Research on Information Security for Cloud Computing

This project aims to study information security issues related to cloud computing, and intends to propose new methods and accordingly make adjustments to the Service Level Agreements (SLAs) between the two sides of cloud computing--service provider and user. Kandukuri et al. (2009) point out the difficulties in ensuring information security for cloud computing. It is essential that the service provider must clearly and comprehensively state its security policies in the SLA. The SLA, however, remains hard to be accepted by users if their critical information is stored in the cloud side. Therefore, this project intends to propose new approaches for privacy protection that will enhance user acceptance toward the SLA. Furthermore, digital signature is an important method for inspecting information integrity and achieving signatories’non-repudiation. It is therefore necessary to study methods that use two- and mutiple-stage digital signature, which are suitable for user-end devices with low computing capability. This project further suggests that a business model that eliminates the use of public key certificates for Public Key Infrastructures (PKI) is more suitable for cloud computing. The expected study period is three years. In the first year, this project plans to investigate the information privacy issues and present new methods. Two basic ideas are: (1) users control their information privacy by themselves through controlling critical information for privacy protection such as cryptographic keys, and (2) when the user-side devices have insufficient computing capabilities, another cloud service provider should offer critical information security services. This provider must coordinate with the application service provider with regard to information security functions. In the second year, this project plans to study the methods of two- or multiple-stage digital signature so that only part of the signature computation is executed at the user end, with the rest executed by the cloud side. Alternatively, a specialized information security cloud service can be set up for computing the remaining partial digital signatures. In the third year, this project plans to study a business model that eliminates the use of public key certificates for Public Key Infrastructures (PKI) and that, as a result, does not require the user end to execute complicate computations. Last, this project plans to make adjustments to SLAs according to our research results. Exemplary SLAs including the adjustments, which are expected to bring about better user acceptance, will also be presented.

Project ID:PF9907-7885
External Project ID:NSC99-2410-H182-025-MY2