Research On Task Scheduling Algorithms For Cloud Computing Platform

Data storage and computing power are becoming more and more important for artificial intelligence supported by large data.Cloud computing model is an important information infrastructure that can meet the needs of high-performance computing power and massive storage space.More and more data are stored in the cloud,therefore,there are lots of resources and tasks in the cloud computing environment.Reasonable allocation of resources through task scheduling algorithms is an important way to improve the service quality of cloud computing,resource utilization efficiency,task execution efficiency,and reduce user usage costs.It also provides potential possibilities for reducing energy consumption in the cloud.Therefore,it is of great significance to carry out task scheduling algorithm and its optimization method in cloud computing platform.This thesis proposed a cost and energy-aware task scheduling algorithm.First,the cost-aware task scheduling algorithm is proposed.The task model,virtual machine model,scheduling problem model and used cost model are established.Based on DAG(Directed Acyclic Graph),a priority calculation method based on critical path node priority was proposed,and the tasks were sorted into a basic task queue according to priority with the key path as the core.Then tasks are allocated to the virtual machines based on the lowest cost allocation algorithm,the interval insertion allocation algorithm,and the precursor node allocation algorithm.Then,based on the interval insertion allocation algorithm,this thesis further proposed the optimization of energy consumption-oriented scheduling algorithm,and reduced the power consumption of task execution by analyzing the task characteristics.This thesis verifies the proposed algorithm by randomly generating DAG applications and real applications respectively.The experimental results showed that the proposed cost-aware algorithm was superior in both cost and execution time.The low-power optimization based on cost-aware algorithm can reduce the energy consumption of task execution at low cost,and realized the goal of reducing system power consumption.