LighTS

Abstract

The tuple space model inspired by Linda has recently been rediscovered by distributed middleware. Moreover, some researchers also applied it in the challenging scenarios involving mobility and more specifically context-aware computing. Context information can be stored in the tuple space, and queried like any other data.Nevertheless, it turns out that conventional tuple space implementations fall short of expectations in this new domain. On one hand, many of the available systems provide a wealth of features, which make the resulting implementation unnecessarily bloated and incompatible with the tight resource constraints typical of this field. Moreover, the traditional Linda matching semantics based on value equality are not appropriate for context-aware computing, where queries are often formulated over value ranges, and where there is a prominent need to deal with imprecise information coming from multiple sources.In this paper, we describe a new tuple space implementation called LIGHTS. Originally developed as the tuple space core of the LIME [11] system, LIGHTS provides a flexible framework that makes it easy to introduce extensions to the tuple space and in general to customize the tuple space implementation. The design and programming interface of LIGHTS is presented, and its flexibility demonstrated by illustrating extensions that proved useful in the development of context-aware applications.