Cost-driven scheduling method for workflow with deadline constraints in cloudy environment

Abstract

The invention relates to a cost-driven scheduling method for workflow with deadline constraints in a cloudy environment in the field of parallel and distributed high-performance computation. According to the method, adjacent subtasks with 'directed cut edges' are iterated and combined according to structural characteristics of the workflow to reduce the algorithm execution time and preliminarily reduce the data transmission cost; a partial critical path finding policy for the workflow is designed to ensure that the deadline constraints and service quality demands of the workflow are met; a 'critical' path is found and integrally scheduled based on partial critical paths to further compress the data transmission time and the execution cost; and according to characteristics of the cloudy environment, a scheme of mapping from the 'critical' path to instance execution is designed to ensure that the task service quality is satisfied and the execution cost is reduced. The scheduling method can be effectively improved in execution efficiency on the premise of meeting deadline constraints of existing real workflow, and greatly reduces the execution cost of the workflow in the cloudy environment.

Description

technical field [0001] The invention belongs to a workflow scheduling method in the field of parallel and distributed high-performance computing, and specifically relates to a scheduling method based on cost-driven and local critical path characteristics of a workflow with deadline constraints in a multi-cloud environment. Background technique [0002] With the popularization of cloud computing technology in recent years, many different cloud service providers have appeared in the current cloud market, such as GoGrid, AmazonEC2 and Rackspace. According to the differences in the content of specific services, service types can be divided into: Infrastructure as a Service (Infrastructure as a Service, IaaS), Software as a Service (Software as a Service, SaaS) and Platform as a Service (Platform as a Service, PaaS). End users guarantee the service quality of their own applications through service level agreements (ServiceLevelAgreements, SLAs) concluded with service providers, a...

AI Technical Summary

Problems solved by technology

However, many research works only pursue cost optimization in a multi-cloud environment, while ignoring other constraints, such as deadlines; or only consider workflow cost optimization scheduling with constraints in a single-cloud environment, and do not involve multi-cloud environments with different bandwidths. The impact on workflow execution cost; or ignore the complex dependencies between tasks and only consider the batch task scheduling problem in a multi-cloud environment; or in the process of pursuing the optimal workflow cost, do not consider the basic principle of cloud environment pay-as-you-go and price-by-range nature

Therefore, in a multi-cloud environment, the optimal scheduling problem of large-scale workflow execution costs with deadline constraints has not been properly resolved

Images