What are you looking for ?
DataOn
Komprise

R&D: Modeling and Analyzing for Data Durability Towards Cloud Storage Services

Approach validated to be effective through series of quantitative evaluations in simulation environment

Lecture Notes in Computer Science book series has published, in ICA3PP 2020: Algorithms and Architectures for Parallel Processing proceedings, an article written by Feng Jiang, Yongyang Cheng, Zhao Hui, China Telecom Cloud Computing Corporation, Beijing, China,and Ruibo Yan,State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing,China.

Abstract: Based on the consideration of economic cost and system performance, the distributed storage technology using multiple data replicas has been widely applied in cloud storage applications. This redundant storage mode could ensure that the data loss event occurs only when all data replicas deployed on disks are damaged. In this case, the data durability is determined by the failure recovery model and replica organization strategy. However, the traditional approaches have poor performance in resisting data loss when associated failures occur. In this paper, we propose a novel modeling and analyzing approach for data durability towards cloud storage services. Initially, we model the processes of data failure and recovery using a no-retrogressive Markov chain. Furthermore, we present a routing table-based replica organization strategy to reduce the data loss caused by associated failures. Finally, our approach presented in this paper has been validated to be effective through a series of quantitative evaluations in the simulation environment.