Implementation Of Commonj.WorkManager On Top Of Platform Symphony

This paper presents a Implementation of Commonj.WorkManager(a threadpool architecture) on top of Platform Symphony APIs ,by this way can enhance the performance of Web Enterprise Computing Based on J2EE.To entend application of Commonj.WorkManager from WEB Server to computer cluster.To distribute the user works into different hosts in a cluster instead of a multithreading manner on a single computer.The WEB Application Server Cluster which is use of multi-threading technology, distributed mission to different node server, and the node server processing tasks, respectively. Multithreading is the advantage of rich-useful interfaces, more mature technology, a large number of engineers. But it can not be used to deal with a host collaborative tasks, or migrated the task from the high load server to free server if any server have no task to processing. And these are precisely the advantages of the technology cluster, but no clustering technology like multi-threading technology which have a variety of useful interfaces, and the need to prepare special procedures for clusters. The goal of this paper is combining the advantages of both technologies to avoid its shortcomings.Thread, which is inappropriate for use by applications hosted in managed environments such as EJB containers and Servlet containersThe WorkManager for Application Servers specification provides an API for application-server supported concurrent execution of work items. This enables J2EE-based applications, including servlets and EJB apps, to schedule work items for concurrent execution. This will provide greater throughput and better response time. After an application submits work items to a WorkManager for concurrent execution, the application can gather the results. The WorkManager provides common "join" operations, such as waiting for any or all work items to complete.Cluster is welcomed more and more because it has higher performance. Cluster is one kind of parallel or distributed system. It is composed of a set of independent computers which are linked. They worked as a single integrated computing resource. This computers may be a single computer of multiple processors system ( such as personal computer, workstation or SMP). Every node has its own memory , I/O devices and operating system. Cluster is a single system to the suers and applicationg. It may apply high performance environments which are low costs and high efficiency and quick and avalaible services.Many applications have a large number of independent calculations that can be done in parallel. Platform Symphony can significantly reduce processing time by distributing the calculation in parallel. These calculations are in the form of loops over many iterations where each iteration is independent. For example, a financial application may evaluate the value of a portfolio under different scenarios. Instead of iterating through the calculation sequentially on one CPU, the iteration is handed off to Platform Symphony and distributed across the grid to run in parallel across a number of CPUs.Platform Symphony is a service-oriented grid middleware for application orchestration and workload management. If you want to deploy a user-defined programs onto Symphony you must follow the specification of Symphony APIs. Symphony WorkManager's target is to integrate with the specification of Commonj.WorkManager with the Symphony Java APIs.To reach the goal of entend application of Commonj.WorkManager from WEB Server to computer cluster, we analysised the Commonj.WorkManager Architecture and Asynchronous Symphony client model.we found the Implement method and design a Symphony WorkManager architecture.Moreover, our implementation of WorkManager -- Symphony WorkManager, is able to distribute the user works into different hosts in a cluster instead of a multithreading manner on a single computer. The Symphony WorkManager is a middleware between user and Symphony Clusters, and it is built on top of the platform of Symphony APIs. It can be regarded as the peremptory input/output message transferred between Symphony Cluster and client.Compared case of used Symphony WorkManager (asynchronism method) to not used it(synchronization method) on same condition, the results approve used Symphony WorkManager more efficient saved the cost of times.This paper studied and implemented of Commonj.WorkManager on top of Platform Symphony. We exended application of Commonj.WorkManager from WEB servers to computer cluster, but we do not consider how to manage work items and thread pools efficiently.