Providing a new approach to increase fault tolerance in cloud computing using fuzzy logic

Abstract

ABSTRACT It is generally accepted that with the ever-increasing need of users to various resources, cloud computing is rapidly evolving as one of the new practical technologies. Cloud computing can be categorized as a computing solution in which required technology or services allow users to access computing resources on demand. Moreover, fault tolerance is one of the major concerns to ensure availability and reliability of services as well as to perform the tasks. In order to minimize the impact of failures on the system and ensure correct task execution, they must be anticipated and managed. Few newly developed methods for fault tolerance have focused on fault detection dimension. Therefore, in this paper, a detailed analysis of the nature of the error and its detection will provide as well as a fuzzy-based method to prepare an appropriate response to the error tolerance. In order to increase the error tolerance and load balancing when the error occurs, the requesting a task re-execution and migration techniques through the checkpoint are used. The migration technique overlaps with time, so check-point use can avoid re-execution as well as tasks as much as possible. The results of the experiments also indicate a mean superiority of 6.49% for accuracy criterion in comparison with ABFT method and 2.27% in comparison with FFD algorithm.