Research On Data Storge Schema And Mapping Mechanism In Multi-Tenants Environment

As a new software development model, Software as a Service (SaaS) has become a focus of research. In the SaaS schema, service providers responsible for the provision, management and maintenance of the necessary IT infrastructure and software to run basic support services platform, while the tenants use the customization features in the service platform to build exclusively for their own personal business systems, using it online. This application schema system has changed the traditional software compliance needs analysis, system design, code development, testing and transfer of the development process, instead of that, it constructed universal system platform for application fields and invited tenants to customize for their software products with personal requirements. On one hand, it can reduce the cost of the software product, and allow for more SMEs having the opportunity to join the information technology for the enjoyment of IT applications to enhance the productivity. On the other hand, SaaS application providers provide application services for the large number of tenants, exploiting economies of scale continue to reduce operating costs.Data has always been highly valued because it is the core of enterprise applications. The introduction of a multi-tenant environmental by SaaS applications requires different tenants to share database, operating system, hardware and other resources and infrastructure. In the new environment, how to ensure that the data storage structure not only to meet specific storage needs of different tenants, but also to maintain high resource utilization and provide a good overall data access performance has been the important goal of SaaS application data storage designers. Especially when large number of tenants increases, by adding hardware resources or the deployment of the cluster can meet the application requirements without changing the data storage structure to meet the system scalability.This paper mainly studied data storage pattern and the mapping transformation mechanisms in the environment of multi-tenants. Including the data storage schema of sharing database and schema, schema mapping and query conversion algorithm, and resources dynamically adjust the optimal strategy. Proposed multi-level data classification and storage database table structure. The main work and achievements include:(1) In shared database and schema data storage style, this paper made a certain amount of expansion on the metadata model and introduced the degree of class attribute that to share, getting metadata to more accurately describe the business data model data objects and fields. According to this type of property sharing, business data storage will be further divided into three areas and build the appropriate storage and index structure. By distinguishing different categories of business data, get a good overall performance and throughput to visit.(2) In the schema mapping and query translation of the proposed design, it proposed the design to meet the tenants of the view layer, the tenant exclusive data model, no effect on each other.(3) In addition, in order to constantly adapt to the process of data traffic operations change, in this paper, based on access frequency statistics, frequently presented the concept and calculation method attributes, as adjusted basis. Prototype system was done by some of the evaluating, validating the effectiveness of adjustment strategy.