Auction based resource allocation is a major challenging problem for cloud computing. However, the existing research is based on untruthful, single resource, single requirement for the premise. In this paper, a truthful auction mechanism is designed for Virtual Resource Allocation and Payment (VRAP) in cloud computing. In this mechanism, users can submit multiple requests at one time, but only one request can be satisfied, known as multi requirements single mind. It is proved that the resource providers can obtain more social welfare under this mechanism than before, and it can guarantee the users' bids are truthful. The mechanism is still compatible with the traditional auction which the user can only submit one request. For the resource allocation problem, a heuristic algorithm is proposed to get the allocation result in a short time, through the reallocation strategy, the social welfare of the cloud resource provider can be maximized. The payment algorithm takes into account critical value to ensure that the machnism is truthful. In the experiment, it is analyzed in terms of social welfare, execution time, resource utilization and so on. Experimental results show that the proposed scheme has good effect for virtual resource action.%使用拍卖方式来进行资源分配可以使得资源提供商获得更大的收益,是云计算领域近年来研究的重点.但现有研究多是基于非可信、单资源、单需求的前提.该文提出一种基于拍卖方式的云计算虚拟资源分配和定价机制(VRAP).这种机制的特点在于,用户在一次拍卖中可以提出多个资源需求.证明了在这种机制下,资源提供商可以获得较以往拍卖机制更大的收益,同时能够保证用户出价是可信的.进而在具体资源分配问题上,提出一种单调的启发式算法能够在很短时间内计算出分配结果,通过资源稀有度概念设计了再分配策略,可以保证云资源提供商的收益极大化;在支付价格计算算法设计中,基于临界值理论计算支付价格,从而保证机制的公平可信.在社会福利、执行时间、资源利用率等多个方面对VRAP进行了测试分析,取得了很好的效果.
展开▼
