分布式实时多数据流的并发协作处理

基于C/S模型分析了分布式实时计算环境中多数据流并发处理的事务问题.采用复杂事件驱动技术,设计了一个适于多数据流并发协作处理的事务机制ARTs-MDS,能够自组织地将对分布式实时数据产生作用的并发任务组成一个原子的事务单位.系统维护着一个永久性的请求/响应队列将物理的设备的实操作与实时事务相分离,通过保证并发协作任务的事务特性获得以分布式局部采集数据为基础产生的微观层操作行为协作产生的整体效果的原子性.测试结果分析表明:该系统能提高对连续的外部采集数据进行协作处理的实时响应,减少了数据丢失.

Coordinate concurrent processing over distributed real-time multi-data streams

Based on Client/Server model, the transactional issues in distributed real-time multi data streams processing are first addressed. By using complex event processing, a transactional mechanism ARTs-MDS was designed to support coordinate concurrent processing on multi-data streams. The system is capable of reforming a set of concurrent tasks as a transactional unit in a self-organization way, isolating real operations on physical devices from the logical operations in the tasks by a persistent input/output queue, and achieving an atomic global effect caused by distributed coordinate tasks processing collected local data. Our performance evaluations show that by applying the system, response time on continuous collected data can be significantly enhanced and data lose can be reduced correspondingly.