After briefing our management teams with the preliminary data, we were ready to enable the power management feature of NightWatchmanSever Edition on our servers. This was an interesting moment in the project as we realized that many of the servers we had selected to date did not have power management capabilities built in and therefore, were too old to make use of allow use of the 1E tool’s power management functionality.

This is because the power management capability of NightWatchman Server Edition relies on the capability of the CPU to “throttle” – technology that has only been available in servers for about the last 5 or 6 years.  It so happened that on our pilot to date, the CSC servers using the tool were not our most recent hardware, and only 7 of our 79 servers had the ability to be power managed. Industry norms are that roughly 25%-40% of servers in average data centers will have this capability. Only 10% of the servers on our pilot are capable of taking advantage of power management.

To ensure that the tool capabilities were fully leveraged during the pilot, we decided to verify that the remaining servers onthe pilot had this functionality before allowing them to participate.  Where necessary, we refused older servers, including only servers that could be power managed.  Our end result would be 28% of our pilot was on servers that would allow power management – much more within the industry normal range.

Read the complete post at http://www.1e.com/bizblog/post/2010/09/07/Part-7-CSC-on-Power-Management-with-NightWatchman-Server-Edition.aspx