The key hardware considerations (in order of priority) for the project repository are the following:
-
Memory Requirements: Memory is one of the cheapest ways to improve the performance of a CM Tool. A
rule of thumb for how much main memory is required in the server machine is to add all the database space used by
the project repository, and divide by two.
-
Disk Input / Output Requirements: The second most likely performance bottleneck in the CM
environment is the speed at which the data can be written to disk. Read/write intensive operations are check-in,
check-out and baseline creation. It is a good idea to have a dedicated controller and channel per disk.
-
Network Bandwidth: Since the CM tool is usually a distributed application, adequate network
capacity and reliability are required for good performance. The recommendation is to put machines hosting the
project repository and views on the same subnet. And if the local area network (LAN) is too saturated as indicated
by time outs and poor response, the idea is to increase network capacity or add a subnet for the CM tool hosing
machine.
-
Project Repository Disk Space: Depending on the size of a project there could be multiple project
repositories, and each project repository could contain tens of thousands of files and directories. The number of
files in any given project repository will depend on the size of the machine on which the repository server is
running, and the number of users expected to concurrently access data.A good rule of thumb is to allow disk space
for growth, and have about 50% free space by allocating 2 giga-bytes of storage per project repository.
|