Partially Ordered Knowledge Sharing and Fractionated Systems in the Context of Other Models for Distributed Computing

Abstract

The latest sensor, actuator, and wireless communication technologies make it feasible to build systems that can operate in challenging environments, but we argue in this paper that the foundations needed to support the design of such systems are not well developed. Traditional models based on strong computing primitives, such as atomic transactions, should be replaced by weaker models such as the partially ordered knowledge sharing model, which we motivate in this paper and put into context of existing research. We also introduce a general probabilistic semanticsfor our model and the flavor of its specialization to characterize fractionated systems, an interesting class of systems with a potentially large number of redundantly operating components that can be programmed independently of the actual number that is deployed or operational at runtime.

Keywords: Shared Memory, Fractionate System, Distribute Hash Table, Tuple Space, Distribute Shared Memory.