| The Mainframe of IBM Company is a high stable and high performance product.Many high level enterprises like its classic hierarchy and the ability of the parallel.The Mainframe's hardware has been update several times in about ten years,the advance not only on speed but also on reliability.But for some reason,the hierarchy didn't change a lot in the past.To compare with it, the enterprise also not changes their applications which run on the mainframe.Then the problem raise now,for some history reason,some application has problem when they were being coding. And now due to the boost of the customer,these programs and the VSAM which they use can not meet the performance requirement.It's also impossible to change the original program.The only two ways to resolve the problem is to optimize from the system level or to migration data from VSAM to database.To migration is a long work,it can not be resolve the current problem.So to optimize is the only way which can release current press.As the major part of the Mainframe,VSAM always takes the largest resource.And for the attribute for the VSAM,it needs to visit the disk equipment.So the performance of the VSAM impacts the application performance directly.If we can decease the times for the visiting from VSAM to hard drive,then it will be enhance the whole system's performanceThis article focuses on the method for how to enhance VSAM performance automatically from system level.Base on the weakness of the man made enhance(human mistake,high workload),I give a new method to enhance the performance of VSAM.Through a cycle which out of the Mainframe,it can collect information from VSAM,analyze the collected data,and give a response to how to improve the VSAM and then give a job flow to realize the enhancement.The major contribution of this text as follow:â—Base on the theory of finite automata,focus on the report of the VSAM performance, create an automatically method the resolve the problem of translate the file data into database.â—Used Data mining method to find the similar file and select the best learning file.And then modify the original file to improve.â—Though the time computing and task controlling to perform the max effect in turned file in limited time. |
