Anomaly Detection of Manufacturing Equipment via High Performance RDF Data Stream Processing
The ACM DEBS Grand Challenge 2017 focuses on anomaly detection of manufacturing equipment. The goal of the challenge is to detect abnormal behavior of a manufacturing machine based on the observations of the stream of measurements provided. The data produced by each sensor is clustered and the state transitions between the observed clusters are modeled as a Markov chain. In this paper we present how we used WSO2 Data Analytics Server (DAS), an open source, comprehensive enterprise data analytics platform, to solve the problem. On the HOBBIT (Holistic Benchmarking of Big Linked Data) platform our solution processed 35 megabytes/second with an end-to-end mean latency of 7.5 ms at an input rate of 1 ms, while the events spent only 1 ms time on average within our grand challenge solution. The paper describes the solution we propose, the experiments’ results and presents how we optimized the performance of our solution.