An SLA-aware Data Dispersion Method for Short-Secret-Sharing Cloud Storage System

• Main Authors: Yu-TingHuang, 黃鈺婷
• Other Authors: Yau-Hwang Kuo
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/24484662428767207595

碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 101 === In the era of information explosion, people usually share various kinds of information through networks. As more information is going to be shared and stored, the maintenance costs of data also increase. Since cloud storage systems have large storage space, high transmission bandwidth, and the supports of data maintenance, the management costs of clients can further be reduced, and therefore more companies would like to outsource their data into cloud storage systems. However, because of the lacks of physical controls of data, the data security and the data availability are always concerned. Hence, this thesis mainly focuses on proposing a method to leverage between the data security, the data availability, the access performance, and the client cost. Nevertheless, clients often cannot be served with what they actually want because the Service Level Agreement (SLA) provided by existing cloud storage systems is not customized. Therefore, this thesis first defines the corresponding SLA elements according to the four major concerns of clients. Based on the customized SLA negotiated with clients, their actual requirements can be discovered and exploited to adjust the system parameters for completing services. Among existing cloud storage systems with the data security support, Short-Secret-Sharing (SSS) cloud storage system has the characteristics including the smallest storage consumption, the highest performance and key management supports. By adjusting the parameters (n, k) of SSS, the cloud storage system can provide different levels of protection in the data security and the data availability whereas obtaining different access performance and client costs. This thesis would like to propose a method to obtain the optimal result by adjusting the parameters (n, k) based on the customized SLA of clients. Finally, by analyzing several scenarios, the customized SLAs negotiated with different clients are obtained. Moreover, optimal results are presented when clients are served by different cloud storage systems.