Font Size: a A A

Research On Workflow Scheduling Algorithms For Cloud Computing

Posted on:2018-12-11Degree:DoctorType:Dissertation
Country:ChinaCandidate:F H WuFull Text:PDF
GTID:1368330623450328Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
As a new computer system architecture,cloud computing enables solving large scale computing and massive data processing problems.In order to programme this new architecture,workflow scheduling method is used to model a lot of complex applications.Workflow scheduling is becoming an effective programming paradigm in cloud computing environment.Currently,cloud computing workflow scheduling is an important aspect of system and software research.Different from traditional workflow scheduling,some new features of cloud computing bring new challenges to workflow scheduling.These new features include resource dynamic access and supply,availability of resources,performance instability and green cloud computing center.In order to meet these challenges,this topic has explored relevant scientific issues and key technologies.Specifically,following five aspects are focused:First,we select closely related papers from academic,and summarize research works of workflow scheduling in different stages of computer development.Some basic theory,algorithm and practical application of workflow scheduling problem are reviewed.According to the new features of cloud computing,a workflow scheduling model for cloud computing is proposed,and 9 main problems related to workflow scheduling in cloud computing are summarized,and the solutions to each kind of problems are compared and analyzed.Based on the analysis and evaluation of existing solutions,the problems that need to be solved in the future are explored.Second.Compared with traditional workflow scheduling,Cloud Computing enables workflow applications flexibly choose resource according to their scale.This feature freed developers from resource management.However,it also bring new challenges to resource supply and job scheduling.How to select reasonable resources and schedule workflow is a new problem introduced by cloud computing for budget constraint workflow scheduling.In this dissertation,the theoretical optimal budget allocation mechanism of the workflow composed of two special structures is obtained.Based on the theory optimal mechanism,a heuristic cluster workflow scheduling algorithm is proposed.Third,the construction of green cloud computing center has become a consensus in both academia and industry.The research of energy optimization for workflow scheduling is researched using hardware technology of DVFS.A energy aware cloud workflow scheduling model is established.Theory energy optimal scheduling is studied for specially structured workflow.Then,we design an energy aware heuristic algorithm.In addition,this dissertation studies the energy consumption problem caused by workflow data transmission and presents an energy aware data transmission mechanism.Fourth,up to now,most workflow scheduling techniques for cloud computing assume that user pay theoretically for resource they have used really.From the IaaS infrastructure layer that forms the base of cloud computing,we get that current technology can not achieve this target.We use packing method to optimize workflow scheduling problem for IaaS cloud computing.A packing model of workflow scheduling in IaaS cloud computing is definition.A compaction method is designed to optimize already generated schedule,and a deadline constrained cloud workflow scheduling algorithm is proposed.Fifth,complex workflow applications in practice often face with multiple QoS constraints and scheduling objectives.Workflow scheduling method subject to different constraints and objectives are summarized.We design a unified multi-constraint multiobjective scheduling framework.The framework adopts economics pareto optimal theory to evaluate the merits of schedule.Based on this framework,a heuristic algorithm for workflow with execution deadline and budget constraints is proposed to optimize energy consumption and schedule reliability.
Keywords/Search Tags:Cloud Computing, Workflow Scheduling, Budget Constrained, IaaS Cloud Computing, Energy Optimize, Multi-constraint Multi-objective Scheduling
PDF Full Text Request
Related items