数据访问的事务处理中间件中等待时间的研究

Research on Waiting Time in Data Access Oriented Transaction-Processing Middleware

数据访问的事务处理中间件主要用于分布式的数据库访问的事务处理 ,采用两阶段提交技术。在两阶段提交中当协调者把任务分支分配到所有的参与者后 ,需要在等待一定的时间再发出准备消息。等待时间的大小会对分布式系统的并发性产生影响 ,为了提高系统的并发性 ,需要对提交等待时间进行估计。该论文对于数据访问的事务处理中间件任务的类型较少的特点 ,提出了一种对于提交等待时间进行较准确估计的方法 ,且针对不同类型的数据访问 (单个操作和批量操作 )采用不同的时间估计 。

Data access oriented transaction-processing middleware is used for distributed database transaction processing, which uses two-phase commit. In two-phase commit, coordinator distributes transaction branches to all participating sites, and after a prespecified interval of time sends preparatory message. Because the interval of time will affect the concurrency of distributed system, the estimation of the interval of time is needed. This paper provides a method to estimate the interval of time based on that data access oriented transaction-processing middleware has a few kinds of tasks. Moreover according to the kinds of data access(single operating and batch operating),the paper uses different estimating method, and brings forward the concept of long transaction.