I have friends who are constantly trying something new. Whether it’s a new restaurant or some activity they haven’t done, they are constantly on the prowl for change. A few months back, I received a Groupon email from them that said, “Skydiving, $89!” When I first read it I thought, “I can do that. I’m under 50, I have a zest for life, and I can definitely afford it. Why not?” After thinking about it a while, I asked myself, “But should I do it?”

Over the 20 years that I have been designing data centers (yes, I was designing data centers before data centers were cool), I have noticed several decisions that should not have been made. Each and every time I walk into a data center and see a “bad decision,” I shudder. Most of the time the decisions are based upon cost, but every now and then I see a decision made that is more costly and less effective. Here are the Top Five “Just because I can, doesn’t mean I should.”

Just because you can, doesn’t mean you should… put generators outdoors. What came first, the generator or the generator housing? Generator housing was created in response to many companies’ lack of room inside their building for a generator. The fact of the matter is the most likely time a generator is needed is during a snow storm, ice storm, or bad weather in general. In the event of an outage and the generator doesn’t start, immediate access to the generator without going out into the elements of the storm is essential. Also, generator housing (2,500 kW) can cost upwards of $100,000 to $200,000 per housing. With a five generator installation, the housing runs between $500K to $1M in total installation cost. Constructing a building (2,000 sq ft per generator) of 10,000 sq ft at $100/sq ft, brings the total cost to $1M to house the building. In addition, the housing extends the lead times by two to four weeks.

Just because you can doesn’t mean you should… build data centers without raised floor. From 1997 to 2001, during the .com era, data centers were often built using overhead duct work distribution instead of raised access floor (RAF). Today there is always an argument that RAF is not necessary. Yet, prior to this era, all data centers were built with them for several reasons. RAF is far more flexible when reconfiguring rack layouts than no raised floor. The duct work above is not flexible when considering changing your environment. In addition, only 174 W/sq ft can be effectively cooled via overhead distribution. Data centers with RAF can cool a much higher volume (275 W/sq ft). The only reason data centers were built without raised floor was that RAF lead times were 45 weeks! During this era, speed-to-market was a major factor in data center construction, and companies could not afford the lead times of 45 weeks. Finally, cost analysis has shown that using RAF is actually less expensive than overhead duct work.

Just because you can, doesn’t mean you should… design remote radiators for generators as a primary design. Remote radiators should only be specified as a last resort. Similar to generator housing, remote radiators are designed only when intake air is not achievable. Data center design practices frown upon exterior equipment installations such as remote radiators. The equipment is exposed to the outdoors and the cost is an approximate $100,000 add-on to the generator purchase.

Just because you can, doesn’t mean you should… consider containers as a primary source of housing a data center. Containers are nothing new to the industry. The first container I designed was for Comdisco’s Comroc data centers. These were containers housing UPS and cooling supporting computer equipment operations. They were used as a disaster recovery solution (many of you more seasoned data center managers remember this solution). While containers today have a purpose in the market place, they should not be considered as a primary solution. They only work well as “swing space” or expansion until a permanent solution is constructed. In addition, cost per rack is often more costly than building out a new data center.

Just because you can, doesn’t mean you should… use equipment vendors to design a solution. Engineers within the industry design with an objective point of view based upon performance. Equipment vendors design based upon the sale.

Now, let me go back to my original story. My friends did go skydiving that day, but not all of them ended up jumping. The jump was delayed due to the team before them. A pair (instructor and first time jumper) went to land and got side swept by the wind. They slammed into the ground, breaking the instructor’s leg. Only half of my friends got to jump that day.

Last week, I received another email from the same group of friends that said, “Bungee Jumping in Starve Rock!”

“I can do that.”