Research And Implementation Of Resource Tuning For J2EE Application Server

J2EE Application Servers are popular servers side solutions for many Web-enabled e-business applications. Software Application performance directly linked to the allocation of resource used by the J2EE Application Server. However, due to a lack of resources self-tuning capability in J2EE application servers, J2EE application servers have to be configured manually to achieve good performance. This kind of manual tuning is time consuming and requires highly skilled personnel. Web-enabled e-business applications can be affected by many unexpected factors during runtime. At the same time, it has different performance characteristic under different workload pattern condition. In addition, since the workloads of e-business sites tend to vary dynamically and exhibit short-term fluctuations even if the system resources is tuned well at one point in time, it will show poor performance at other times.The thesis presents a solution for automatically optimising the resource of J2EE application servers. The proposed approach is based on feed-back control theory and some experience. J2EE Application Server performance is regulated by adjusted J2EE Application Server resource parameters. The automatical mechanism uses runtime monitoring data to detect performancea nomalies and significant variations in the application's execution environment. It automatically use the optimal resources configuration under the current running conditions. We present the optimization algorithm used to adjust these parameters of J2EE Application Server. By comparing the experience stored in a database, the mechanism of tuning system resources is able to find appropriate parameter configurations more rapidly. An architecture of JBoss Application Server has been analyzed, A prototype based on JBoss Applciation Server (JROS) has been implemented.A prototype based on JBoss Applciation Server has been tested for automatically managing a J2EE application. Obtained results prove the framework's capability to successfully manage a software system without human intervention.