Transaction Management in a Multidatabase Environment

多数据库环境中的事务管理

• 批准号: 8904932
• 负责人: Yuri Breitbart
• 金额: $ 11.37万
• 依托单位: University of Kentucky
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1989
• 资助国家: 美国
• 起止时间: 1989-10-01 至 1992-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8904932&HistoricalAwards=false
• 关键词: Transaction; Management; Multidatabase; Environment

This project deals with transaction processing in a multi- database environment. A Multidatabase System (MDBS) is a collection of several, locally controlled, databases. It is assumed that local database management systems (DBMS) are not aware of each other and cannot exchange concurrency control information. The objective is to develop systematic methods for maintaining semantically related information in a consistent and reliable manner. The investigation is focused on transaction management issues that arise when global and local transactions are processed concurrently in a multidatabase system. Specific research problems include: 1. Extension of the existing theory for concurrency control and recovery in distributed database systems to the case of multidatabase systems. 2. The design of a mechanism for transaction management in a multidatabase environment, such that each local DBMS uses the same concurrency control mechanism. 3. Performance evaluation of various transaction management systems in a multidatabase environment. 4. Development of a theory of transaction management for cooperating multidatabase systems in a heterogeneous hardware and software environment. The results of this research will form theoretical foundations for transaction handling that will be beneficial to large-scale industrial multidatabase systems development.

这个项目处理多数据库环境中的事务处理。多数据库系统（mdb）是几个本地控制的数据库的集合。假定本地数据库管理系统（DBMS）彼此不知道，不能交换并发控制信息。目标是开发以一致和可靠的方式维护语义相关信息的系统方法。调查的重点是在多数据库系统中并发处理全局和本地事务时出现的事务管理问题。具体研究问题包括：1。将现有的分布式数据库系统并发控制与恢复理论扩展到多数据库系统。2. 在多数据库环境中设计一种事务管理机制，使每个本地DBMS使用相同的并发控制机制。3. 多数据库环境下各种事务管理系统的性能评估。4. 异构软硬件环境下协作多数据库系统事务管理理论的发展。本研究的结果将为事务处理提供理论基础，有利于大规模工业多数据库系统的开发。