Automatic interpretation of a declarative cloud service description

Abstract

The accelerated move from on-premise enterprise software to a cloud service model introduces certain challenges to service providers. The difference between enterprise customer’s workloads and the complexity of software products create the need for a formal service description. It increases abstraction level, encompasses business requirements and eliminates misunderstanding between a service provider and its customers. However, the declarative nature of the description does not allow to determine precise implementation components. Furthermore, each provider is interested in customization and portability of its services to meet requirements of several customers, at the same time allowing automatic selection of service topology. The major objective of this work is to show how declarative cloud service description can be automatically processed, analyzed and mapped to the service topology matching customer’s workload and original business requirements. In order to decrease service delivery time and to eliminate the manual selection of application components, a method of automatic interpretation of a declarative cloud services is proposed. In this work, several solution concepts of interpretation of a declarative service description were discussed. As a result, a method of automatic identification of service components matching to a given business requirements was elaborated. It is based on the gradual reduction of the list of possible service components considering their compatibility and deployment sequence. Moreover, it includes an optimization algorithm that selects exact service topology for the case of several matching options. Additionally, a prototype of an interpreter that transforms declarative cloud service description into the exact topology of components was implemented and discussed.