云环境下数据中台能耗感知的微服务任务调度算法

ENERGY AWARE TASKS SCHEDULING FOR MICROSERVICE IN DATA CENTER PLATFORM CLOUD ENVIRONMENT

针对云计算环境下数据中台具有依赖关系的微服务任务调度导致能耗浪费问题,根据虚拟资源节点的能耗表现,构建相应的能耗效益函数,在此基础上,提出一个融入能耗优化的启发式微服务任务调度算法。算法主要由两个子算法构成:任务映射算法和串行任务合并算法。任务映射算法为每个工作流任务选择最优虚拟机类型,串行任务合并算法用于合并工作流中两个满足约束条件的子任务。仿真实验表明,该算法与同类算法相比,在同等条件下具有较好的能耗优化性能。

In order to solve the dependent tasks scheduling problem which did not take higher energy consumption into consideration of data center platform under cloud computing environment, the energy effectiveness function was constructed according to the energy consumption performance of virtual resource nodes. A novel heuristic tasks scheduling algorithm for microservice which considered energy consumption was proposed based on the function. The algorithm consisted of two sub-algorithms: task mapping algorithm and sequence task merging algorithm. Task mapping algorithm was utilized to select optimal virtual machine type for each task and sequence task merging algorithm was used to merge two tasks which satisfies constraints in workflow. Simulation results show that this algorithm has a better performance in comparison with similar algorithms under the same conditions with respect to energy consumption.