Dynamic Workflow Scheduling based on Autonomic Fault-Tolerant Scheme Selection in Uncertain Cloud Environment
Chenyang Zhao, Junling Wang · 2020
Most of the existing methods cannot deal with the problem of fault-tolerant workflow scheduling in uncertain environment. Therefore, the paper proposes a dynamic workflow scheduling based on autonomic fault-tolerant scheme selection in uncertain cloud environment. In order to make full use of various cloud resources, the tasks that constitute workflow are divided into various types. And a computing capacity fluctuation model based on multiple states and jump points is established to describe the instability of cloud resource. Moreover, an extended fault-tolerant model is built to deal with various computation and transmission faults. Based on the two models and task classification, tasks are firstly assigned statically to appropriate virtual machines, then the fault-tolerant strategies and the actual execution for tasks are automatically adjusted in a dynamic manner. Experiments show that the proposed method can perform effectively in uncertain cloud environment while ensuring fault tolerance.