Accountability in Cloud Computing and Distributed Computer Systems

Accountability in Cloud Computing and Distributed Computer Systems Hongda Xiao 2014 Traditionally, research in computer security has focused on preventive techniques such as passwords, authentication protocols, and encryption. With the rise of internet technologies, especially cloud computing and distributed computer systems, preventive measures have proven inadequate, and security researchers have sought to complement them with accountability mechanisms. Despite widespread agreement on its importance, “accountability is not yet a unified research area. This thesis advances the state of accountability research by systematically comparing a significant amount of existing work in the area and designing practical accountability mechanisms for realistic scenarios in cloud computing and distributed computer systems. First, we propose a framework to categorize research on accountability mechanisms with respect to time, goal, information, and action. Our systematization effort shows that more sparing use of the word “accountability” is desirable, because it is currently used by different researchers to mean different things. Our conception of the term dispels the mistaken notions that accountability precludes anonymity and privacy and that it requires centralized authority. Second, we present a privacy-preserving structural-reliability auditor (P-SRA) for cloud-computing systems. P-SRA enables evaluation of the reliability of cloud infrastructure without compromise of cloudinfrastructure providers privacy. We present the privacy properties of P-SRA and evaluate a prototype implementation built on the Sharemind SecreC platform [BK13]. P-SRA is not only a mechanism for holding cloud-service providers accountable but also an interesting application of secure multi-party computation (SMPC), an extensive body of privacy technology that has not often been used on graph problems, which are inherent in structure-reliability auditing. Third, we extend our study of the accountability of cloud-service providers to cloud-service users; rather than focusing only on reliability, we consider general properties of cloud infrastructure. We develop the