Abstract

The shared dataspace metaphor is historically the most prominent representative of the family of coordination models. According to this approach, concurrent processes interact via the production, consumption, and test for presence/absence of data in a common repository. Recently, the problem of the accumulation of outdated and unwanted information in the shared repository has been addressed. Typical garbage collection algorithms cannot be adopted in this context because there is no notion of unaccessible data. The most promising solution to this problem consists of the introduction of the notion of temporary data, intended as data with an associated expiration time. In this paper, we investigate the impact of different mechanisms for expired data collection on the expressiveness of shared dataspace coordination systems with temporary data.