Research And Implementation Of Data Integration Middleware For Automobile Industry Chain Based On Web Service

With the rapid development of the economy,the number of Chinese automobile enterprises is showing a sharp increase.In order to survive and develop in the fierce competition,many enterprises have their own unique information system,but also because of this unique characteristics,hindered the information exchange and cooperation between the various automobile enterprises,formed an "information island." Because of the rapid development of automobile industry,the constant accumulation of data,the urgent need for cooperation between enterprises,"information Island" has become an urgent problem to be solved.After years of research,data integration is the best way to solve this problem,but it needs to solve the problem of heterogeneous heterogeneity of distributed data.How to solve the data semantics and structure difference caused by heterogeneous heterogeneity is an important research topic in the field of data integration application.The reason of data heterogeneity is that most enterprise business is different,the database is not type and the data representation is different.By contrasting the solution model,it is clear that middleware is a best solution to solve the interoperability problem between heterogeneous data.Based on Web service technology,this paper studies the data integration middleware system of heterogeneous database.The system uses global mode to map the information of each child database and uses an XML schema to define an xml-based global schema.The middleware system passes the global request to the request decomposition module to decompose to the child service request,after the child service request is received by the service invocation module,locates the service provider registration related service in the Web Service registry,connects the database,and then transfers to the Scheduler module to arrange the order of the task execution.Finally,all the results of the query result synthesis module merged into the complete query results.In order to solve the problem of user concurrent access performance,a multistage feedback queue scheduling algorithm based on process execution time is designed in the Scheduler module.Based on the multilevel feedback queue scheduling algorithm,the algorithm improves the number of context switches,compares the process execution time with the queue time slice,puts in the ready queue according to certain rules,and then assigns the CPU to perform all processes.In order to test the performance of the algorithm,the experimental data areobtained by using the process of the schedule simulation algorithm,and the experimental data are obtained by using JMeter to simulate individual user access and multiple concurrent user access system in order to test the system performance.Through the analysis of experimental data,it is shown that the multilevel feedback queue scheduling algorithm based on process execution time has advantages in turnaround time,average waiting time and context switching times,and achieves the expected target of algorithm design.After analyzing the data of the JMeter experiment,we get the conclusion that the system has strong load capacity,and the concurrent access response is stable,finally,it can meet the needs of the users.