基于数据依赖和触发器的简单子任务调度算法

Simple Subtask Scheduling Algorithm Based on Data Dependency and Trigger

针对在数据库管理系统环境下实现的分布式子任务计算平台,提出了一种基于数据依赖、采用触发器实现的简单分布式子任务调度算法,可有效保证分布式子任务调度的准确性和一定程度的及时性.首先介绍了采用人工划分的基于执行阶段的子任务调度思想,并结合数据依赖调度算法,证明了两者之间的相似性和密切联系,然后进一步提出基于执行阶段的分布式子任务调度算法,为在数据库环境下实现的分布式子任务计算提供了一种简单、快捷、正确的调度算法.

Now many subtask scheduling algorithms are based on DAG model and its variations,all these algorithms are relatively complex.For distributed computing platform built on DBMS,we proposes a simple distributed subtask scheduling algorithm based on data dependency and implemented by trigger,which guarantees the accuracy and efficiency in a certain extent for distributed subtask scheduling.Furthermore,close relations and similarity between manually decomposed execution phase and data dependency based distributed subtask scheduling algorithm are proved, which lead to the suggestion of a distributed subtask scheduling algorithm base on execution phase, which provides a simple, easy to build and accurate subtask scheduling algorithm in the DBMS environment.