Research on dependability of cloud computing systems

TLDR: A generic definition of cloud computing system dependability, which including availability, performability, security, recoverability and so on is provided, and a novel simulating approach is proposed with well-designed framework and procedure.

Abstract: With the continuous growth of application requirements and a significant advance in the research of cloud computing systems, a large number of cloud computing systems nowadays based on different structures and virtualization technologies are still being developed. However, dependability of cloud computing system is always a critical issue for all the cloud service providers, brokers, carriers and consumers around the world. How to ensure the dependability of cloud computing systems is still not well solved by former researchers. This paper first introduces the infrastructure of a cloud computing system and its major actors. Based on the research of the running mode of a cloud computing system, main influencing factors of cloud computing dependability and its normal failure modes are given. It provides a generic definition of cloud computing system dependability, which including availability, performability, security, recoverability and so on. We also discuss some methods to establish cloud computing dependability models, such as analytic method, state space method and simulating method. In order to evaluate the validity of the cloud dependability measurement method, a novel simulating approach is proposed with well-designed framework and procedure. Through reasonable simulating and analysis, proper solutions can be promoted to find and solve the problem related to cloud computing systems dependability. Our future work will focus on an integrated software platform for cloud computing system dependability simulating and verification.