Font Size: a A A

Research On The Optimal Scheduling Of Multi Tenant Service Requests In Cloud Environment

Posted on:2017-05-01Degree:MasterType:Thesis
Country:ChinaCandidate:B QianFull Text:PDF
GTID:2308330509956907Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the gradual maturity of the cloud computing industry, the software as a service(Saa S) as an important service delivery method is widely used in cloud computing. In Saa S mode, mature service providers usually single-instance multi-tenancy approach to reduce operating costs, which means the same application instance is leased to multiple tenants. The multi-tenant Saa S application mode can provide on-demand rental service for small and medium-sized enterprises, and has broad prospects.In multi-tenant mode, the same application instance is required to be shared by multiple tenants. Each tenant demands for different services. How to design Saa S applications to embody the multi-tenant sharing characteristics and how to meet the individual demand of each tenant while sharing application instances need to research. In cloud environment, how to allocate the corresponding shared resources for multiple tenants to meet the needs of multiple tenants with minimum cost. It’s necessary for service providers to adjust the resource allocation strategy in time according to the dynamic changes of the service request in the runtime. So that the resource utilization and the service quality can be improved. To solve the above problems, this paper researches the multi-tenant service request optimal scheduling in cloud environment, mainly includes the following aspects:(1) In order to meet the needs of multi-tenant application services, this paper will establish a multi-tenant application architecture and design the multi-tenant service request scheduling framework in cloud environment. In the service request scheduling framework, the scheduling is divided into two stages, which are the optimal placement of service requests and the dynamic scheduling of service requests. The scheduling strategy of each stages are analyzed respectively.(2) In the multi-tenant application architecture, service providers need to reduce the number of the component instances, which need to be deployed, to reduce operating costs. In this paper, the service request placement is optimized. The problem model is established the quantum genetic algorithm and the hybrid genetic simulated annealing algorithm are used to realize the service request placement algorithm. By comparing with the traditional genetic algorithm, the solution quality of the two algorithms is analyzed.(3) On the basis of the service request placement scheme, service requests will be dynamic scheduled in this paper. For the changes of the multi-tenant service request at runtime, the dynamic scheduling will maximize the dynamic QoS earnings of the platform. A dynamic QoS revenue model is established to evaluate the validity of the scheduling. The algorithm steps of the service request scheduling algorithm are presented and the analysis is carried out by the comparative experiments.(4) On the basis of the optimization placement and the dynamic scheduling of service request, combine with the multi-tenant application architecture, the architecture, database and multiple data sources of the cloud service platform are designed and developed.
Keywords/Search Tags:cloud computing, multi-tenant, optimize placement, dynamic scheduling, QoS
PDF Full Text Request
Related items