Study On Key Techniques Of Personalized Tenant Customization In The Interaction Based Multi-Tenant SaaS

With the rapid development of web environment, the applications of cloud computing and SaaS are increasingly rich. SaaS as an on-demand software delivery mode which provides software as service with a variety of use patterns to implement the individual requirements of different users, has become a hot issue in the research field of cloud computing. Currently, for the problem of multi-tenant SaaS customization, lots of research works have been done from different facets conducted by researchers from both academia and industry, and great achievements have been gained. But at present, most of multi-tenant SaaS customizing approaches are on the basis of process orchestration business organization form, which adopts the customizing approaches either based on variation modeling or metadata to be configured. The business process in the form of orchestration needs a control flow to control the business activities and coordinate the execution sequence of these activities. Consequently, when the business process needs to face the change of the personalized tenant business requirement, it needs to design and plan the business process change from the perspective of the overall consideration. For the multi-tenant SaaS personalized customization, these methods are very difficult to change and customize the local implementation of business process, which reduces the flexibility of SaaS personalized tenant customization.Therefore, this dissertation proposes an approach for customizing personalized multi-tenant of interaction based multi-tenant SaaS, which is on the basis of the interaction based application mode of multi-tenant SaaS, with the aim of flexible and effective customization of personalization, by the business description based on interaction and business organization form based on process choreography to meet the users’personalized business requirements much more better. For this approach, this dissertation proposes a model of business and tenant for interaction based multi-tenant SaaS, which is a basis for customizing the personalized tenant of SaaS. The key issues in customization framework of the interaction based multi-tenant SaaS, such as graphical operation based customization approach for personalized tenant business process, effectiveness detection of personalized tenant business process, the graph matching based tenant recommendation approach are studied.(1) For the problem of the flexibility of the business organization, on the basis of analysis of the maturity model of SaaS application and the problems in the orchestration business organization form, an application mode of interaction based multi-tenant SaaS is proposed, a model of business and tenant for interaction based multi-tenant SaaS is defined, and the description language SaaSCL for this model is designed by extending WSCL. Based on this, the process of constructing interaction based multi-tenant SaaS is presented, and the customization framework of personalized tenant customization of multi-tenant SaaS based on SBTM model is proposed, which lays a foundation for personalized customization of multi-tenant SaaS.(2) For the problem of personalized tenant customization and modeling with diversity of business variations, the business views as well as their customization operation behaviors are proposed by analysis of the customized elements and actions which can affect the change of tenant business process in the multi-tenant SaaS. For each of the business view, the establishment algorithms are designed. Based on this, graphical operation based customization approach for personalized tenant business process is proposed, and by a concrete business instance, the performance of the approach is shown.(3) For the problem of effectiveness detection of personalized tenant business process, by the analysis of current study in the field of process validation, the basic concepts of effectiveness detection of personalized tenant business process is presented. For the problem of diction of correctness of interaction relation constraint and transition guard include-relationship, the detecting activities, detection rules as well as the TLA description of the detection behavior is presented. The approach of correctness detection of the business relationship is proposed. The BRVT algorithm to support this approach is designed. For the problem of interaction reachable-ness, process completeness and death cycle of the business process, the approach of directed graph based business process rationality verification is proposed, and the directed graph depth-first traversal algorithm DFS_SCBM to support this approach is designed. On this basis, the SCBM_AT algorithm for comprehensively effectiveness detection of the personalized tenant business process is designed, which can ensure the effectiveness of the customizing result of the personalized tenant.(4) For the problem of the recommendation from the existing tenant business models, by the analysis of the importance of the tenant recommendation, the graph matching based tenant recommendation process is presented and the basic concepts in the approach of tenant recommendation are defined. Taking the case database optimization techniques in CBR as the core idea, the index table based recommendable tenant business model filtering approach is proposed and the TI_Create algorithm for establishing the index table and the RSTMF algorithm for filtering the tenant business model are designed. Taking the case retrieval technique in CBR as the core idea, the ECGM cost detection based graph matching approach is proposed for model matching, and the ECGM_CT algorithm based on A*for computing the cost distance between models is designed and the node matching as well as the edge matching methods are presented in detail. By the experiments, the factors influencing the quality of the algorithm are analyzed and the importance of the recommendable tenant business model filtering in improving the performance of the tenant recommendation.