Cloud computing is a calculation model, which provides dynamic scalable virtual resources by the Internet as a service. With an increasing number of large-scale IT enterprises begin to provide cloud resources, the size of the cloud service markets have a sharp growth. At the same time, the cloud resources trading has been the main method of achieving the cloud resource supply and hire between cloud resource providers and users. In the open and dynamic cloud service markets, establishing what kind of negotiation mechanism for cloud resources trading, which can effectively support the cloud resources trading between the cloud resource users and providers, and giving a suitable strategy of price and penalty for break contracts have been becoming a pressing problem.On the basis of the system analysis of cloud resources trading mechanism, combined with the characteristics and influencing factors of cloud resources trading, for the problem that the study of negotiation mechanism for cloud resources trading has not yet considered how set reasonably the contract price and penalties for breaking contracts, the paper proposes a kind of auto-negotiation mechanism for cloud resources trading, in which buyers and sellers negotiation on the contract price and penalty at the same time. The simulation experiment results show that our negotiation mechanism effectively improve the social benefits and transaction success rate of trading of cloud resources trading. For the problem of the cloud resources trading, the paper explores and studies of the following aspects:Firstly, the paper has studied on the literature of home and abroad, and according to the characteristics of the cloud resources trading, has analyzed the defects of applying fixed pricing and auction mechanisms in cloud resources trading, and then worked out that auto-negotiation mechanism based agent is more suitable for cloud resources trading.Secondly, through in-depth studied on the negotiation mechanism of cloud resources trading, in the negotiation of cloud resources trading, the paper found that the existing researches mainly considered sole attribute of time slot or multiattribute negotiation of cloud resources which consisting of price and time slot, SLA etc., and has not yet consider how set reasonably the contract price and penalties for breaking contracts.Thirdly, The paper has overall considered the factors impacting on the contract price and penalty for breaking contract, including cloud resources competition and cost, trading of multi-resources, selfness as well as breaking contract, and improved the alternate-proposed negotiation protocol, and used the technology of auto-negotiation based-agent and presented a kind of auto-negotiation mechanism for cloud resources trading, in which buyers and sellers negotiation on the contract price and penalty at the same time.Finally, through the simulation experiment, the paper has compared the fixed pricing mechanism and our negotiation mechanism on breaking contract, with the change of the three aspects:supply and demand of resources, the number of resources demanded and the flexibility of the task execution, comparatively analyzed the changes in social benefits and success rate of resource trading. The final experimental results show that our negotiation mechanism of cloud resources trading based on breaking contracts negotiation better than fixed pricing mechanism, it is able to achieve higher social benefits and success rate of resource trading. |