Data Replication for Top-k Query Processing in Mobile Wireless Sensor Networks

Abstract

In mobile wireless sensor networks (MWSNs), since sensor nodes are highly resource constrained, it is effective to retrieve sensor observations using a top-k query, in which the observations are ordered by the value (or score) of a particular type of sensor, and the query-issuing node acquires sensor observations with the k highest scores. In our previous work, we proposed a message processing method for a topk query in mobile ad hoc networks (MANETs) for reducing traffic and keeping the accuracy of the query result. This method basically works well, however, it cannot avoid the deterioration of the accuracy when network partitioning occurs. In this paper, we assume MWSNs and propose data replication methods for keeping high accuracy of the query result even at the point of network partitioning. The proposed methods replicate sensor observations with high scores when each node sends and relays sensor observations to the query-issuing node.

Efficient cluster-based top-k query routing with data replication in MANETs

“…It is therefore non-trivial to extend the existing methods for top-k data retrieval in MANETs. Hara et al (2010) proposed a data replication technique for top-k data retrieval in MANETs. Basically, this method executes top-k query processing using the method proposed in Hagihara et al (2009), and caches data items with high scores when receiving the data.…”

“…Also, we integrated the naïve method in Sect. 6.1 and the data replication method in Hara et al (2010), and used this integrated method for comparison. Although a top-k data retrieval method in the environment where nodes hold replicated data has been proposed in Sasaki et al (2014), we do not use this method as a comparison method, because its performance is obviously worse than the integrated method.…”