Usage analysis involves identifying the users of your system and determining the usage patterns for those users. In doing so, a usage analysis enables you to determine expected load conditions on your directory service.
Your reasons for offering Sun Java System Directory Server Enterprise Edition as an identity management solution have a direct effect on how you deploy the server.
During usage analysis, interview users whenever possible. Research existing data on usage patterns, and interview builders and administrators of previous systems. A usage analysis should provide you with the data that enables you to determine the service requirements that are described in Chapter 5, Defining Service Level Agreements.
The information that should come out of a usage analysis includes the following:
Number and type of client applications. Identify how many client applications your deployment must support, and categorize those applications, if necessary.
Administrative users. Identify users who access the directory to monitor, update, and support its deployment. Determine any specific administrative usage patterns that might affect technical requirements, for example, administration of the deployment from outside the firewall.
Usage patterns. Identify how various types of applications access the system, and provide targets for expected usage.
Answer the following questions, for example:
Are there times when usage spikes?
What are usual business hours?
Are client applications distributed globally?
What is the expected duration of application connectivity?
Client application growth. Determine if the number of client applications is fixed or expected to grow. If you anticipate additional applications, try to create reasonable projections of the growth.
Application transactions. Identify the types of transactions that must be supported.
These transactions can be categorized into use cases, for example:
What tasks are performed by the applications?
When applications bind to the directory, do they remain bound, or do they typically perform a few tasks and unbind?
Studies and statistical data. Use preexisting studies and other sources to determine patterns of application behavior. Often, enterprises or industry organizations have research studies from which you can extract useful information about users and client applications. Log files for existing applications might contain statistical data that is useful for making estimates for a system.
For more information about usage analysis, see the Sun Java Enterprise System Deployment Planning Guide.