分布式区域农业信息系统元数据设计研究

Design of Metadata Schema for Distributed Regional Agricultural Information System

该文讨论了在分布式的异构网络环境下,针对区域农业信息管理与服务系统的的体系,如何进行元数据设计的问题。文章分析了目前元数据标准存在的问题,根据农业信息管理与服务系统的体系架构,提出分别从数据、软件平台、网络服务(WebService)3个方面来考虑元数据的设计问题,认为它们的元数据内容也就构成了整个系统的元数据框架。该文建议采用面向对象技术进行元数据的设计,将元数据与它所要描述的对象(数据、软件模块、网路服务)从逻辑上捆绑在一起,把元数据内容当作对象的一个组成部分。通过对象抽象过程,确立描述该对象的内容以及作用于对象上的操作,从而定义出相应的元数据结构。这样,元数据设计的过程就转化成为面向对象的分析与设计过程。采用面向对象的元数据设计方法可以有效地避免元数据的丢失或者与数据不一致的情况。最后,文章简单给出了分布式农业信息系统的基本操作元数据内容。

Design of metadata schema for Distributed Regional Agricultural Information System (DRAINS) is discussed in this paper.DRAINS runs on the internet and integrates web service, object-relational database management system and geographic information system technologies.It includes four ties: database tier, platform tier, application tier and user tier.This paper examines the current metadata standard firstly.And then, base on the architecture of DRAINS, it proposes metadata design should concern data, software platform and web service, metadata of which also make a architecture of metadata.An object-oriented method is introduced to develop metadata schema.In this method, metadata are binded in their associated objects tightly and become a part of data objects, software module objects and web service objects.In the process of abstract object analysis, objects are formed and their attributes and operations of objects are defined.As a part of object, metadata structure for attributes and operations also are built during the process.Thus, metadata analysis and design become a process of OOA and OOD.This paper discusses the metadata schema of data objects, software module and web service.1)By OO method, many data objects can be defined to support various data types, and metadata are defined based on definition of objects.The method can prevent metadata from missing or mismatching with their associated object; 2)Modularization of software platform can also use the method.metadata of software describe the functions, interfaces, data requirements, parameters, host server, version information and etc..Differences on functions and interfaces between versions can be recorded on metadata; 3)Web services are objects or components deployed on network and have good encapsulation feature.Generally web services are proxies of software modules and are distributed on network nodes.Metadata of web service describe functions, interfaces, parameters, authorization and register information.Finally, a brief operational medata schema is gaven for distributed agricultural information system as a sample.