Servers Per Fte

Metric Type

Productivity

What Does It Tell You?

Unix servers per FTE, like all productivity metrics, is a measure of efficiency. The higher Unix servers per FTE is, the less labour is required for each server, and thus Cost per Unix server and Total Cost of Ownership will be lower.

How Easily Can It Be Compared?

Unix servers per FTE is difficult to compare easily, unless there is a very precise definition about which server management tasks and activities is included. Unix server management can include any of the following:

  • Setting technical architecture and standards for Unix servers
  • System integration and engineering work to set up a new application on a Unix server
  • Operational console management (monitoring and acting on alerts)
  • Unix system administration (user id administration, patch management, capacity monitoring and performance tuning)
  • Reactive incident and problem management
  • Database management
  • Storage management
  • Backup, archiving, and recovery.
  • Batch management (setting up, monitoring, and resolving problems on batch jobs)
  • Applications operations.

A simple way of calculating Unix servers per FTE is to take the head-count in the Unix server management team and divide by no-of-logical-servers. However, the results will depend primarily on how many of the tasks above are the responsibility of the server management team in this particular organisation. A team doing only console management and system administration will report a very high Unix servers per FTE, perhaps 200 or more. A team where database management and applications operations is integrated will report a much lower Unix servers per FTE, perhaps 30 or less. However their productivity is not necessarily lower, since they are taking on many more responsibilities.

How Do You Collect It?

There are two main components:

  • FTEs for Unix server management, which must be collected either through time-recording, or through FTE calculations.
  • Number of logical Unix servers

How Is It Calculated?

Number of logical Unix servers@@
------@@
//FTEs for Unix server management

How Frequently Should It Be Reported?

Annually, or quarterly at most. Expect minor fluctuations from month to month, as new software releases or operating problems cause peaks of work. Monthly reporting is therefore not very informative, whereas trends over several quarters or years are very useful.

Dangers, Traps, & Pitfalls

  • Managers working this number down
  • Isolating teams to increase the metric for each
  • Playing the "out-of-scope" game
  • Managing development environments
  • General FTE utilisation issues

How Is It Related To Other Measures?

.
See also

  • FTEs per hundred Unix servers
  • Hours for Unix server management per Unix server per month

both of which measure the same thing using a different calculation method.

Rules Of Thumb (Red/Yellow/Green)

Public Domain Sources

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License