Research On Dynamic Construction Method Of Collaborative Application In PaaS Platform

With the maturity of network and software technology, especially the popularity of cloud computing, great changes have taken place in the pattern of software development, delivery, deployment and usage. Software developers no longer use the traditional design and development model, it's a new trend of software development model based on service composition; software developers needn't concern on the underlying technology of web applications, but focus on implementation of business logic in the software delivery, and deliver the application business logic to the cloud computing-oriented network platform based on platform as a service (PaaS),we called it PaaS support Platform, in other word this PaaS platform can make business logic become web application for the large-scale user; it's convenient for software vendors to deploy software and application due to the scalability of cloud computing and the low cost of leasing; different users adopt software as a service (SaaS) model to customize and use those delivered software in the platform to meet individual requirements.The current PaaS platform focuses on support for a single application, for example support users with customizing individual application. However, the current PaaS platform is general lack of support for collaborative applications. Now the business requirements of industry users become increasingly complex, the large and complex information system has webified, collaborative and cross-domain growing trend. The popularization of collaborative applications is also becoming more and more widely. However, the development, delivery, deployment and usage of such collaborative applications need to solve many underlying technology, and need support of application development, delivery, operation, maintenance and other key technologies in cloud computing. In particular, how based on network, based on the PaaS theory, based on massive software services delivered to PaaS platform, dynamically construct collaborative applications on demand to meet requirements will face many challenges.This paper pays more attention on collaborative applications in Cloud Computing, and concerns the following aspects:1) Propose a PaaS Platform supporting collaborative applications. Platform supports software vendor (ISV) with delivering applications, data and constraints to the platform in the form of service registration or full deployment; supports collaborative application integrators with building collaborative applications based on collaboration relationships and constraints to meet business objectives; support industry customers with customizing collaborative applications.2) In order to automatically and quickly construct large-scale collaborative web application to fulfill complicated application requests with applications in PaaS platform delivered by different ISV, this paper proposes a dynamic construction method of collaborative applications based on automatic planning. Based on collaborative semantic model for application collaboration and the association and constraints between applications, our method uses HTN Planning to construct the logic of software service orchestration on demand.3) It is necessary to verify the correctness of application collaborative association and whether service orchestration logic from different ISV can collaborative with each other. The process of verification process contains the following two aspects. Matching verification of service orchestration logic is the use of the execution path of the application in graph theory to fulfill application matching principles. Matching verification of data availability is the use of Graph Planning to meet data matching principles.PaaS Platform supporting application collaboration has great utility value in the domain which needs collaboration such as e-commerce. We design and implement a prototype system. Our method can ensure not only the semantic consistency between the business objectives and application logic, but also data consistency in the course of application execution. The correctness and necessity of this method are verified through experiments and some cases.