Social-optimized win-win resource allocation for Self-organizing Cloud

With the increasing scale of applications and the number of users, we design a Self-organizing Cloud (SoC) which aims to make use of the distributed volunteer computers or dedicated machines to provide powerful computing ability. These resources are provisioned elastically according to user's specific demand, by leveraging virtual machine (VM) resource isolation technology. Based on such a framework, we propose a social-optimized auction-based resource allocation scheme, which mainly tackles two issues: (1) how to make full use of the widely dispersed multi-attribute idle resources to construct a win-win situation, such that each task schedule could let both sides (resource providers and consumers) be satisfied with their final payoffs. (2) The total resource utility welfare should also be optimized to guarantee the overall performance around the global system. The key challenge of getting the win-win effect with social optimization is its provable NP-completeness. Finally, we validate that our approach can effectively improve the resource contributor's payoffs up to about five times as the level without our method via simulation work. Meanwhile, our approach can also keep a high level of the processing rate for the task scheduling.