Local Reputation Management in Cloud Computing

In the Cloud computing community, the calculation of the reputation using the feedback of cloud customers is widely adopted to address the issue of trustworthiness of cloud services. Currently, most methods pursue a global reputation score essentially assuming that the value of a cloud service´s reputation is the same for every consumer. However, depending on the expectations and needs of a consumer, there can be significant deviation of perceived reputation for the same cloud service. In this paper we propose a trust management framework that differentiates reputation for various user groups thus providing what we term local reputation. To achieve this we compute the similarity of consumers based a decision-tree model which is used to cluster feedback into localised scores. To refine the result, a time decay factor applicable to feedback is also to be considered. The simulation results illustrate that our approach is feasible and also effective for consumers to choose reputable cloud service.