The Algorithm Research And System Implementation Of Business Instruction Cluster Discovery For Complex Mainframe

With the rapid development of civil aviation business,the early mainframe system built by a civil aviation information service company based on mainframe gradually shows some shortcomings in adaptability and scalability,and the open system based on distributed technology showed advantages.In addition,the company’s system architecture is a binary system,which brings difficulties to the company’s system operation and maintenance.In order to solve these problems and reduce operating costs,the company has established a strategic direction to systematically promote the migration of system business function instructions from host system to open system.However,if the cooperation between instructions is not clear,there may be a situation that the migration of one instruction affects other services in the process of migration,and there is a major hidden danger of business system failure.Therefore,before instruction migration,the cooperation between instructions must be combed out to find the accurate business instruction cluster.At present,companies usually use manual methods to find business instruction cluster,but manual methods are time-consuming and hard to carry out,and only rely on the experience of business personnel,without actual data support.Therefore,in order to improve the efficiency of discovering the business instruction cluster,we dig out the cooperation relationship between instructions from the instruction operation sequence data constructed in the execution log of the host system,and then discover the business instruction cluster,which lays the foundation for the migration of the host system instruction.Because the business instruction cluster based on the automatic discovery algorithm cannot guarantee 100% accuracy,in this paper,we divide the solution of finding the correct business instruction cluster from the instruction operation sequence into two stages: Firstly,the service instruction cluster based on the algorithm can be found as accurately as possible;Secondly,a visual prototype system of business instruction cluster based on the data of business instruction cluster,is designed and developed to assist the business personnel to check and further sort out,and then get the accurate business instruction cluster.The specific research of this paper is as follows:(1)The solution for automatic discovery of business instruction clusters based on algorithm.In view of the fact that automatic discovery of business instruction clusters based on algorithm,which can greatly reduce the workload and difficulty of manual combing,this paper proposes a solution based on sequence multi-hop structure network strategy.In this scheme,we first use the sequential multi-hop strategy to capture the cooperation relationship between successive instructions in the sequence,so that the constructed instruction cooperation network has an obvious sub cluster structure,and then use the community discovery algorithm to discover the business instruction cluster.The verification of real and manual instruction operation sequence data shows that this scheme can effectively improve the accuracy of automatic discovery of business instruction clusters.(2)Design and development of a visual prototype system.The visual prototype system is a tool to assist the mainframe’s business personnel to check and sort out the business instruction cluster based on automatic discovery algorithm.In this paper,we analyze the actual needs of business personnel in detail,and design the functional modules,organizational structure and operation process of the system,and elaborate the design of system data model from the three levels of conceptual model,logical model and physical model,and show the effect of system operation at the end of the paper.