query to. In this paper, we present our experiences in devising and testing a mechanism for effective query routing in a PDMS. In particular, we describe a distributed index mechanism where each peer is provided with a Semantic Routing Index (SRI) for routing queries effectively. We illustrate SRIs' structure, their use and the framework we devised for their incremental update, then we provide an extensive evaluation of their effectiveness through a set of query routing experiments on a variety of scenarios. This work is partially supported by the PRIN WISDOM and FIRB NeP4B national projects.
Semantic Query Routing Experiences in a PDMS / Mandreoli, F.; Martoglia, R.; Penzo, Wilma; Sassatelli, S.. - ELETTRONICO. - (2006). (Intervento presentato al convegno 3rd Italian Semantic Web Workshop on Semantic Web Applications and Perspectives (SWAP 2006) tenutosi a Pisa nel 18-20 December 2006).
Semantic Query Routing Experiences in a PDMS
PENZO, WILMA;
2006
Abstract
Querying a PDMS means either flooding the network with messages to all peers or taking advantage of a routing mechanism to reformulate a query only on the best peers selected according to some given criteria. As reformulations may lead to semantic approximations, we deem that such approximations can be exploited for locating the emph{semantically best directions} to forward a query to. In this paper, we present our experiences in devising and testing a mechanism for effective query routing in a PDMS. In particular, we describe a distributed index mechanism where each peer is provided with a Semantic Routing Index (SRI) for routing queries effectively. We illustrate SRIs' structure, their use and the framework we devised for their incremental update, then we provide an extensive evaluation of their effectiveness through a set of query routing experiments on a variety of scenarios. This work is partially supported by the PRIN WISDOM and FIRB NeP4B national projects.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.