Performance Tuning Your Application

   

The key performance issue for distributed applications is not tuning raw code execution speed, but rather determining the optimum physical packaging of components, the best resource sharing model, and the best network topology for an effective distributed solution. With APE you can model various deployment options by putting all of the services on the same machine, putting the business processes and the client services together on one machine with the data access on another machine, or by defining many other physical deployment options. Additionally, you can model the performance and scalability benefits of both queuing and object pooling.

What APE is really measuring isn't so much the work that your application will do (which APE can only simulate), but rather the capacities and throughput performance of the distributed infrastructure your new application will use. Because the performance of this infrastructure is often a critical factor in the success of any enterprise application, the use of APE early in the design phase of a new application can significantly improve its chances of success.

To understand the significance of the various design factors APE can benchmark, you need to learn how remote components can be deployed either synchronously or asynchronously. Each of these approaches has its advantages and disadvantages, and different parts of your application can use different models, depending on the needs of the tasks being implemented.

The following topics introduce the common performance issues of a distributed application, describe the basic APE remote deployment models, and explain certain performance tuning concepts that are important to your enterprise application.

Section	Description
Understanding How Distributed Components Affect Performance	Describes how component size and location can affect the general performance of your application.
Understanding APE's Remote Deployment Model	Describes the three APE deployment models: synchronous, asynchronous, and queued.
Common APE Performance Comparisons	Examines the test results of three basic APE profiles.

For More Information   For more information on performance tuning from a data access perspective, see .