Research On Key Technologies Of Information System Integration Methods And Complex Event Processing

With the rapid development of information technology, the complexity and scale of integrated information systems are increasing. Cooperations between several systems are required to support a complex business activity. Therefore, new integrated system architecture and event processing approach are needed. Currently, complex event processing has become a hot issue in real-time event processing field. It has vital significance to improve the capability of event processing in integrated systems by using this technology.There are several known issues in this field, such as the complexity of integrated architecture, the poor performance of event processing method, as well as the slow speed of data analyzing in integrate systems. With proposed methods like information system integration architecture, permission certification method, and automata based incremental processing approach are proposed, an intelligent data analysis system is designed and implemented in integrated systems. The main research work of this dissertation is shown in the following:(1) To solve the problems of the complex architecture of integrated systems, the low efficiency of portal, and the low speed of responding the permission authentication in integrated systems, an SOAE (Services Oriented Architecture with EIP) architecture based on SOA (Services Oriented Architecture) is proposed. SOAE is an information system integration architecture which adopts a modularized EIP (Enterprise Information Portal) as its core component, and uses ESB (Enterprise Service Bus) as its information transportation component. A permission authentication method is presented to support multi-user concurrent access integrated systems. SOAE integrates information systems through a high performance portal, which could deal with concurrent access requests from multiple users. The experimental results of pressure tests show that EIP of SOAE has the capability to process concurrent access requests efficiently for permission authentications.(2) Focusing on the architecture complexity of cloud services integrated platform, and the low efficiency of authentication mechanism, a cloud services integration architecture based on CSP (Cloud Service Portal) is proposed. CSP is a layered and modularized portal, which supports different storage patterns of run-time data model of users. In order to improve the respond speed for end users, an authentication mechanism based on CSP is presented, which is efficient to deal with the concurrent access requests of massive users. The experimental results of pressure tests show that the integrated architecture based on CSP does not only achieve the rapidly respond for user requests, but also achieve the rapid permission authentication for cloud services.(3) To solve the problems of the multisource and heterogeneous events in integrated information systems, a novel event processing approach LAIP (Lightweight Automata based Incremental Processing) is proposed. In order to model the events of integrated systems, adopting an event algebra in integrated systems, LAIP uses event models, operators and expreesion to define the relationship of events. Based on event algebra, an event processing language LAIPL (Lightweight Automata based Incremental Processing Language) and event processing rule structure are proposed, which are used to define the complex event in LAIP. By improving the traditional AIP (Automata based Incremental Processing) approach, LAIP translates LAEPL rule into automaton, and leverages LAIP algorithm to achieve the real-time and high performance processing of a great number of input events.(4) Focused on the problem of the low efficiency of event processing engine, a LAIPE (Lightweight Automata based Incremental Processing Engine) based on LAIP approach is proposed. Leveraging the advantages of LAIP in integrated systems, LAIPE engine adopts rule translation algorithm in LAIPL rule translation, and uses LAIP algorithm to maintain each automaton instance to process the input events in real time. The layered architecture of LAIPE engine describes its design and implementation. Comparing with a normal CEP engine Esper, experimental results show that LAIPE engine can deal with massive input events with higher throughput and better performance.(5) Aiming to solve the problem of real-time analyzing massive data in integrated systems, an intelligent data analysis method based on complex event processing technology is proposed. In integrated systems, data is packaged as events, whose operators and processing rules can be used to describe the data processing conditions. Improving LAIPE engine, the approach adopts LIDA (Lightweight Intelligent Data Analysis) algorithm to build a lightweight intelligent data analysis engine (LIDA-E). Based on the intelligent data analysis approach, a lightweight intelligent data analysis system (LIDAS) is proposed to real-time analyze massive data in integrated systems. LIDAS adopts an event transmission component of agent and filter to collect and transport data in real-time. According to the experimental results, adopting the intelligent data analysis approach, LIDA-E engine has better efficiency of analyzing data by comparing with Esper engine.