The Research Of Cloud Workflow System Resource Allocation Based On Combinatorial Reverse Auction

Cloud computing is an emerging business computing model, its basic idea is to transmit all kinds of resources through the Internet, such as:storage resources, computing resources, bandwidth and so on. So users do not need to purchase a large of computing systems to manage their business, on the contrary, they only need to purchase the resources according to their needs, so that it can reduce the cost greatly.Cloud workflow is the workflow management system in the cloud, which can optimize the cloud computing costs and improve performance. During the execution of cloud workflow, a lot of resources are needed for its services. And an effective resource allocation scheme can effectively improve resource utilization and reduce the execution time of workflow.However, as users’ demand for cloud resource increasing and the number of cloud providers increasing, the traditional cloud resource allocation scheme failed to meet the dynamic change of the cloud resource market and the competition between providers. So users can’t get favorable resources and providers cannot adjust the price according to the dynamic changes of the market in a timely manner either, which reduce the efficiency of the deal. In addition, in the cloud market the users’ demand for resources is diversity, and there are also different types of resources provided by provider. In traditional resource allocation scheme, only one resource is allocated in an auction, which increases the cost of resource allocation and reduces the efficiency of the resources allocation.The market-oriented reverse auction method is efficient and economical to allocate resources for cloud workflow systems since it dynamically allocates resources depending on the supply-demand relationship of market. However, during the process of auction the cloud resource price is usually fixed, and the resource allocation mechanism cannot adapt to the changing market properly. This results in low efficiency of resource utilization. In this paper, a dynamic-pricing reverse auction based resource allocation mechanism is proposed to improve resource utilization. Firstly, a dynamic pricing strategy in reverse auction is proposed, and then an efficient resource allocation mechanism based on dynamic pricing reverse auction is presented. During the auction, resource providers change prices according to the trading situation so that our novel mechanism can increase chances of making a deal and improve efficiency of resource utilization. In addition, providers improve their competitiveness by lowering prices. Simultaneously users are easier to get resources with the increase of deal. Therefore, users can obtain cheaper resources in shorter time, which would decrease monetary cost and completion time for workflow execution.When there are different types of tasks in workflow, combinatorial auction can improve the efficiency of auction. During workflow execution, the user select the task which is served according to the partial relation of workflow t, and only after the execution of parent nodes, the child task begins to execute. For separate auction, only one task is in each auction, and thus increases the time of the auction greatly. In combinatorial auction, the auction order task is the partial order relation of workflow tasks, and tasks in a separate auction is brother node can be grouped together for auction, an auction can has multiple tasks, thus greatly reduces the task completion time, improve the efficiency of the auction, so the completion of the entire workflow execution time and cost of product is lower than the separate auction.In the market of cloud resources, there is also fierce competition between resource providers. To make their own benefit maximization, the providers must improve the competitiveness of the resources and the probability of auction winners. Because changing the resource price is the most direct and most effective to improve the competitive ability of the resource, so the combination reverse auction is proposed based on of dynamic pricing. The providers change the current resource prices according to the situation of the last round of the auction. If a provider lost in last round of auction, he will decrease the price of resources which can improve the competitiveness of resources. Otherwise, he will keep the resource price unchanged. In addition, due to the competition of the provider, the better resources will emerge and users will obtain cheaper resources for its service, thus reducing the completion time of workflow.Finally, the simulation of resource allocation is giving in the end of every article to implement the giving algorithms as well as the comparison algorithm respectively. Experimental results confirm that resource allocation scheme based on reverse auction has higher market efficiency, and the combination of the strategy is applied to make the resource allocation efficiency in the allocation of resources was improved, and user costs have been reduced.