File/storage tier

Dimensions CM securely manages the content of assets. It stores files and retrieves tasks using the agent technology.

Item library

When assets are checked into Dimensions CM, their content (the versions of source code, documents etc) is stored in a secure location called an item library. The machine hosting the item library must have a Dimensions CM server or agent installed on it. Different item types can be placed in different item library locations, for example:

  • Separate directories on the same machine.

  • Separate physical machines that host item libraries for different item types.

When an item library is accessed, the library server process (dmlibsrv.exe) runs as the following user:

  • Windows: SYSTEM

  • UNIX (when the item library is on a different machine to the server): root

You can also configure Dimensions CM to specify a non-privileged user account to run the library server process and to access item libraries that are local to the server directly from application server processes. For details about using the following configuration symbols, see Administration.

  • DM_LIBRARY_USER_USER_<product id>_<item type> <username>


You can use the following methods in an item library to store the content of files:

Flat files One file for every revision of an item held in Dimensions CM. This is referred to as a normal item library. This is the recommended method for storing file content as it offers the highest performance and lowest overhead.
Delta storage One file per item that contains all revisions of the item. Only the differences between each revision are stored. This method is useful if disk space is limited.
Compressed Content is stored in flat, compressed files. This type of item library is referred to as a normal, compressed item library. Some CPU overhead may occur when accessing and extracting data.

Back to top

Remote work and deployment areas

You can use remote work areas in addition to local ones. For example, a user is running the desktop client but is developing software on a mainframe therefore assets need to accessed from partitioned datasets (PDS) on a mainframe MVS file system. A Dimensions CM agent must be installed on the mainframe so that Dimensions CM can interact with it.

When a remote work area is accessed, the pool manager on the Dimensions CM agent node attempts to log in with the credentials you used to log in to the server. If the login fails, the server prompts the user for credentials to access that work area. After a user has entered their credentials, they are not prompted again unless they log out or their session expires.

When a deployment area is accessed, if the Dimensions CM administrator chose to store credentials for that area then those credentials are used. If they were not stored the user is prompted.

Back to top

Item library cache

Software development teams are often geographically dispersed, with remote users connecting to a main site across a Virtual Private Network (VPN) utilizing a Wide Area Network (WAN). However, VPN connections over a WAN can have low bandwidth and suffer from a high latency. A single application message may take a long time to travel over a slow connection. If the protocol utilized by the users' applications at a remote location is "chatty," the remote users typically experience more delays and slower response times.

To address the problem of WAN latency, Dimensions CM uses a library cache that improves performance by caching file contents on a network node that is geographically close to the remote users. When users update files, they are transferred from the library cache not the Dimensions CM repository, reducing transfer times by eliminating the geographical distance and resulting network latency. If a project has many large files it is more likely to benefit from using library cache areas.

Personal library cache directories (PLCD) and delta compression on file transfers can also provide significant performance improvements for geographically distributed development teams.

For details, see Improve WAN performance.

Back to top

See also: