IEEE Access (Jan 2018)
An Agreement Under Early Stopping and Fault Diagnosis Protocol in a Cloud Computing Environment
Abstract
With its explosive development, the Internet has come to offer an increasing multitude of online service applications. In order to better facilitate such services, a cloud computing environment, composed of a large number of processors and memories, high-speed networks, and various applications, has been developed, and continues to grow, providing convenient and quick network services. In this cloud computing environment, each server processor in the environment must cooperate with other processors to satisfy various user demands. As a result, the issue of fault-tolerance needs to be revised in order to ensure the reliability of cloud computing environments. One of the most important issue of fault-tolerance is the Byzantine agreement (BA), which requires that, even if some components are damaged, a set of fault-free service processors are able to agree on a common value. Furthermore, the faulty service processors must be detected and eliminated. Therefore, a fault diagnosis agreement (FDA) issue for such environments must also be revised simultaneously. The goal of FDA is to enable each fault-free service processor to detect/locate a common set of faulty service processors. However, a lot of messages need to be collected to solve the BA and FDA problem in the previous works. Thus, this paper also uses the concept of an early stopping protocol (ESP) to allow its participants to obtain common values early, during different rounds. Furthermore, the result of ESP can then be used to detect/locate the maximum number of faulty service processors with dual failure modes efficiently, using the minimum number of rounds. As a result, the early diagnosis cloud agreement protocol can be proposed to solve the BA, ESP, and FDA problems simultaneously to provide greater computing abilities by enhancing the reliability of a cloud computing environment.
Keywords
