Research On SaaS Service Relevance Calculation Method Under Dynamic Evolution Environment

SaaS software creates huge economic benefits for services providers,because of its characteristic of single instance,multi-tenant.When the SaaS software can't satisfy the requirements of tenant or other reasons,it is necessary for SaaS software to have an activity of evolution.Because of the characteristic of single instance,multi-tenant,the sharing of runtime is increased.All of services,that tenants require,are integrated in the same SaaS software.And multi-tenant is consecutively severed in the same time.So,in this pattern dynamic evolution needs to be implemented to the SaaS software otherwise plenty of users would be influenced.Thus,it affects the interests of the providers and causes economic losses.To ensure the reliability of evolution,and at the same time reduce the evolution cost,impact scope needs to be controlled as far as possible.Determine the boundary of impact scope.At the beginning,analysis the relevance of service to get the minimal set of relevance services so that relevance services can be derived to the static state to reduce the evolution cost and ensure the reliability of evolution.It is non-ignorable for controlling the boundary of impact scope to calculate the relevance between the services.In order to quantitatively describe the degree of the relevance between services,it is necessary to calculate the relevance degree.So,it is important to solve the problem of calculation of relevance degree before implement dynamic evolution.To calculate the relevance degree,firstly,this paper according to the characteristic of single instance,multi-tenant,create the model basing on the hierarchical colored Petri net.By the static structure of single instance,get the service relevance between services.By the dynamic operation of multi-tenant,get the tenant behavior set.Then,according to the data mining technology,propose the method of calculating service relevance underling dynamic evolution by using the service relevance set and tenant behavior set.At last,the experiment is carried out to calculate the relevance degree of CRM system internal service,providing new ideas and methods for software evolution correlation analysis.