Design And Realization Of Optimized Query Strategy About Multi-Tenant Saas Based Application

SaaS, software as a service, is a new software development and deployment way in recent years. It’s not like traditional software, SaaS itself has an agile reaction for marketing, it has a short period development, low price of operating process attributes, and can offer service to many companies and people together. So it’s very fit to accomplish many small and medium-sized enterprises online office systems. Because SaaS itself is powerful, SaaS is put forward high request about SaaS design and deploy architecture. How to satisfy many users request in common time, guarantee each user’s service quality, it’s also a severe problem for SaaS to face.As to discuss SaaS’s characteristics, the article describe thorough SaaS software architecture, including SaaS metadata layer, authority management layer and user data request layer, also describe and discuss to usage technology and search performance. In order to realize highly customizable configuration and stability, it’s provided SaaS metadata architecture. Metadata is data to generate data. It can store in SaaS operating system side, based on SaaS metadata query commonplace, In metadata layer, we realize our Cache system, save all of the metadata. So as to more loose coupling system, we split SaaS system, authorize management system, user data request system. Through online communication, guarantee SaaS system data security characteristic, make enterprise more likely use SaaS model to accomplish their work. In authorize management layer, we adopt LDAP directory to query in order to confirm query performance. Because authority is hardly changeable, we use read performance higher than write performance system LDAP. Moreover, in user data storing layer, MIT provided the concept of DaaS, is database as a service, storing data is considered as online service. So user data can store in cloud, or store in company’s local disk, it can enhance user’s data security confidence. As a result of splitting user data and metadata, SaaS request must point to user data node online position, so we put the pointing work into authority management system. Others, in order to have better query performance, we add user data a Cache layer also. However, because of complex of user data, our Cache can only store part data, it is not like metadata layer do. We use hit and miss hit to judge Cache data value and provide a new way of our system usage to value data, put forward several ways to enhance performance of query Cache.In all, we analyze the SaaS architecture, as to every procedure realization provide our design and solvable scheme, and guarantee the query and storage performance. We use many new technologies like Coherence Cache to realize query Cache system, LDAP to realize base on directory authority management system, etc.