SaaS多租户数据存储技术研究与应用
发布时间：2018-10-19 08:39
【摘要】：SaaS是一种伴随云计算的兴起而产生发展的新型软件交付模式。这种模式下，软件由服务提供商开发，通过互联网交付，不需要在本地安装部署。多租户是SaaS应用的核心技术之一。企业通过网络向软件服务提供商租赁所需服务，被称为租户，SaaS模式决定了云计算是为多个租户提供服务的。多个租户共享一个软件实例，通过软硬件资源的复用，实现资源利用的最大化，降低运营成本。 租户数据的合理高效存储与管理是SaaS应用的基础。本文采用分层数据存储方案并对现有模式映射技术进行改进，主要解决多租户数据存储的如下问题：1.在租户的业务数据结构不尽相同的情况下，满足不同租户的需求；2.在租户规模扩张和租户数据量增长的情况下，保障数据检索效率和数据存储利用率；3.在租户数据共享物理存储的情况下，保障租户数据的逻辑隔离性。具体工作及成果有以下几个方面： 首先基于共享数据库共享模式架构，设计具有扩展性的多租户分层数据存储方案，在租户逻辑视图层和物理存储层之间增加模式映射层，管理租户元数据，进行查询转换，结果重构和数据迁移。上层应用通过租户逻辑视图获取租户相关数据，不需了解中间映射层和底层存储机制。 然后研究数据模式映射机制，，提出基于快折叠（Chunk Folding）的优化数据模式映射技术——泛型扩展表映射。并在存储效率和重构操作上，将其与现有数据映射技术扩展表映射，通用表映射，透明表映射和块折叠映射进行比较。 最后，通过将SaaS多租户应用到校园能源云平台，来验证优化数据模式映射技术具有较优的存储效率和存取性能，以及分层数据存储方案的合理性与扩展性。
[Abstract]:SaaS is a new software delivery model with the rise of cloud computing. In this mode, software is developed by service providers, delivered over the Internet, and does not require local installation and deployment. Multi-tenant is one of the core technologies of SaaS applications. Enterprises rent the required services from software service providers through the network, known as tenants, and the SaaS model determines that cloud computing provides services to multiple tenants. Multiple tenants share a software example, through the reuse of software and hardware resources to maximize the use of resources and reduce operating costs. Reasonable and efficient storage and management of tenant data is the basis of SaaS application. In this paper, the hierarchical data storage scheme is adopted and the existing schema mapping technology is improved to solve the following problems of multi-tenant data storage: 1. In the tenants' business data structure is different, meet the needs of different tenants; 2. In the case of tenant scale expansion and tenant data growth, data retrieval efficiency and data storage utilization are guaranteed. 3. In the case of shared physical storage of tenant data, the logical isolation of tenant data is guaranteed. The specific work and results are as follows: firstly, based on the shared database sharing schema, a multi-tenant hierarchical data storage scheme with extensibility is designed. A schema mapping layer is added between the logical view layer and the physical storage layer to manage the tenant metadata, query conversion, result reconfiguration and data migration. The upper application acquires tenant related data through the tenant logical view without knowing the intermediate mapping layer and the underlying storage mechanism. Then, the mechanism of data schema mapping is studied, and an optimized data schema mapping technique based on fast folding (Chunk Folding), generics extended table mapping, is proposed. The storage efficiency and refactoring operation are compared with the existing data mapping techniques, such as extended table mapping, general table mapping, transparent table mapping and block collapse mapping. Finally, SaaS multi-tenant is applied to campus energy cloud platform to verify that the optimized data schema mapping technology has better storage efficiency and access performance, as well as the rationality and expansibility of hierarchical data storage scheme.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP333
本文编号：2280652
[Abstract]:SaaS is a new software delivery model with the rise of cloud computing. In this mode, software is developed by service providers, delivered over the Internet, and does not require local installation and deployment. Multi-tenant is one of the core technologies of SaaS applications. Enterprises rent the required services from software service providers through the network, known as tenants, and the SaaS model determines that cloud computing provides services to multiple tenants. Multiple tenants share a software example, through the reuse of software and hardware resources to maximize the use of resources and reduce operating costs. Reasonable and efficient storage and management of tenant data is the basis of SaaS application. In this paper, the hierarchical data storage scheme is adopted and the existing schema mapping technology is improved to solve the following problems of multi-tenant data storage: 1. In the tenants' business data structure is different, meet the needs of different tenants; 2. In the case of tenant scale expansion and tenant data growth, data retrieval efficiency and data storage utilization are guaranteed. 3. In the case of shared physical storage of tenant data, the logical isolation of tenant data is guaranteed. The specific work and results are as follows: firstly, based on the shared database sharing schema, a multi-tenant hierarchical data storage scheme with extensibility is designed. A schema mapping layer is added between the logical view layer and the physical storage layer to manage the tenant metadata, query conversion, result reconfiguration and data migration. The upper application acquires tenant related data through the tenant logical view without knowing the intermediate mapping layer and the underlying storage mechanism. Then, the mechanism of data schema mapping is studied, and an optimized data schema mapping technique based on fast folding (Chunk Folding), generics extended table mapping, is proposed. The storage efficiency and refactoring operation are compared with the existing data mapping techniques, such as extended table mapping, general table mapping, transparent table mapping and block collapse mapping. Finally, SaaS multi-tenant is applied to campus energy cloud platform to verify that the optimized data schema mapping technology has better storage efficiency and access performance, as well as the rationality and expansibility of hierarchical data storage scheme.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP333
【参考文献】
相关期刊论文 前2条
1 丁文定;;基于服务集成管理的SaaS解决方案[J];计算机与现代化;2010年01期
2 昌中作;徐悦;戴钢;;基于SaaS模式公共服务平台多用户数据结构的研究[J];计算机系统应用;2008年02期
相关博士学位论文 前1条
1 孔兰菊;SaaS应用交付平台中多租户云数据管理关键技术研究[D];山东大学;2011年
本文编号：2280652
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2280652.html
