一种基于云的S aaS分布式多租户数据库研究

摘要

给出一种基于云平台的SaaS分布式多租户数据库架构，探索了SaaS多租户应用面临的租户自定义数据访问控制和租户海量数据存储两大挑战的解决方法。使用云平台虚拟机作为局部存储节点，借助数据切分和读写分离，实现了可用性和扩展性强的租户海量数据的存储。采用灵活的键值对模型存储租户自定义数据，提出了基于元数据的租户自定义数据访问控制方法，并设计了访问控制算法。将典型应用TPC-W转换为多租户模式并进行原型实验分析，结果表明该架构是有效和可行的。%We proposed a cloud platform-based SaaS distributed multi-tenant database architecture and explored the solutions of two challenges faced by SaaS multi-tenant applications,which are the tenant mass data’s storing and the tenant self-defined data access control. For the first one,we achieved the tenant mass data storage with high availability and scalability by applying virtual machine running in cloud platform as local storage nodes and with the support of data segmentation and read/write splitting.For the second one,we adopted flexible key-value pair model to store the tenant self-defined data,and presented a metadata-based access control method for tenant self-defined data, the access control algorithm was designed as well.To carry out prototype experiment analysis,typical application TPC-W was converted into the multi-tenant pattern,and the results showed that this architecture were effective and feasible.