Research On Massive Real-time Data Stream Parallel CEP In Manufacturing Internet Of Things

Internet of Things in Manufacturing can solve the problem of poor integration of traditional manufacturing in terms of data perception,intelligent processing,human-computer interaction,and collaborative manufacturing,and it is a key technology for realizing manufacturing production intelligence.Manufacturers in the actual production process will produce a large number of raw data representing real-time production status.These data streams with multi-source,massive,and real-time characteristics add to the burden of data computing services.Therefore,how to efficiently and accurately obtain valuable information from the collected data of the Internet of Things is a key issue for realizing efficient and rapid development of the manufacturing industry.In order to extract valuable information from massive real-time data streams,domestic and foreign researchers have proposed many solutions.Most of the schemes use complex event processing mechanisms to convert the original data in the data stream,that is,the original event,into a composite event containing a certain business logic relationship,and provide the input data to the upper application.However,the manufacturing environment contains multiple mobile,multi-association and multi-cooperative production equipment,which brings many difficulties to the real-time and accuracy of data processing in the production process.The research of this paper is mainly used to manufacture mass information processing in the Internet of Things.Firstly,a high-priority event detection algorithm based on probability is proposed to solve the problem that high-priority events generated by multiple sources in the Internet of Things cannot be detected in time.Analyze the priority scheduling mechanism in real-time event flow,establish a priority event processing model,and propose a probability-based event priority decision algorithm for high-priority events.The algorithm can quickly handle high-priority events and reduce detection delay.Then,for the event detection method of centralized serial event detection in manufacturing Io T,there is a problem of low detection accuracy and slow detection efficiency.An event tree-based event partitioning algorithm and parallel complex event processing architecture are proposed.This method takes full account of the multiple dependencies among atomic events in the original event stream.The event tree is used as a node to construct a linked list of dependent event trees,which improves the number of valid detection times of the complex event processing engine.Meanwhile,the method increases the event processing throughput while reducing the use of memory space.This paper mainly focuses on the efficient and complex event processing method of massive real-time data in the manufacturing Io T environment.It uses the test data actually produced by manufacturing companies.Chapter 5 illustrates the superiority and feasibility of the proposed method on mass data processing problems through simulation experiments.Therefore,applying the algorithm proposed in this paper to the manufacture enterprises can effectively improve the company's ability to handle massive amounts of real-time data streams,which also can solve the bottleneck problems encountered by enterprises in a timely manner,and make certain contributions to the manufacturing of IoT technology.