Improving Systems Performance › Performance Architecture › Main Performance Architecture Components › Performance Distribution Server

Performance Distribution Server

Performance Distribution Servers request configuration data from the Performance Domain Server and deliver it to the Performance Agents. As this data is centralized at the Domain Server level, Distribution Servers operate without the need for any local persistent information. Therefore, you can install or re-install them without the need to backup and restore data.

A key function of a Distribution Server is to manage performance data for the machines for which it is responsible and maintain this data in its local cube store. Each Distribution Server communicates with the others and is aware of the contents of their cube stores. In fact, to operate efficiently, each Distribution Server must have knowledge of:

• All the performance cubes to which it has access in its local cube store.
• All the monitored machines for which other Distribution Servers have data.
• Which Distribution Server is the primary source of data for each monitored machine.

To achieve this, each Distribution Server examines its local cube store and builds up an index of the machines for which it has cubes, the cubes that exist for each machine, and the resource metrics for which there is data in the cubes. The Distribution Server then passes its index to the other servers with which it is registered so that they all have up-to-date information on each other.

When you use an application like Performance Scope to request performance data, the application submits the request to any available Distribution Server. The server examines its local cube store and then either returns the requested data or, if it is not the most appropriate server to handle the request, forwards the query to a more suitable server.