A Genetic Programming Based Scheduling Approach For Hybrid Community Cloud

Hybrid cloud framework combines the advantages of private cloud and public cloud and it has been widely applied now.So the research of hybrid cloud is very significant.This paper proposes a new cloud resource sharing framework based on the disadvantage of the traditional hybrid cloud.The framework consists of a community cloud combined by multiple private clouds and multiple public clouds.In the community cloud,private cloud owners strip out the resources that have a high idle rate to form the community cloud,where users can gain profit by rent resources to others.Community cloud can provide cheap computing resources to users in the community.When the resources in the community cloud cannot meet the demand of computing peak,the framework will then ask for resources in the public cloud to meet the computing demand.To describe the workflow scheduling problem in hybrid cloud,this paper establishes a mathematical model to minimize total cost within the deadline of workflow application.To solve the scheduling problem,a two-stage algorithm is proposed.In the first stage,a resource reservation algorithm is proposed to reserve resources in the community cloud.Then,the problem is transformed to the traditional scheduling problem.In the second stage,we choose genetic programming algorithm(GP)to get the scheduling rule on simulator,and scheduling with the best rule.Finally we designed an experiment to verify the effective of the new framework.We designed experiments to verify the effective of the algorithm compared with rule based scheduling algorithm and the generalization ability of the algorithm compared with PSO based scheduling algorithm.