Design And Implementation Of Business Decision System Base On Business Engine

In today’s society, companies in the development process continue to accumulate more and more data, but for the utilization of these data, it is very inefficient, so every company always expects to get some information from these data to help him to analyze business information, and the information from these data always can help the leaders to make the auxiliary decision in a way of favorable judgment, along with the needs of these companies,Business intelligence technology is attended, over time, all the companies make great progress in the field of business intelligence and develop a great many business systems or software, such as open source software for data integration named Kettle, data mining software named Weka, etc., not only that, but also develop a traditional architecture of business decision system, these technologies,to some extent, help each company or organization to obtain information from lots of historical data in a more efficient way. But in the architecture of traditional BI systems, they still don’t implement business rules separated from the operation in a BI process management, resulting in the requirements of a professional skill to user these tools, so that while a Business Analyst can not only focus on business analysis but also take great effort to become familiar with these specialized tools to analyze which is obviously harmful to improve our effectiveness. For this reason, businesses and companies have a very urgent need to find a new way to realize business rules separated from operation of the BI system, and then we analyzed the workflow technology in business management above, so in order to meet this demand, in our BI system, To complete the design of a service-based workflow engine technology is our primary problem to solve.There are still existing problem in traditional BI systems, although the BI system is enough to meet the basic needs of business analysis, but due to the traditional BI systems for BI operations is scheduled for execution with a serial mechanism, so that it comes to making business analysis on efficiency bottlenecks and especially when we are dealing with the amount of data becoming extremely huge, such a bottleneck will appear as obvious, so Traditional BI System is clear lack of the a BIlity to handle massive amounts of data, generally speaking, concurrently scheduled execution’ efficiency is always higher than the efficiency of serial schedule, so we consider to change the way of the serial schedule in the traditional BI systems, we decide to use the way-Action Workflow concurrent Schedule to improve operational efficiency of the entire BI system.This paper as a research project in cooperation with a company, and the major work is as follow:The first study is to learn the basic principles of workflow engine and its architecture, followed, we should make a deep analysis on the implementation of the execution mechanism of solution files in traditional BI system, according to the results of our analysis, we complete to bound the workflow technology and the execution mechanism of solution files, and then design business engine in our BI system.And then in order to optimize the concurrent scheduling action workflow,we make a detailed analysis of the implementation of mechanism with the action operation under the traditional BI systems, and for ETL operations,we are doing a full analysis,and then We presented rules with dividing unrelated set of ETL operations, and presented concurrent scheduling algorithm, not only do this but also we still make a full analysis on the new way of schedule Action operation, and we found that BI system can’t response error timely while we use the new way of scheduling action, so we present a new Centralized Detection Mechanism, then we completed a ActionFlowEngine design and implement.Finally we verify the correctness of our improvement through the experiments.