Cisco Logo


Inside Cisco IT

What are typical power and cooling densities in a Data Center these days?  What do you think they’ll be in the future?

I routinely host Data Center tours at Cisco’s headquarters.  During those walkthroughs I often discuss how the densities in our server environments have increased over time.  Facilities with 50 to 60 watts per square foot gave way to those of 100 to 120 watts per square foot, which in turn gave way to those of 175 to 200 watts per square foot.

“I’m your density.  I mean… your destiny.”

- George McFly, Back to the Future

That’s where Cisco’s demand has settled in, with no obvious need for greater densities thanks to the efficiencies of virtualization technology and our Unified Computing System technology.  I recently chatted with a Data Center manager who runs a huge co-location facility.  He mentioned several tenants needing 400 watts per square foot and one pending hardware deployment anticipated to require 1,000 watts per square foot.

As Back to the Future’s Dr. Emmett Brown is fond of saying, “Great Scott!”  That’s a lot of power.  Perhaps not the 1.21 jigawatts needed for the good doctor’s time-traveling DeLorean, but a lot just the same.  Doc Brown and Marty McFly aren’t due to arrive from 1985 for a handful of years yet.  What sort of power and cooling densities do you think the world’s latest and greatest Data Centers will be offering by then?

Grab your Huey Lewis and the News cassettes and your hoverboard and watch the video below for discussion of future Data Center trends.

In an effort to keep conversations fresh, Cisco Blogs closes comments after 90 days. Please visit the Cisco Blogs hub page for the latest content.

2 Comments.


  1. Hi Douglas,

    When you design our data centers and visit customer data centers, how much emphasis is being placed on having from to rear airflow designs on the “box level”? Do you ever see our smaller routers such as the Cisco 2800, 3800 and ASR100x series being deployed in the data center?

    I am wondering if we should be designing front to rear airflow for our next generation products.

    Thanks,
    Brent

       0 likes

    • Douglas Alger
      Douglas Alger

      Hi, Brent -

      Hot- and cold-aisle designs are standard in Data Centers these days, so hardware models that have front-to-back airflow – and therefore align with hot and cold aisles – are certainly preferable. For side-venting machines, Data Center managers have the choice of living with the less-efficient airflow or else deploying the machines into server cabinets that can redirect air into a front-to-back pattern.

      I encourage anyone designing Data Center hardware to aim for front-to-back airflow. That’s not always feasible – I know some models have a different configuration to permit greater port density, for instance – but it’s certainly desirable. Also, continue to focus on energy efficiency – more computing capability relative to the power consumed.

      Thanks,
      - Doug

         0 likes

  1. Return to Countries/Regions
  2. Return to Home
  1. All Inside Cisco IT
  2. Return to Home