Extending query languages for in-network query processing

Abstract

Sensor networks have become ubiquitous and their proliferation in day-to-day life provides new research challenges. Sensors deployed at forest sites, high performance facilities, or areas striken by environmental, or other, phenomena, are only a few representative examples. More recently, mobile sensor networks have made their presence and are rapidly growing in numbers, such as the successful ZebraNet project or PDAs and smartphones. Nevertheless, such networks have mainly been used for data acquisition and data are being processed externally instead of in-network. Basic research problems that arise in the in-network setting include how to adjust in a timely and efficient manner to changing conditions and network topology. In this paper, we present a methodology, based on declarative query processing to alleviate the aforementioned problems, by making the deployment and optimization of a data analysis application as automatic as possible, which also helps execution in mobile environments. Our proposed solution focuses on extending a state-of-the-art sensor network platform, SNEE, with builtin data analysis capabilities.

Distributed Database Management Techniques for Wireless Sensor Networks

“…The work in [100] is an enhancement over SNEE. The basic idea is based on declarative query ptimize, and deploy automatic data analysis techniques in a sensor network.…”

Mining and managing user-generated content and preferences

“…The occurrence of an inten- sional extent triggers its substitution by a templated subplan, which performs its algorithmic computations. We collectively refer to this process as query refactoring [208]. In essense, we reformulate an initially posed query into an equivalent one, that is also expressed in the same declarative language (SNEEql).…”

“…Our approach is to model data mining algorithms as extensional extents, allowing us to include them in declarative queries like any other extent (e.g., relation, view, etc.) [208]. For the same reason, query execution engines are able to optimize these constructs, as long as they can be expressed with the declarative language.…”