多视图业务模型存储及访问框架研究

发布时间：2018-02-25 16:00

本文关键词： 模型驱动开发多视图业务模型业务模型库数据存储访问多租户　出处：《上海交通大学》2014年硕士论文　论文类型：学位论文

【摘要】：在以多视图业务建模为基础的模型驱动开发（Model Driven Development, MDD），面向服务架构（SOA）的信息系统实施过程中，业务模型库承担了业务模型数据的管理，实现业务模型到服务的转换并对服务的全生命周期进行支持的职责。云平台下的业务模型库为了应对大量企业的业务模型数据存储和访问需求，，需要支持海量业务模型数据的存储访问，并实现多种模式的复杂应用。然而传统的以关系数据库为核心的存储技术在面临海量结构化、半结构化业务模型数据时，在效率上难以满足需求。而现有的业务模型库技术关注处理基本的业务模型数据访问，而不支持更复杂的查询请求。针对上述的问题，本文提出了多视图业务模型存储及访问框架，作为一个中间件在数据存储、访问、管理、应用等多方面支持云平台下业务模型库，并构建原型系统对其进行了验证。重点研究了多视图业务模型数据本身的特点以及应用模式，结合多种数据存储平台及技术实现多视图业务模型数据的高效存储及管理，并构建多视图业务模型数据的关系元模型，实现对多种复杂应用模式的支持。本文的主要完成的工作主要分为如下几个方面：（1）提出多视图业务模型存储及访问框架框架作为一个中间件支持云平台下的业务模型库进行业务模型数据存储、管理、应用。框架结合多种数据存储平台，采用多租户数据隔离等技术实现云平台下海量业务模型数据的管理，并且通过构建业务模型的关系元模型来实现多视图业务模型的复杂应用模式。（2）结合多种数据存储平台实现底层数据访问通过研究多视图业务模型数据的特点，针对不同类型数据，采用有机结合的多种数据库（包括关系数据库以及非关系数据库），以及经过封装和扩展的HDFS实现底层数据存储访问。并在此基础上进行扩展和封装，以实现高效的数据存储和管理。（3）构造关联元模型并实现复杂应用模式构建用于描述分散存储的多视图业务模型数据间相互关联的关系元模型，实现系统对不同业务模型数据之间关联（包括直接关联以及间接关联）的自动识别。在此基础上设计并实现复杂查询，支持对业务模型数据的多种应用模式。（4）构建原型系统并对其进行验证基于本文所提出的方法，设计并实现了基于多视图业务模型存储及访问框架的业务模型库原型系统。采用某大型物流公司配送流程模型数据为实例，通过对业务模型数据进行存储和执行复杂查询来对原型系统进行验证。验证结果表明多视图业务模型存储及访问框架能够为云平台下的业务模型库提供了从数据存储访问到数据应用的多方面支持，为在模型驱动服务开发中对业务模型数据的管理和应用提供了一个高效可行的方案，具有一定的应用前景以及理论价值。
[Abstract]:In the process of developing information system based on multi-view business modeling model driven by Model Driven Development (MDD), Service oriented Architecture (SOA), the business model base is responsible for the management of business model data. In order to meet the data storage and access requirements of a large number of enterprise business models, the business model base under the cloud platform is responsible for implementing the transformation from business model to service and supporting the whole life cycle of service. It is necessary to support the storage and access of massive business model data, and to implement complex applications of multiple patterns. However, when the traditional storage technology based on relational database is faced with massive structured and semi-structured business model data, It is difficult to meet the requirements in efficiency, and the existing business model library technology focuses on processing basic business model data access rather than supporting more complex query requests. In view of the above problems, this paper proposes a multi-view business model storage and access framework, as a middleware in data storage, access, management, application and other aspects to support the business model base under the cloud platform. And build a prototype system to verify it. The characteristics of multi-view business model data and its application mode are mainly studied. The efficient storage and management of multi-view business model data are realized by combining various data storage platforms and technologies. The relational metamodel of multi-view business model data is constructed to support various complex application patterns. The main work of this paper is divided into the following aspects:. 1) A multi-view service model storage and access framework is proposed. As a middleware, the framework supports the business model database under the cloud platform to store, manage and apply the business model data. The multi-tenant data isolation technology is used to realize the management of massive business model data under cloud platform, and the complex application pattern of multi-view service model is realized by constructing the relational metamodel of the business model. Implementation of underlying data access based on multiple data storage platforms. By studying the characteristics of multi-view business model data, aiming at different types of data, It uses a variety of databases, including relational database and non-relational database, and encapsulated and extended HDFS to realize the underlying data storage access. To achieve efficient data storage and management. Constructing association metamodel and implementing complex application pattern. Build a relational metamodel to describe the correlation between the data of a decentralized multi-view business model, The system realizes the automatic recognition of the association between different business model data, including direct association and indirect association. Based on this, complex query is designed and implemented, and various application patterns for business model data are supported. Build and verify the prototype system. Based on the method proposed in this paper, a business model base prototype system based on multi-view service model storage and access framework is designed and implemented. The prototype system is verified by storing the business model data and executing complex query. The verification results show that the multi-view service model storage and access framework can provide the data from the business model library under the cloud platform. Storage access to multiple aspects of data application support, It provides an efficient and feasible scheme for the management and application of business model data in model-driven service development. It has a certain application prospect and theoretical value.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP333

【参考文献】