The Study Of Model And Algorithm Based On Complex Event Processing

With the rapid expanding of demand for real-time processing of large scale complex data,Complex event processing(CEP)is a technology which is used to deal with a large number of data streams in real time to meet these needs.It considers each set of data which has specific meaning in stream as an event,maps the query requirements defined by user to a specific recognition model structure.Then the CEP analyzes and extracts the event sequences that meet the query demand from the continuous events stream,to achieve real-time and efficient analysis of large amounts of data.But the existing CEP model is insufficient and disadvantages of inefficiency for algorithm.Therefore,improving the performance of CEP is the focus of research.Based on the analysis and summary of the existing CEP models and algorithms,the following research work is carried out:1.Aimed at the original model of non-deterministic finite automate with matching buffer NFAb is only able to processing the kleene closure operator,this paper proposed the model of kleenePlus Non-deterministic Finite Automate with Run Buffer kNFAR,to support kleenePlus operator.The construction of the model and implementation process of the running buffer are adjusted to realize the processing of more complex relationship events.2.Aimed at the processing efficiency is influenced by a large number of run that created in the process of the kleene operator,the postponing algorithm P-kNFAR based on kNFAR model is proposed:Only process the timing constraints evaluation but not copy run buffer when handling the events of kleene operator to reduce number of run and storage cost.Predicate constraints and kleene constraints are executing when the run reached the final state to improve the throughput of complex event processing.The experimental results show that the proposed algorithm P-kNFAR has more efficient processing performance compared with the algorithm of NFAb model.3.Aimed at the feature of event streams and different constraints defined in the pattern matching based on the pattern matching of tree method,this paper adjusts the order of pattern matching to improve processing speed,proposing a tree based optimization sequence matching algorithm of OMSTree.Analyzing the factors that affect the processing cost,finding the order of minimum cost and the matching results are constructed by scoping parameters,which can effectively improve the processing efficiency.Experimental results show that the algorithm OMSTree is effective in improving the efficiency of CEP processing compared to the traditional pattern matching algorithm.