A Framework for Adaptive Optimization of Remote Synchronous CSCW in the Cloud Computing Era

Abstract

The systems for remote synchronous computer supported cooperative work (CSCW) are significant to facilitate people’s communication and promote productivity. However, in the Internet, such systems often suffer from the problems of relatively large latency, low bandwidth and relatively high cost of wide-area networking. Previous works tried to improve various mechanisms of communication, but till now we still cannot get rid of these problems due to the nature of the Internet data transmission mechanism. Rather than making optimizations based on the traditional CSCW computing style as previous work did, this paper proposes an idea of moving appropriate collaborative instances to the proper computing nodes which are just born in the emerging Cloud computing environments. Moreover, the paper presents a formal framework AORS to optimally organize the collaborative computing upon the emerging computational resources from the perspectives of both performance and cost. The formulization of the framework is proposed, and an analytic theory is developed. Directly solving the modeled problem has to refer to the exhaustive search, which is of exponential computational complexity; so we develop two heuristics. The experimental evaluations demonstrate the high efficiency and effectiveness of the heuristics. Furthermore, we conduct extensive simulation experiments on the current collaborative computing style and AORS. They illustrate that AORS brings the CSCW applications better communication quality and lower cost.