Research On Job Scheduling Strategy In Cloud Computing

Cloud computing, which is a new computing model and provides dynamic and on-demand service to users in the form of a data center, has caught the scholars and companies’ attention. The job scheduling is one of the key technologies in cloud computing, it is of significance to meet users’ needs and improve service quality and economic benefit of cloud service providers. However, domestic and foreign scholars’ study on job scheduling is not sufficient, some studies only from the point of user, some only from the perspective of cloud service provider,consider single objective optimization, and some consider job scheduling from the perspective of both but with incomprehensive considerations. Meanwhile, admission control refers to a strategy of whether to receive the arrival job request for the cloud computing data center, it’s an effective mechanism to avoid resource overload in cloud computing but with less studied.To solve the above problems, this thesis will carry out the research on job scheduling in cloud computing. The main work:1. From the user and cloud service providers’ perspective,a job scheduling algorithm based on improved ant colony algorithm has been proposed in cloud computing to minimize the total cost.Besides achieving this goal,it takes into account not only the user’s quality of service(Qo S), such as job finishing time and expense,but also virtual machine’s load balancing.2. For these cases:there are a lot of job requests,the deadline requirements are tight or there are less resources in the cloud computing data center,this thesis has proposed a job scheduling strategy based on admission controlling.The main goal of it is by increasing the number of job requests to receive,namely job request throughput, to maximize profits of cloud service provider.At the same time,the strategy uses utility computing’s penalty mechanism,priority-based preemptive strategy,resources’ scalability in data center.3. Use Cloud Sim’s internal way to modify its source code to program,realize the job scheduling algorithm based on improved ant colony algorithm and the job scheduling strategy based on admission controlling.Use the Cloud Sim platform to simulate,firstly compare the job scheduling algorithm based on improved ant colony algorithm with job scheduling algorithm based on base ant colony algorithm,then validate the job scheduling algorithm based on admission controlling and make comparative experiments.Experimental results of the former show that the proposed algorithm in this thesis has lower total cost, the job completion time and expense cost are lower, and when considering data center resources’ load balancing, resource load balancing also has some degree of improvement; the latter comparative experiments have demonstrated the effectiveness of penalty mechanism and preemption strategy,namely increasing the throughput of job requests and improving profits of the cloud service provider.