Optimizing Concurrent Queries in Wireless Sensor Networks

Abstract

Wireless Sensor Networks (WSNs) are a promising area of research with a number of compelling potential applications. However, the capabilities of these applications are limited by the power-limitation, which is an inherent limitation for almost any WSN. Nowadays, it is the trend to share aWSN among multiple users, because not everyone can afford to deploy and use a WSN. In a shared-WSN, the power-limitation problem could get even more critical as the power-consumption would increase with the number of users. Furthermore, in a shared-WSN, multiple users may want to query and get data from the WSN at the same time: concurrent queries are possible to occur. Fortunately, unlike in traditional database context where concurrency is an overhead, in WSNs concurrency seems to be advantageous. As such, several researches have been carried out in the history to yield this advantage by optimizing concurrent queries. However, these researches assume WSNs with only one base station, thus have limited applicability in the present world, as many WSNs have multiple-base stations nowadays. This research introduces a concurrent-query optimization scheme which addresses any WSN: one with a single-base station or with multiple-base stations. The suggested scheme performs concurrent-query optimizations at two levels in the WSN: (1) at the base station-level, (2) at the node-level. The base station-level optimization scheme satisfies newly submitted queries from already executing queries, with the aim of reducing communication cost, sensing cost and processing cost. The node-level optimizer which operates on every sensor node in the WSN, optimize concurrent queries in two main ways. Firstly, whenever multiple queries demand a data acquisition at the same time, it will be done only once and the acquired value will be shared among those queries, with the intention of saving sensing cost. Secondly, whenever more than one query-result packet need to be sent to a certain base-station, to satisfy multiple concurrent-users, all results will be sent in one packet, with the aim of saving communication cost. The evaluation-results show that the suggested scheme does a significant energy saving in both single-base station and multiple-base station scenarios. The most amount of energy saving was achieved from saving communication cost.