A NOVEL CUSTOMER-ORIENTED JOINT NEGOTIATION MECHANISM FOR CLOUD RESOURCES RENTING

• Main Authors: Shin-Tsung Lee, 李信璁
• Other Authors: Chia-Ying Tseng
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/30303596836220373747

博士 === 大同大學 === 資訊工程學系(所) === 103 === Cloud Computing is a popular topic in recent years, it is different from Distributed Computing, Grid Computing or Parallel Computing. The customers of cloud computing only pay their attentions to the services they requested, and they do not consider execution platform and infrastructure about their services. There are a variety of dynamic or static allocations of resources or energy savings about cloud resource management, but resource providers usually do not consider the user's quality of service (QoS), because it is difficult to use a single mechanism to satisfy all the different QoS requirements of users. The cloud resource supply markets may consist of few providers composed currently, but there may be more providers to provide services for the cloud users in the future. How to increase the loyalty of original users and attract new users to rent is a very important key to maintain the benefit of cloud resources' supply. In our research, we applied game theory to construct a customer-oriented joint negotiation mechanism of cloud resource renting for decision-making, such that the cloud users can negotiate with resource providers without adjust the content of the work and the constraints of deadline. By applying the proposed mechanism, the final results of negotiation can let users to obtain extra remuneration and allow resource providers to have more complete resource section for new users to rent. In our simulation, the results of negotiation for all users compared to the original solutions obtain the better recommendation and increase resource providers 25.93% of the full time sections for new users to rent. Although the context of different workloads could bring the different benefit, but the final decision recommendations provided can produce better rental decisions compared to the pre-existing history of rental record for cloud environment.