This all comes down to exploring different potential policies for each type or category of servers in the data centre.
While exploring different possible power capping policies, the data centre planner can take advantage of understanding the potential positive or negative implications of their deployment by running individual what-if scenarios with their professional DCIM toolset.
This can be done for the various options and thereby test drive the potential outcome in a safe environment. These what-if scenarios can cover different areas of the energy and performance equation, like power, cooling, and floor space capacity. If done correctly, this could allow the data centre planner to free up capacity to enable the deployment of more servers into the same rack.
Step 3: Test possible strategies and determine approach
With potential policies determined, it's time to start testing each and every one to determine how applicable they are in the live data centre.
For mission critical installations, the first part of this testing might be best suited for a contained testing or staging environment outside of the production data centre. However, since the worst-case scenario for applying an improper policy is a decrease in application performance, which can be detected fairly quickly, it might be appropriate to apply the policies directly to the production environment. Starting with the least aggressive policy allows for verification before moving on to more aggressive policies.
During the testing period it is imperative to keep a close eye on the performance of each server with changed policies. Again, this is best done using professional toolsets which allows definition of alerting thresholds and according actions. With this capability at hand, the data centre planner can easily and safely deploy each policy to servers or groups of servers, while monitoring the results along the way.
Step 4: Tune the physical infrastructure environment
A last action, after all policies have been applied, is a tuning of the physical infrastructure. To understand what can be tuned and to what level, the use of smart DCIM tools is imperative.
With FNT Command, for example, the data centre planner is provided with a holistic view of the data centre's true power and cooling needs. Furthermore, is it possible to determine the operational capacities needed from the supporting physical infrastructure in order to support the continuing operation at minimum risk.
In a modern data centre the ability to engage and disengage various components of the physical infrastructure might be available, while older installations will be forced to tuning the runtime conditions of the equipment. However, no matter what the situation is, the decision-making support of the right DCIM tool will help the data centre planner to mitigate the risk of committing the various changes.
Sign up for Computerworld eNewsletters.