Efficient Service Re-composition Using Semantic Augmentation for Fast Cloud Fault Recovery

Today, the success of a cloud system significantly relies on its availability and reliability. To achieve a better availability, fault recovery must be performed in an efficient way. When no backup is available, service re-composition becomes the only option. Although many semantic approaches for service re-composition have been proposed, they all seem to be time-consuming, and service granularity still remains as an open question. In this paper, we introduce an efficient service re-composition approach using semantic augmentation and a fast service granularity selection so that a faulty service can be replaced promptly. A semantic service matching mechanism is also proposed for discovering appropriate candidate services for re-composition. The experiment results from a cloud environment confirm the advantages of our approach.