摘要
目前信息集成技术对于新的数据库应用来说非常重要,对象代理模型可以作为信息集成的一种通用的数据模型。对象代理模型是一种能够很好解决各个异构数据源集成的一种数据库模型。它不仅有传统面向对象模型的丰富的语义及抽象机制,而且它还能解决面向对象模型可扩充性不强及很难实现视图机制的弱点。对象代理模型的主要特征如下:
1) 一个代理对象享有其自己的标识符,而且除了继承源对象的属性和方法外,还能拥有独自的属性和方法。
2) 一个对象的属性和方法可以被它的代理对象所继承,通过切换操作实现。在切换操作中能改变所继承的属性和方法的名字和类型。
3) 在对象和它的代理对象之间存在着双向连接,这样既能实现继承,又能实现更新迁移。
由于信息集成是基于对象代理模型,这就要求直接访问处理的数据必须是00数据,如果能在非00数据源上实现00接口,把这些非00数据转化成00数据。这样,对于任意数据源,就可以采用基于对象代理模型去集成它们。而针对当今主流数据库仍是关系型数据库的特点,我们在Smalltalk环境中实现了对于Oracle9i这种最常用的关系型数据库以及PostgreSQL这种开放源代码的颇具前景的对象关系型数据库的00接口,使得可以直接用面向对象方法去访问这些基于关系型的数据库
通过建立代理对象和源对象,查询处理的对应关系也能够较好的实现,应用可以将不同的对于代理对象的查询处理翻译成对于局部数据源源对象的查询处理,也可以把对局部数据源的查询结果以用户应用想要的方式返回。我们已经在Smalltalk环境中实现了对象代理模型,在这篇论文中,我们讨论在Smalltalk环境中如何实现基于对象代理模型的异构信息源的集成。
Information integration is very important for advanced database applications,Object Deputy Model can be used as common data model for information Integration.It can solve the problem of all kinds of heterogeneous database sources. The object deputy model is more flexible than the traditional object-oriented model. It can provide inheritance for specialization, generation and aggregation. In addition, the view mechanism is easy to be implemented by the object deputy model. The use of the object deputy model as the common data model can facilitate the creation of the federated schema.The major features of deputy objects are summarized as follows.
1. A deputy object has its own persistent identifier, and may have additional attributes and methods that are not derived from its source objects.
2. The attributes and methods of and object can be inherited by its deputy objects.The inheritance is realized through switching operations that can change the names and types of the inherited attributes and methods.
3. There is a bilateral link between an object and one of its deputy objects, which allows not only inheritance but also update propagation between them.
Because information integration is based on object deputy model, accessed data must be object-oriented data. If we can implement a interface between object-oriented data and non object-Oriented data, we can tranverse the data according to query data. Nowadays the most popular database is relational database, so we implement the oracle9i database system interface and the object-oriented interface of postgresql database that is a land of object-relaiton open-source database system, and then we can access the data from the database system, using the object-oriented methods.
Through establishing the source objects and deputy objects, the corresponding relations about query handle can be implemented better. Accroding to object deputy model, application can translate the query disposal of the deputy object into the ones of the local database sources, it also return query results from local database sources, according to the application user's requirements. We have implemented object deputy model in Smalltalk. In this paper, we discuss how to realize integration of heterogeneous databases based on object deputy model in Smalltalk.
引文
[1] Zhiyong PENG, An Object Deputy Model for Advanced Database Applications (1994)
[2] Zhiyong PENG, Yahiko.Kambayashi,Deputy Mechanisms for Object-Oriented Databases, Proc. of IEEE 11th hat.Conf, on data Engineering,pp.333-40(1995)
[3] Zhiyong PENG, Yahiko.Kambayashi, Resolving Conflicts and Handling Replication during Integration of Multiple Database by Object Deputy Model,ER 2001,LNCS 2224,pp:285-298,2001
[4] Yannis Papakonstantinou,Hector Garcia-Molina,Jeffrey Ullman, MedMaker: A Mediation System Based on Declarative Specifications, 1063-6382/96 1996 IEEE
[5] Ling Ling Yan,M.Tamer Ozsu,Ling Liu, Accessing Heterogeneous Data Through Homogenization and Intergration Mediators,0-8186-7946-8/97 1997 IEEE
[6] Cincom Systems,Inc, VisualWorks Database Connect,P46-0128-02
[7] 数据库系统实现Hector Garcia-Molina Jeffrey D.Ullman Jennifer Widom
[8] 数据库原理编程与性能 Patrick Oneil Elizabeth Oneil
[9] Roel Wieringa, Wiebren de,Jonge, The identification of objects and roles(1996)
[10] Ioana Manolescu,Daniela Florescu, Answering XML Queries over Heterogeneous Data Sources
[11] Ashish Gupta, Venky Harinarayan,Dallan Quass, Aggregate-Query Processing in Data Warehousing Environments
[12] 数据库设计 何玉洁 武欣 邓一凡 等译 机械工业出版社
[13] W.Effelsberg and M.Manino, Attribute equivalence in global schema design for heterogeneous distributed databases, Inform.systems, vol.9 no.3/4 1994
[14] Laks V.S. Lakshmanan,Fereidoon Sadri,Iyer N.Subramanian,On the Logical Foundations of Schema Integration and Evolution in Heterogeneous Database Systems.
[15] M.W. bright,A.R. hurson,Autoated Resolution of semantic Heterogeneity in
Multidatabases.
[16] EGriffiths sellinger, M.M.Astrahan, Access Path Selection in a Relational Database Management System.
[17] Cincom Systems,Inc, VisualWorks Smalltalk tutorial. P46-0128-04
[18] postgresql Technical document,postgresql application develop group.
[19] Charles w.Baehman and Manilal Daya, the role concept in data models.
[20] M.EReddy, Amar Gupta, A Methodlogy for Integration of Hegerogeneous Databases.
[21] oracle9i数据库高级管理,飞斯科技产品研发中心,电子工业出版社.
[22] 面向模式的软件体系结构Frank Buschmann Regine Meunier机械工业出版社.
[23] Cassio Souza dos Santos,Design and Implementation of Object-Oriented views,Proc.of DEXA 95,pp.91-102(1995).
[24] Hong-Tai Chou,David J.DeWiu,An Evaluation of Buffer Management Strategies for Relaitonal Database Systems.
[25] Surajit Chaudhuri, An Overview of Query Optimization in Relational Systems.