Font Size: a A A

Research On The QOS-aware Service Selection Approach For The Multi-tenancy SAAS Application

Posted on:2014-02-25Degree:MasterType:Thesis
Country:ChinaCandidate:Y L LiuFull Text:PDF
GTID:2248330398460994Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
With the development of SOA (Service-oriented Architecture), SaaS (Software as a Service), as a new service delivery model, which has the advantage of on-demand using, easy to maintenance and extend has changed the way people using software and the way ISV providing software. Multi-tenancy is the core technology to SaaS Application model, which can host a large number of tenants on the same service instance and hardware infrastructure.QoS (Quality of Service) is a description of the services’ non-functional properties, which can determine the effect of a service to complete a business qualitatively or quantitatively. It’s a complex structure that contains multi-dimensional indicators, and it can be used to distinguish homogeneous services, also is an important standard to process service selection.Broadly, service selection is a process including service selection and service ranking, which is relatively independent. Service selection is to obtain a set of services that can fulfill tenant requirements from service library, and service ranking is a optimal selection on the above basis. Currently, an urgent problem faced by QoS-aware service selection is how to combine the characteristics of multi-tenant applications to conduct the process based on a detailed QoS description model.In order to solve the above problems, first, a extended QoS model is presented from the view of the operator and tenant, and also a service library comprised by service asset library and case library, which can provide detailed information for subsequent selection and ranking.When conducting the selection process, item-based collaborative filtering technology is adopted. If there is a match hit, the matched service returns to the tenant directly. If fails, especially for some composite services with complex path structure, the entire execution plan is decomposed into fine-grained fragments. When resolve the emerging service selection problem, through reusing existing cases, the execution path is adjusted to downgrade the problem size and reduce the complexity. For the adjusted execution path, the Genetic Algorithm is used to for an execution plan meeting users’ requirement.With regard to service ranking, a set of functions is established to determine service satisfaction for the tenant and operator respectively and to combine their satisfaction as an overall score based on certain strategy for service ranking, which in order to return the final service.Practically, this paper develops a prototype system to verify the research work under the background of supplier business management (SBM) service, and verify the validity of our approach.
Keywords/Search Tags:Multi-tenancy, Path decomposition, Service selection, Service Ranking
PDF Full Text Request
Related items