A Novel Lightweight Middleware for Distributed Massive PMU Data Mining

PMU data is facing with the problems of processing and application caused by the massive, distributed, multi-source, high redundancy existing in data. Therefore, this paper proposes a novel lightweight middleware to process and mine distributed massive PMU data. There employs light sampling parallel mining algorithm to deal with rules of data association, and proposes round robin algorithm to allocate computing tasks and data streams equally. Then the proposed middleware is applied in a multi-core computer. Test and comparison results from mining massive PMU data verify the feasibility and effectiveness of the proposed middleware technology in algorithm strategy and real-time requirement.