Cloud computing platform-oriented scientific workflow task scheduling method

Abstract

The invention provides a cloud computing platform-oriented scientific workflow task scheduling method. The method comprises the steps of modeling a constraint relationship between tasks as a DAG graph; converting the DAG graph into an in-tree structure task diagram; constructing a scheduling set; adjusting a scheduling set Pn of anode tasks En, wherein the number of scheduling combinations in thescheduling set Pn is the required number of virtual machines; and assigning each scheduling combination in the scheduling set Pn to a corresponding virtual machine for execution. The cloud computing platform-oriented scientific workflow task scheduling method has the advantages that the communication overhead between workflow tasks is reduced through a task replication technology, the task scheduling length is minimized, and the efficiency of task scheduling is effectively improved; through a task grouping technology, the number of virtual machines used is effectively reduced; and by rationally utilizing the idle time of the virtual machines, the utilization of virtual resources is improved.

Description

technical field [0001] The invention belongs to the technical field of scientific workflow-oriented task scheduling in cloud computing systems, and in particular relates to a cloud computing platform-oriented scientific workflow task scheduling method. Background technique [0002] As a hot issue in recent years, cloud computing has attracted more and more researchers in academia and industry. Cloud computing technology is developed from distributed computing, parallel computing, and grid computing. It uses virtualization technology to virtualize computing resources, storage resources, and bandwidth resources into a dynamically scalable virtualized resource pool. , provided to different users on demand in the form of services. [0003] Scientific workflow refers to turning a series of data management, calculation, analysis, display and other tasks encountered in scientific research into independent services, and then combining these services through data links to meet the n...

AI Technical Summary

Problems solved by technology

But this algorithm does not take into account the number of virtual machines used

(2) The OSA algorithm assigns the parent task node and the child task node to the same virtual machine as much as possible, so that the current task can obtain the earliest start time, but this algorithm is only limited to the local area, and does not consider other ancestor nodes from the global perspective. This is limited to minimize the makespan of the entire task

(3) The CPTD algorithm first converts the DAG task graph into a corresponding product processing tree, then finds out the critical path, and shortens the execution time of all task nodes by minimizing the completion time of the task nodes on the critical path, but the algorithm high complexity

However, the CPFD algorithm does not consider the optimization of the number of virtual machines, which takes up more virtual machine resources.

[0007] Therefore, the existing various task scheduling methods cannot meet people's needs for minimizing the completion time of scientific workflows and minimizing the number of computing resources used.

Images