Font Size: a A A

Research On SaaS Multi-Tenant Data Model And Implementation Technology

Posted on:2017-09-04Degree:MasterType:Thesis
Country:ChinaCandidate:W ZhangFull Text:PDF
GTID:2348330512964247Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the rapid development of the SaaS model, support more tenants become a key problem for SaaS application. In data storage layer, the domestic and foreign researchers have proposed many data model and solutions based on relational database, wide table model rely on it's customize flexible, good scalability, high degree of sharing and use of low cost advantages has been widely applied. In the implementation process, wide table model in order to effectively deal with the data in the table storage requirements of different tenants, set aside a large amount of data in the table columns, although such a design for a SaaS application provides a flexible customization and extensibility, but the tenant not customized data listed in the store when it is empty values, this situation lead to the data in the table is very sparse, reduces the storage space utilization efficiency. When growing amounts of data, a relational database in extension problems encounter bottleneck, is not conducive to business data processing and analysis of large data environment.This paper in view of the wide table model under the environment of large amount of data query efficiency and variety of insufficient storage space utilization, use NoSQL database for the underlying database, designed a data model of metadata driving based on column family storage, eliminates the problems of a large number of null values in wide table model, and through a contrast experiment in the distributed environment with wide table model, proved in this paper, the design of the data model in a large amount of data environment is feasible and has good query efficiency. Paper's main work is as follows:1) The wide table model based on relational database have been studied in detail. Associated with wide table model in this paper reference documents, all documents only for the relationship between the tables and each table structure description and analysis, does not relate to the actual business data stored procedure and query transformation logic. In actual operation process of the SaaS application, the query logic to physical query is very important for the transformation, this paper analyzes the wide table query conversion process, and calculated its space utilization.2) Design a Based on column family store and metadata driven data model. For full use of advantages that the column family database system itself, the model on the basis of wide table model was improved, by introducing a secondary index creation and optimization mechanism made up for the inadequacy of the underlying database in terms of conditions query, prove that this data model provides a good query efficiency.3) Through the contrast experiment under the distributed environment analysis in this paper design's data model and wide table model in the performance of the query efficiency and space utilization. The experimental results show that with increasing amount of data, using HBase database implementation based on column family store and metadata driven data model has a good and stable query efficiency, can satisfy the large volume of data environment multi-tenant customization requirements.4) A general log management system was designed and implemented. System adopts the based on column family store and metadata driven data model store multiple source of log data, through the implementation of this system proves the correctness and feasibility of the data model.
Keywords/Search Tags:SaaS, Multi-Tenant, Data Model, NoSQL, Column Family Storage
PDF Full Text Request
Related items