为了满足用户需求,实现云资源能力的动态扩展,提出了一种多云联盟机制。首先,以最大化总收益为目标,以联盟博弈论对多个CRPs之间的合作行为进行建模,旨在形成稳定最优的CRPs联盟结构以满足用户需求;其次,为了得到稳定的联盟结构,引入动态的联盟合并与分裂规则,使得最终稳定联盟中的个体成员无法为提高个体收益而脱离当前联盟结构;最后,为了满足公平性原则,依据个体对联盟的贡献度,提出一种Shapley值法完成联盟总收益在其成员间的分割。实验结果表明,算法不仅可以产生稳定的联盟结构,还可以为个体成员带来更高的收益。
In order to meet user' s requirements and implement dynamic extension of cloud resource capacity, this paper pro- posed a multi-clouds coalition mechanism. First, to maximize the total profit, it modeled the cooperative behavior among multiple cloud resource providers by eoalitional game theory, which aimed to form stable and optimal eoalifional structure of CRPs for meeting user' s requirements. Second, to obtain the stable coalitional structure, it introduced a dynamic coalition merge and split rules, that was, the participating CRPs in the coalition do not have incentives to break away from the coalition. Fi- nally, to meet the principle of fairness, on the basis of an individual' s contribution to the coalition, it put forward a Shapley value method to divide the coalition' s total profit among its members. The experimental results show that this algorithm not only can produce stable coalitional structure, but also can bring higher profit for the participating cloud providers.