基于MapReduce的语义发布/订阅系统

MapReduce-based semantic publish/subscribe system

为了提高已有的语义发布/订阅系统中事件与订阅匹配的时间效率,提出了基于MapReduce[1]的语义发布/订阅系统[2]。对语义发布/订阅系统的处理流程进行了认知与分解,指出订阅与事件的匹配时间效率是系统必须要解决的问题,在语义发布/订阅模型的基础上设计了新的匹配模块。对过去已完成的匹配方案进行了分解,采用基于MapReduce的并行处理技术[3]对事件与订阅的匹配进行并行处理,从而提高系统的时间效率。该系统通过加入多台具有相同处理能力的处理机,搭建了一个能够并行处理的运行环境,验证了该系统的准确性和有效性。

To improve the matching time efficiency of the existing sematic publish/subscribe system,a semantic publish/subscribe system based on MapReduce is presented.Firstly,the process of handling in the semantic publish/subscribe system is recognized and decomposed.It is pointed out that the difficulty in the system is to deal with the matching time efficiency between the event and the subscription.Based on the semantic publish/subscribe system,the new matching model is designed.Then,the past matching schemes are decomposed by adopting the parallel processing technology based on MapReduce to improve the time efficiency of system.Finally,by adding several the same processors and building the running system of parallel processing,the validity and efficiency of the system is demonstrated.