Ensuring Secure and Reliable Interactions between Mobile Applications and Cloud Services

点击翻译按钮获取中文摘要

Motivation Modern mobile apps are commonly integrated with cloud services to provide powerful computation capabilities. Mobile apps frequently exchange data with the cloud services. We name such exchanges as mobile-cloud interactions. It is challenging to build secure and reliable mobile-cloud interactions given the complicated runtime environment and the short release cycle of mobile apps. Software bugs in mobile-cloud interactions can significantly jeopardize the security and reliability of mobile apps. For example, it was reported that the misuse of cloud service credentials in mobile apps has induced massive user private data leaks. Studies were conducted to design secure communication protocols or detect malicious apps leaking user information. However, none of them studied issues in mobile-cloud interactions from the software engineering point of view. Objectives In my research career, I aim to improve the automation level for the QA practice to facilitate the identification of software bugs and reduce the QA costs. In the short term, I will focus on advancing software engineering practices for mobile applications (mobile apps) in the cloud computing era. Specifically, I aim to (1) characterize the diverse mobile-cloud interaction issues in practice, (2) develop effective techniques to identify security flaws in mobile-cloud interactions to protect the mobile apps and their users, and (3) propose novel test coverage criteria and test oracles that can effectively expose reliability issues in mobile-cloud interactions in an automatic manner. Team and Approach In this program, I plan to recruit two PhD students, two MEng (thesis-based) students and four BEng summer interns to achieve the short-term objectives. We will 1) construct and publish a large benchmark of mobile-cloud interaction issues, 2) leverage open coding method to analyze the benchmark and gain a comprehensive understanding on mobile-cloud interaction issues, 3) devise novel models to abstract mobile-cloud interactions and capture the concerned issues, and 4) leverage the models to design automated techniques to detection the issues. Outcome and Impact This program will generate the first public benchmark and empirical study of mobile-cloud interactions. We will also publish a set of techniques to automatically detect mobile-cloud interaction issues. The outputs can benefit both the research and the industrial communities. The study will break new ground for software engineering research on building reliable mobile cloud computing infrastructures. The techniques will increase the automation level in quality assurance process in mobile development companies in Canada, reducing their cost in developing and maintaining trustworthy mobile apps in the cloud computing era. The program will also train HQPs with technical and research skills that are highly demanded in Canada.