Font Size: a A A

Research And Implementation Of The Enterprise Job Scheduling System Based On SaaS

Posted on:2009-08-15Degree:MasterType:Thesis
Country:ChinaCandidate:T B TangFull Text:PDF
GTID:2178360272974522Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
"Job" refers to a task that can be executed on a computer, usually corresponds to an executable program, like exe, jar, bat etc. Job usually has no UI and runs in the background, normally doesn't involve user interaction but needs starting by user.Usually, there are many jobs in modern enterprise applications. It's definitely a waste of resources to start all jobs manually. What's more, as the increase of job number and business complexity, and different requirements for different jobs on execution time, condition, scenario and rate, some jobs can't be operated manually. Meanwhile, it's also difficult to monitor jobs because they run in background. In addition, we must make sure jobs run parallelly for batch jobs and ensure availability of system for significant jobs. All issues mentioned above show, a friendly enterprise job scheduling system is needed for better managing, executing and monitoring jobs.This paper derived from the Service Building Group project "Countryside Bank (CS-Bank)" in IBM China Research Lab. There are many batch jobs in CS-Bank, such as, "calculate interests everyday", "generate financial analysis reports every month", etc. The purpose of this paper is to design and implement an enterprise job scheduling system based on SaaS(Software as a Service) which can be used in different scenarios according to validating basis of CS-Bank.First, the SaaS model and technologies used in this paper were analyzed, like J2EE, Struts, Ibatis, Ajax etc. Secondly, according to the analysis of different scenarios, a job scheduling model that can be used in different situation was raised. Through the research on thread, thread pool, cluster and load balance algorithm, a parallel solution and "Next Least Load" algorithm were given. The feasibility of the parallel solution was proved by an experiment. By the study of SaaS data model, the "Shared Database, Independent Infrastructure" model was selected. Then, by following the standard software engineering workflow, requirement analysis was done; system infrastructure and database design were given; detail design of functional components were described and the system was implemented base on J2EE. And last, the enterprise job scheduling system was verified in CS-Bank and exploration on widespread use was done.
Keywords/Search Tags:Enterprise Application, Job Scheduling, SaaS, Cluster, Load Balance
PDF Full Text Request
Related items