High Concurrent Performance Monitoring System Based On Real-time Visual Data Mining

In software engineering domain Concurrent Operation means that multiple programs run on processor at the same time in micro-analysis, or multiple users operate some system at one time in macroscopic analysis. High Concurrent Operation is that high number of programs or users run on system at the same time, and it will generate high load pressure. Now High Concurrent Operation systems are used in many field including portals station, telecommunication,3G mobile platforms, traffic and some hotspots such as Cloud Computer.Since such kind of system not only pay attention to the function but also if can run normally under high load concurrent pressure, so it is necessary to do performance test before delivery and monitor in formal work envirement. But now most performance test and monitor are operated in respective evirment and tools, it leads to the analysis data and the experience from testing can't be applied in monitor and also can't make good estimation and prediction for the performance situation. The opinion that merge test and monitor to one performance testing-monitor system is held in the paper.At same time many High Concurrent Operation systems need get or process data in Real-Time. The feature of Real-Time not only requires the stable working under High Concurrent envirement but also get performance data synchronously. It also need deirct visual monitor for performance tracking, warning and preditiction and get perfect information expression. In order to satisfied the above technologis requested by High Concurrent Operation system testing and monitor, it need corresponding Real-Time and Visual data mining technical settlements, So it will have practical significancy to develop the high concurrent performance monitoring system with visual data mining under real-time.The paper will adopt the technology of distributed compute base on time sequence and the arithmetic of BP Neural Network in order to process data accurately and analysis performance data effectively. This will make test-monitor data collected and statistics analyzed efficiently according to the requirement of High Concurrent Operation system test and monitor. It also can extract training sample from performance testing analysis data for Visual data mining when make real-time performance monitor. That will gain the goal of visual performance analysis; situation prediction and warning monitor graphic view effectively.The paper state the necessary point to develop and apply the High Concurrent performance monitoring system with visual data mining under real-time, that point just from the promblem met in now industry and during High Concurrent system testing and monitor. The paper will anlysis and research the technology of visual data mining uner Real-Time and the High Concurrent performance testing and monitoring system respectively, dissertate how to apply the technology of distributed compute base on time sequence and BP Neural Network into High Concurrent performance testing and monitoring system, the detail design and implementation will be acted base upon points. VI...