Research On Trust Evaluation Technology For Cloud Computing

After being proposed by Google in 2006, cloud computing has drawn more and more attention from academies and industries. Afterwards, Google and IBM begin to popularize cloud computing in American universities, and other IT giants also follow up and start their own cloud computing programs. It is predicted by IDC that the investment in cloud computing will reach to 42 billion dollars in 2012.In cloud computing, large amounts of cheap computers are bundled into a resources pool, which is delivered to users through networks and can be scaled down or up dynamically according to users'demand. This paradigm not only reduces the IT costs and thresholds for users, but also provides them with an ideal solution to deal with large amounts of data. However, lots of companies and individuals are still reluctant to transfer their business or data to cloud in spite of the advantages brought by cloud computing, because they are afraid of the unpredictable security problems caused by the lost of the control of infrastructure and data. Although lots of security measures, such as encryption, authentication and authorization, can be used to ensure the security of the data, it is still an unresolved problem whether they can get rid of uses'doubts and obtain their trust.In order to solve this problem, this paper proposes a trust evaluation and negotiation method to construct trust relationships between users and providers. Specific works includes:1) Analyze related concepts, such as trust, reputation, credit and their differences, and research on the application of trust in E-Commerce and security;2) Propose a method of evaluating the trust degree of cloud computing services based on their Service Level Agreement and other users'feedback to help new users choose the service that they trust accordingly;3) Propose a method of evaluating the trust degree of users by providers based on users'behaviors, service cost and service time. The degree of the trust relationship can be referred by access control mechanism of cloud service;4) Develop a trust evaluation tool based on the research, and verify the effectiveness of the method and tool through experiments.Compared with existing trust evaluation methods, this above-mentioned method takes full account of the effects of the transaction amount and service time on the accuracy of evaluation results and thus renders them more scientific and accurate. Besides, it also takes users' individual needs into consideration in the computation of trust degree of cloud services.