Sign up for our newsletter and get the latest HPC news and analysis.

CFD simulation shows datacenter when a best practice can be a bad idea

I know this isn’t a datacenter blog, but those of you with big iron have big datacenters, so I hope the rash of datacenter posts is at least somewhat interesting to you.

Anyway, Mark Fontecchio has an interesting article at SearchDataCenter this week, running through a case study for Lab 7D, a 7,000 square foot raised floor datacenter that Cisco uses for testing and QA.

But like many data centers, it also runs out of power. According to Chris Noland, who oversees the facility, the lab was the No. 2 consumer of electricity on the San Jose campus, generating $150,000 a month in power costs, or $1.8 million a year.

Once they figured out how much all those electrons were costing, they started trying to save money. They shut off redundant power supplies and saved about 10%, but that wasn’t enough

Cisco’s data center was already set up in a hot-aisle/cold-aisle configuration, complete with perforated tiles in the cold aisle and, in Cisco’s case, ceiling vents in the hot aisle. Looking to improve on this setup, Noland then talked to Pacific Gas & Electric, the main utility company in San Jose, and the Lawrence Berkeley National Laboratory about cold-aisle containment.

Before heading off into the deep end on containment, though, Lab 7D brought in a team to do CFD simulations to fully understand what was going on. The verdict: containment wouldn’t work for them, and would in fact make things worse.

Interesting case study on how the conventional wisdom isn’t always right for every situation.

Comments

  1. Thanks for the mention. It should definitely be noted that this lab is an atypical data center, mainly because it has quite a bit of Cisco equipment. As any data center manager out there can tell you, Cisco equipment often has odd in and out airflow patterns. Thanks,

    Mark

  2. The odd airflow configuration was only part of the story. The real problem is the volumetric air supply from the room which is 1/4 the airflow requirement of the installed IT equipment. This situation is preventing the segregation of supply and return air.

  3. John West says:

    Sherman – right…the real point of the article to me is that blindly applying any best practice or rule of thumb is never a good idea, although sometimes it does generate the expected result. Thus the focus of my post. I’m so interested in the remediation specifics, but I am very interested in the lesson.

  4. We see tremendous enthusiasm for containment projects. And, after a couple “blind enthusiasm” experiences we have learned to ask the obvious questions about IT load and cooling capacity.

Resource Links: