A Malicious-Resilient Protocol for Consistent Scheduling Problem in the Cloud Computing Environment

[[abstract]]In the cloud computing environment, resources are distributed among processors of a computing cluster. When performing a task, processors have to request needed resources that do not exist in them from other processors in the network. If the requested resource is currently used by another processor, they have to wait until the resource is released. If the resource is not released, a deadlock may occur. Deadlocks can be avoided if the order of jobs is properly arranged in advance. However, there may exist malicious faulty components in the cloud computing environment. Malicious faulty components may disturb job scheduling of fault-free servers, causing them to generate inconsistent schedules. In this paper, we investigate the Interactive Consistency (IC) problem in the cloud computing environment to develop a protocol that can exclude influences from malicious faulty components to help servers achieve consistent job scheduling. The proposed IC protocol is called Efficient Interactive Consistency Protocol with Weight Consideration (EICPweight). In this problem, each server will collect job requests from end-users first, then give a weighted value to each job according to its amount, and finally generate a consistent schedule of jobs based on their weighted values. Moreover, we also propose a new Cutoff-test mechanism to enhance the performance of the proposed protocol.[[notice]]補正完