Research Of Cloud Resource Allocation Based On Continuous Double Auction

Cloud computing is a new business computing model, the commercialization property ofcloud computing makes it focus on the users’ increasingly diversified demands. It aims atflexiblely and reasonably allocating resources according to the needs, improving the datacenter resources utilization. The use of mature virtualization makes resource allocation andscheduling different from the previous parallel distributed computing. Therefore, the study oncloud resources allocation and scheduling can promote the development of cloud computing.This paper adopts the market economics mechanism which mobilizes individualenthusiasm, inspires resources sharing, and realizes the reasonable allocation of resources.Applying continuous double auction in cloud computing with the commercial properties, itdesigns a resource allocation algorithm based on continuous double auction in cloudcomputing. The algorithm considers different cloud users with diversified demands, sets fairand transparent pricing strategies for the resources providers and the user respectively: clouduser agent decides the bid independently based on the available remaining resources and theaverage remaining time in each unit of time, while resource providers proposes the ask priceaccording to their own workload. Cloud users and resources providers submit the bid and askprice respectively to auctioneer, while the auctioneer manages current list of bid and ask priceand matches final deal. In this paper, a few cloud users with high urgent deadline tasks arealso discussed, we assume that these cloud users are allowed to choose and reserve suitableresource providers with the consultation price based on their actual situation. It finally adoptsthe SLA to define the service contract between users and the cloud service providers, specifiesservices level for users to achieve different QoS, and explores the realization of the QoSrequests.After introduction and analysis CloudSim’s system architecture, the technicalimplementation and the concrete use method, the paper expands and recompiles CloudSimplatform to implement the proposed algorithm, makes a simulation experiment andperformance analysis. Comparing with other related scheduling algorithms, it is proved thatthe S-CDA presented by this paper adapts to different users’ QoS at a certain extent and issuitable for open complex cloud computing environment.