Improving scheduling efficiency by probabilistic execution time model in cloud environments

Recently, cloud computing has become a promising paradigm for various kinds of large-scale applications. Due to the unpredictable characteristics of resource availability and workload intensity, execution latency still drastically impairs the performances of cloud applications. In this paper, we model the execution latency by a probabilistic distribution and propose a general task execution model which can be used in most of scenarios. By using the proposed execution time model, cloud administrators can easily refine their resource management or implement some fine-grained task scheduling policies for cloud applications in various cases. Massive experiments are conducted in a real-world cloud platform, and the results indicate the proposed model can be used in many existing scheduling policies for improving the efficiency of task execution.