The Research On Complex Event Processing Method For Internet Of Things

Internet of Things (IOT) is an important part of the new generation ofinformation technology, its application areas has overlaid every trade. How tomanipulate information of IOT application is the key issue for IOT widely used, andIOT middleware can deal with this information effectively. Complex event processing(CEP) is one of the core tasks of middleware for IOT. Event stream of IOTapplications has many features, such as heterogeneous, distributed, mass anduncertainty, et al. Existing work on pattern evaluation, however, assumes that theoccurrence time of each event is known precisely and the events from various sourcescan be merged into a single stream with a total order, but we find that eventoccurrence times are often unknown and imprecise in real applications.In order to resolve these problems, a complex event processing method for IOTuncertainty event is proposed in this paper. Meanwhile, in accordance with the IOTdistributed application characteristic, a distributed CEP architecture is proposed basedon Service Oriented Architecture (SOA), and then presents a Distrbuted ParallelComplex Event Processing (DPCEP) method. Main works of this paper include:At firstly, in accordance with the feature for event stream of IOT application andthe disadvantages of existing CEP methods, this paper proposes an uncertaintytemporal pattern that assigns a time interval to each event to represent its entirepossible occurrence, and then present a CEP Interval-based method of such pattern.This uses K-Slack buffer mechanism to buffer the out of order events. In order toprocess mass event more efficiently, an optimization OptInterval-based method isproposed that discards the useless partial order events as soon as possible. Thefundamental experiments show that the method can support uncertain events in CEPand equiped feasibility and acceptable performance than other methods.Secondly, in accordance with the feature of IOT distributed application, thispaper proposes a distributed CEP architecture based on SOA, and we also set thelightweight Embedded CEP engine processing primitive event first that embeddedIOT gateway or underlying equipment. Based on the method of OptInterval-basedproposes a distributed and parallel DPCEP method, which partition the data streamfrom distributed systems according to the user’s query, and use global parallel andlocal parallel to process event, then global CEP merge the local result into global CEPengine to print the all final event sequence. Simulation results show that the method has obvious advantages in big data, and architecture based on SOA has goodgenerality and scalability.