Look, just like lots of other things in your life, airflow management in your data center is all about control. Unlike other things in your life, there are some simple things you can do to give you some level of control.

Let’s start at the beginning, then some tips.

Early Data Center Design – Room Level
Early data center design relied on control in the form of rack layout. Hot aisle / cold aisle (HACA) was an early approach that faces computer intakes towards the cool air aisle in the data center and directs the hot exhaust into a common hot aisle. It works for lower densities, but in today’s high-density computing data center environment, it’s just not enough.

Row-level
That’s where containment strategies come into play. Aisle containment physically separates and encloses either the hot or the cold aisle in a hot-aisle/cold-aisle arrangement. This technique requires keeping two rows of enclosures cooled to match the single hottest server in the range. While more effective than room level control, it can create a hazardous working environment for technicians, result in poor lighting conditions and pose a potential fire suppression challenge as well.

Enclosure-level
The next level of airflow containment is at the enclosure level. This approach uses an adjustable base-plate to let under-floor air enter directly into the enclosure. It directs the air properly to the server intakes by using diverters, air curtains and blanking plates to ensure proper air segregation in the enclosure. This enables data centers to mix high-density enclosures with lower density enclosures, providing optimal, granular control of the cool air resource. One major benefit is that it greatly reduces the server inlet temperature variations common in traditional data centers. And that allows data center managers to incrementally increase the temperature of the supply air – where every one degree (F) increase can net between 3 percent and 4 percent energy savings.

OK, for those of you who cheated, here are three steps that data center managers can take immediately to achieve the benefits of server-level containment:

Three Simple Steps You Can Take to Improve Cooling in Your Data Center Today
Clear your under-floor plenums. To maximize airflow in your data center and ensure that it is directed where it needs to go, first make sure your raised floors are cleared of any obstructions – especially where your CRACs or CRAHs are pushing out the supply air.

Use best practices. Plug enclosure rack mount openings with blanking plates. Make sure the enclosure air is properly segregated with air curtains. Consider skipping the wasteful use of perforated tiles (subject to bypass cooling) in favor of bringing the air directly into the enclosure base through an adjustable base-plate.

Contain your cabling. Ever notice how your server cabling seems to be everywhere? It is, and it’s also blocking cool airflow to your servers. Look for ways to manage them properly to ensure better airflow. Ideally, route them through a side-car attached to the enclosure.

Thoughts or suggestions?

Among all the coverage, talk and hype over cloud computing, we’ve been interested to see what the financial sector is doing. We work with a number of companies in this industry and, in our experience; they’re a pretty savvy group. Some large banks are testing cloud computing due to its agility and cost-efficiency. Other banks are waiting for answers to security and regulatory questions before they consider the cloud. And a handful of banks still aren’t quite certain what benefits cloud computing will offer them.

Financial services data centers tend to be high-density environments, so along with privacy and security concerns, we (of course) like to think about cooling and power consumption. It doesn’t matter where the data center is located – it still needs to be monitored and maintained.

Opinions of bank CIOs on whether they are comfortable housing their secure data in the cloud vary tremendously. At one extreme is Michael Harte, CIO of Commonwealth Bank of Australia. In a 2010 speech to the Committee for Economic Development in Australia, he said, “We will never buy another data center. We will never buy another rack or server or storage device or network device again.”

At the opposite—and more common—end of the spectrum are those bankers who are skeptical about using the services of an external cloud provider, according to a survey by Bank Systems & Technology.

We think the folks at Ovum Research have it right: there’s a middle ground where financial services can play safely: so-called “hybrid clouds.”

Ovum predicts that hybrid cloud computing will become the norm for large banks and that those with multiple legacy and mainframe-based applications will be the last to move to the cloud. Equally important, the banking industry is not rushing to adopt cloud computing because of regulatory and security concerns. They expect legal and technical clarity and a big move to the cloud by banks in the 2012-2013 timeframe.

We think that means some computing will reside behind lock and key on the financial services organizations’ premises in private clouds, while some computing – perhaps development and non-customer data-intensive applications – can move “to the (public) cloud.”

What do you think?

Stop us if you’ve heard this one: a data center manager puts his computing into the cloud…

Have you heard about Cloud Computing? Probably too much and not enough at the same time. We’ve been laughing about a great post decrying the overload of cloud computing coverage at Cloud Ave: “Cloud Computing is Dead.”

We’ve taken more seriously recent news about Google’s Gmail disruption and what it might reveal about becoming too cloud dependent.

On Wikipedia, the definition of cloud computing seems to say everything and nothing all at once. Plus, the definition contains a helpful note that “this article seems to contain a large number of buzzwords” which is always a warning sign:

“Cloud computing describes computation, software, data access, and storage services that do not require end-user knowledge of the physical location and configuration of the system that delivers the services… Cloud computing is a natural evolution of the widespread adoption of virtualization, service-oriented architecture, autonomic and utility computing. Details are abstracted from end-users, who no longer have need for expertise in, or control over, the technology infrastructure “in the cloud” that supports them.”

Just Like Your Data Center
We especially like the part that says “details are abstracted from end-users.” It sounds so fluffy and cloud-like that it makes you forget an important fact: the “cloud” isn’t a cloud at all. It’s comprised of physical space, just like your data center. It requires electricity and cooling, just like your data center. It houses high-density computing, just like your data center. And it provides many mission-critical applications and services, just like your data center.

But, unlike your data center, public cloud computing is off premises “somewhere.” Unless you’re a cloud services provider, you don’t get to see and touch your cloud computing infrastructure too frequently. But, it’s still a physical infrastructure. And it has all the requirements and challenges of a physical infrastructure.

Our point? If you’re considering cloud computing – and there are many great reasons to use this technology – then conduct the due diligence! Check to see that the facility is SAS70 compliant, find out where your data will be stored (and who will have access to it), ask about the cloud computing provider’s business continuity and disaster recovery (BCDR) strategy, and make sure you know the specifics of the physical cloud computing data center infrastructure.

We think this last one is very important: let us say it again:

“Make sure you know the specifics of the physical cloud computing data center infrastructure.”

If you are planning on placing your data and applications “into the cloud” you should take every step to ensure that the cloud infrastructure is as good as – or preferably better – than the data center you have built for yourself. That means power, cooling, rack and cable access, and a host of other concerns are all met.

What do you think?