一种Agent数据库系统框架及其规则并行算法

面向对象方法描述的对象是被动的,与主动机制存在矛盾,不能很好地定义分布式主动数据库中对象的功能和特性.分布式主动数据库中的主动规则还带来了终止性、合流性的问题.首先分析面向对象方法的局限性,将Agent技术、分布式数据库、主动数据库相结合,给出一种面向Agent的分布式主动数据库系统框架.然后在这个框架基础上,提出扩展事件规则图方法和改进的Coffman-Graham规则并行算法,并分析它们的性能.分析结果表明,前者在一定程度上解决了数据库系统的终止性问题,后者在保持合流性的基础上提高了规则并行处理效率.这个框架对研究Agent技术在分布式主动数据库中的应用有一定的启示.

An Agent-Based Database System Framework and Its Parallel Algorithm of Rules

The object-oriented method is passive, so it is not consistent with the active rules and can not easily define the functions and features of the objects in a distributed and active database system. And also the active rules lead to the problems of termination and confluence. In this paper, the Agent method, distributed database, and active database are combined after the constraints of the object-oriented method are analyzed. On the basis of the combination, a distributed and active database system framework based on the Agent-oriented method is proposed. As to the system framework, a method of expanding the event-rule graph and an improved Coffman-Graham parallel algorithm are presented, and then their performances are analyzed. Shown by the analysis results, the former is used to solve the problem of database system termination, while the latter, on the foundation of remaining the confluence, improves the efficiency of parallel rule process. Besides, the database system framework is useful for applying the Agent method to the distributed and active database.