Design And Simulated Implementation Of Resource Double Multi-Attribution Auction Mechanism Under The Circumstance Of Cloud Computing

As a new emerging technology, the cloud computing is a product of development and integration of distributed computing, parallel computing, grid computing and many other technologies. It considers computing as a kind of service. Through the high-speed data transmission ability of Internet, it can transfer data from terminal to computer cluster in cloud environment to deal with it. The entity which can be used through Internet in the cloud computing environment is called cloud resource. The emergence of cloud computing can integrate all kinds of cloud resources effectively and reasonably and make them used fully. It is necessary to study a kind of trade mechanism which can urge the resource provider of the cloud resource to provide resource and resource consumer to purchase resource.The thesis puts forwards cloud resource double multi-attribution auction mechanism through analyzing current popular auction mechanism, choosing the continuous double auction mechanism as the basis of auction mechanism design and combing the qualities of the cloud resource and the double characteristics of resource provider and resource consumer. The main concerns and achievements of this paper include the following six aspects:(1) Study the characteristics of the cloud resource, resource provider and resource consumer; putting forward the detailed description methods, the depict methods of the behavior of resource provider and consumer and the depict methods of the tender of resource provider and consumer;(2) Study and confirm the multi-attributive characteristics of the resource provider and consumer. Through improving error back propagation algorithm, it transfers the other attributions of resource provider and consumer except price to quality index. Based on improving error back propagation algorithm, it puts forward the multi-attributive processing methods of resource provider and consumer;(3) Research and confirm the factors in need of considering when providers and consumers of resources predict the price, realize their prediction of price by improving the SVM algorithm and put forward their methods of predicting respectively;(4) Research and confirm the measuring mechanism of allocation performance in the process of cloud resources allocation, realize cloud resources allocation optimization by improving the MVO algorithm and the WDO algorithm and put in the ways of resources allocation on the basis of improving MVO algorithm and WDO algorithm;(5) On the basis of continuous double auction mechanism, raise cloud resources bidirectional multi-attribute auction mechanism by combining the contents of the former four points;(6) Use Java language to realize cloud resources bidirectional multi-attribute auction mechanism based on the GridSim simulation platform, and do a test and valuation for the performance and the functions supposed to be realized in auction mechanism.