Subscribe by RSS

Subscribe by Email

Your email:
Smartnet E-Paper

Browse by Tag

The XSi Data Center Blog:

Current Articles | RSS Feed RSS Feed

Data Centers – Where Does All the Energy Go?

  
  
  
  
  
data center energy reduction

Today’s on-demand society assumes nearly universal immediate access to real-time data and analytics in a resilient, secure environment. Anything short of that standard is unacceptable. These demands are being driven by a proliferation of data sources, mobile devices, radio frequency identification systems, unified communications, Web 2.0 services and technologies such as mashups. These rising expectations are creating demands of data centers that IT administrators are challenged to satisfy.  The impact of rising data center energy consumption and rising energy costs has elevated the importance of data center efficiency as a strategy to reduce costs, manage capacity and promote environmental responsibility.

To understand how to reduce energy consumption it is imperative to learn where and how energy is used in a data center

Data Center energy usageThe first key aspect to take a look at is how energy is distributed between IT equipment and supporting facilities.  Or in other words, which portion of the energy is being consumed by servers, storage and network equipment as opposed to power, cooling and lighting.  Research shows that in a typical non-optimized data center the IT equipment load consumes about 45% while the supporting facilities consumes 55%.  That means that 55% of the energy that is brought into a data center is not producing calculations, data storage, and so forth.

 

server hardware energy usage

 

The second key aspect is to understand how energy is distributed between the different components of the IT equipment.  Specifically,  within each specific piece of IT equipment what is consuming more, the processor, memory, disk, fan, power supply, etc.  In a typical server the processor uses only 30% of the energy.  The rest of the system consumes the remaining 70% of the energy!  Clearly, it’s imperative that efficient hardware design is considered when planning out a new piece of equipment for your data center.

 

 

server load energy usage

 

The final aspect to understand is how the energy in a data center is allocated to produce business results.  Often times, idle resources are powered on using energy without producing results.  Commonly, servers are underutilized yet they are still consuming the same amount of energy as if they were running at 100%.  Research shows that a typical server only utilizes 20% of its capacity.  This creates a big issue as a lot of energy is wasted on non-business purposes wasting a major investment. 

 

 

Power and cooling costs for data centers have skyrocketed by 800 percent since 1996, and the escalating costs see no end in sight, yet data center resources have low utilization (many below 20 percent).  Over the next five years, industry watchers predict that U.S. enterprise data centers will spend twice as much on energy costs as on hardware and twice as much as they currently do on server management and administrative costs. Moreover, many data centers are realizing that even if they are willing to pay for more power, capacity constraints on the grid mean that additional power is unavailable.

By understanding where all the energy is going in your data center you are on the road to significantly cutting consumption and reducing your overall data center costs.  Through analysis of your data center it is possible to reduce your equipment and floor space usage by up to 65% and cut your energy consumption in half. 

If you would like to find out more about data center consolidation you can download our Federal Data Center Consolidation and Data Silo Removal Case Study. 

Or to find out more about Data Center Relocation you can download our whitepaper “How to Carefully Plan & Execute a Data Center Move – Reduce Data Center Relocation Risks”

Comments

For proper management of Data Center Resources different capacities must be managed with utmost precision. The foremost issue is mismanagement of 'Energy'.The servers should be properly utilized i.e.the servers should be utilized according to the volume of job.In a nutshell-the power and cooling equipment should be fine tuned for optimization of physical infrastructure.Proper infrastructure can avert loss of data, overheating and processor overload.
Posted @ Sunday, June 03, 2012 7:42 PM by Chandrachur Bose
The problem can be simplified to the amount of "Load" to support and the distribution of "Power" required to support the Load. In the USA power, until recently was always available and relatively cost effective. Those two assumptions have changed. In some areas of this country it has been significant (just look at Nothern Virginia or specifically the Dulles Corridor). Bottom line the approach is a truly Systems Engineering approach that looks at the "Chip to Grid" and must look at available technologies that simplify infrastructure, reduce footprint and reduce loss. One of those available technologies is the implementation of power distribution. Specifically, replacing the current AC (alternating current) solution with a DC (direct current) solution. But this problem cannot be addressed in a piece part approach nor does have to be a highly precisioned approach. It just needs some basic Systems Engineering a more holisitc view of all the technologies that are needed to drive, support, manage, maintain and recover from issues the information that controls the business.
Posted @ Tuesday, June 19, 2012 6:17 PM by Andrew Clancy
We have been successfully modifying chillers for data centres in the UK. 
In some cases we have achieved improvements of 35%+ 
Our website features some case studies which operators may find interesting.
Posted @ Wednesday, June 27, 2012 2:54 AM by Lawrence Leask
Comments have been closed for this article.