A Multi-tenant Data Management Method And Its Application In Intelligent Transportation System

Multi-tenant data storage is an important research in Software as a Service (SaaS). The characteristics of "Single-instance, multi-tenancy" make many research scholars committed to solve the problem of great interest. Currently, there are some popular methods such as the private table method, extended table method, universal table method which based on relational databases and Key-Value table method which based on non-relational databases, etc. Of which, universal table method use a shared data table to storage all the tenants’data, effectively reduced the storage space. However, the data stored in the universal table show a relatively sparse characteristic, it also waste the storage resources to a certain extent.As the present situation of traffic data show the features of multi-source and large amount, the traditional Multi-tenant data storage method is difficult to realize the effective management. This paper analyzes the popular multi-tenant data storage methods, for improving both the utilization of data storage space and the query efficiency, proposes an improved multi-table storage mechanism in universal table method, including multi-tenant data hierarchical management model, and implemented a traffic-aware data management platform. The main work and results of this study are as follows:1. The paper analyzed the popular multi-tenant data storage and extension mechanisms, and proposed multi-tenant data management hierarchical model, which helps developers shielding the underlying multi-tenant data management system implementation.2. The paper analyzed the characteristics of storage data in single universal table, and used multi-universal tables that have different number of column instead of a single universal table, according to the data customization need of tenants, stored the tenants’data into the appropriate universal table. This method improved efficiency of storage space, and reduced the impact to the data query.3. The paper analyzed the data migration problem when the tenant data model changed, and proposed the storage mode that multi-universal table combined with key-value pair to solve the data migration problem.4. The paper designs a multi-tenant data management application for traffic-aware data platform, analyzes structure and function of the overall framework of the platform, describes the implementation of the write and query functions, and finally with experimental verification shows the high storage density, and good writing and query performance.