HP ML150 HP Power Capping and Dynamic Power Capping for ProLiant servers techn - Page 17

Using power capping in data center provisioning - proliant g5 server

Page 17 highlights

Using power capping in data center provisioning Power capping is a tool to help IT organizations manage infrastructure size and costs in the data center. By setting a power cap for a server or group of servers, a data center manager can be assured that unexpected changes in workload or environment will not cause servers to consume more than the specified amount of power. Appropriate use of basic Power Capping in the data center lets administrators provision the cooling infrastructure at an effective level-a level that meets the cooling requirements associated with the server power consumption at a given application load and power cap rather than at the level needed to handle all servers running at full power. Alternatively, using basic Power Capping to limit average server power consumption will allow more servers to be safely operated within a pre-existing cooling infrastructure. Dynamic Power Capping takes this one step further. Its ability to control server power consumption in real time allows administrators to use power capping in planning and managing their electrical provisioning as well as their cooling infrastructure. Choosing effective power caps Setting an effective power cap involves determining the lowest value for the power cap that will still meet the power requirements for a server or group of servers running their application workload. Setting such a power cap should have no impact on server application performance but would reduce the required cooling and electrical provisioning for the servers. Administrators can use the HP Power Calculator Utility to estimate the maximum input power for a given server configuration. This information provides a starting point for considering power caps since it enables the data center administrator to determine the maximum power and cooling needed to support a server or a group of servers. Additionally, the reporting features in IPM and iLO provide historical power consumption information that can be used to select the most effective power caps to achieve specific power savings or capacity planning targets. Figure 10 shows the output from the HP Power Calculator utility for a ProLiant DL380 G5 server configured with two Quad-Core Intel Xeon X5460 3.16-GHz processors, one 72-GB disk drive, and 8 GB of system memory. The results indicate that the Total System Input power requirement for this server configuration is 423 watts. This is the predicted maximum amount of power that this particular server configuration would use under all environmental and load conditions. It is one of the numbers that an administrator could use when calculating the maximum power and cooling requirements to support this system in the data center. For a rack containing a group of eight of these servers, the total requirements would be 3384 watts. 17

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26

17
Using power capping in data center provisioning
Power capping is a tool to help IT organizations manage infrastructure size and costs in the data
center. By setting a power cap for a server or group of servers, a data center manager can be
assured that unexpected changes in workload or environment will not cause servers to consume more
than the specified amount of power.
Appropriate use of basic Power Capping in the data center lets administrators provision the cooling
infrastructure at an effective level
a level that meets the cooling requirements associated with the
server power consumption at a given application load and power cap rather than at the level needed
to handle all servers running at full power. Alternatively, using basic Power Capping to limit average
server power consumption will allow more servers to be safely operated within a pre-existing cooling
infrastructure.
Dynamic Power Capping takes this one step further. Its ability to control server power consumption in
real time allows administrators to use power capping in planning and managing their electrical
provisioning as well as their cooling infrastructure.
Choosing effective power caps
Setting an effective power cap involves determining the lowest value for the power cap that will still
meet the power requirements for a server or group of servers running their application workload.
Setting such a power cap should have no impact on server application performance but would reduce
the required cooling and electrical provisioning for the servers.
Administrators can use the HP Power Calculator Utility to estimate the maximum input power for a
given server configuration. This information provides a starting point for considering power caps since
it enables the data center administrator to determine the maximum power and cooling needed to
support a server or a group of servers. Additionally, the reporting features in IPM and iLO provide
historical power consumption information that can be used to select the most effective power caps to
achieve specific power savings or capacity planning targets.
Figure 10 shows the output from the HP Power Calculator utility for a ProLiant DL380 G5 server
configured with two Quad-Core Intel Xeon X5460 3.16-GHz processors, one 72-GB disk drive, and
8 GB of system memory. The results indicate that the Total System Input power requirement for this
server configuration is 423 watts. This is the predicted maximum amount of power that this particular
server configuration would use under all environmental and load conditions. It is one of the numbers
that an administrator could use when calculating the maximum power and cooling requirements to
support this system in the data center. For a rack containing a group of eight of these servers, the total
requirements would be 3384 watts.