项目名称: 基于SysML和MARTE的异构数据模型转换方法研究
项目编号: No.61472180
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 自动化技术、计算机技术
项目作者: 张天
作者单位: 南京大学
项目金额: 82万元
中文摘要: 大数据给传统的软件开发技术带来了新的挑战,而其中异构数据的集成问题被认为是充分发挥大数据优势的核心问题。同时,由于在大数据应用下,变化速度、真实性、价值等数据特征普遍受到用户关注,而这些特征会在不同的领域中以不同的方式进行捕获和描述。因此异构数据之间的转换必然要解决如何保持此类特征在转换前后保持不变的问题。本课题从模型驱动工程的视角看待此类异构问题,从异构数据自身的特征入手,基于元建模技术进行面向方面的特征建模,并构造异构数据模型转换框架,采用特征制导的思想建立异构数据之间的转换。此外,在转换框架中将语义映射和语法转换相互隔离,分而治之,以降低转换规则构造难度,提高转换框架的通用性。最后,基于系统工程建模语言SysML和实时嵌入式建模语言MARTE以及大数据平台Hadoop进行具体的转换研究。
中文关键词: 异构数据;模型驱动工程;模型转换;大数据
英文摘要: Big Data brings new challenges to traditional software development technologies, among which heterogeneous data integration is a key issue considering full advantage of Big Data. Meanwhile, some of the data features, such as velocity, veracity and value, are widely concerned and described in different ways according to different areas. Therefore the problem of keeping these features unchanged before and after the transformation should be solved. The project tries to solve the problem from the viewpoint of MDE. To build the aspect-oriented feature models using metamodeling techniques, to construct transformation framework of heterogeneous data, to perform transformation between heterogeneous data using feature oriented idea. In addition, the semantic mapping and syntax conversion are isolated from each other so as to increase the generality of the framework and decrease the complexity of it. The project is performed based on SysML, a system engineering modeling language, and MARTE, an embedded system modeling language, and finally implemented on the Hadoop platform.
英文关键词: Heterogeneous Data;MDE;Model Transformation;Big Data