I heard a funny story about how camel dung came to be used in making seats for British war planes. This post on Snopes describes it, but the punch line is that the practice started way back in WWI when they used camels for transportation across the desert. The practice was carried forward without anyone ever questioning why it was needed on war plane seats.
I read another example of “the way it’s always been done”, where a beef roast recipe is handed down a few generations. The daughter asks the mother why the steak was folded into the pan in a certain way and the mother said that’s how she was taught. So the daughter goes to the grandmother and the grandmother explains that the only pan she had was always too small for the steak!
What’s my point? Last week I was editing a white paper on how to fix hot spots in the data center. One section talks about some ways people address hot spots. One of these ways is to place perforated tiles in the hot aisle. Of course we’re not proposing this as a correct method for fixing hot spots, but it may seem like a perfectly good solution to someone who managed a data center full of mainframes. The question we should ask is, “Why is it done this way?”
Sometimes the answer is, “it’s the way it’s always been done”, but not enough of us ask why. So I’m asking “why” to the following questions. If you have some WHY questions, please include them in your comments.
- Why do many data centers control cooling units off of the return air temperature, when the IT supply temperature is what matters?
I’ve heard this has to do with the fact that when data centers housed large main frames there was no hot / cold aisle configuration. Instead the air temperature was so well mixed that cooling off of the return air temperature worked just fine.
- Why do some data centers built today use signal reference grids?
This is a throwback to the days where IT equipment was highly susceptible to electrical ground noise. Modern IT equipment has improved significantly in terms of immunity and does not require a signal reference grid.
- Why are oversized neutrals specified in data centers with harmonic power factors near 1?
Before power factor correction mandates were placed on IT power supplies, over sized neutrals were required to prevent excessive heating due to the harmonic currents. Today it’s nearly impossible to buy modern IT gear without a power factor-corrected power supply which translates into copper savings on smaller neutral wires.
- Why do we need withdrawable medium voltage circuit breakers long after the days of oil-filled breakers?
In the early to mid 1900’s medium voltage breaker contacts were immersed in oil in order to break the arc and cool the contacts. As you can imagine, these breakers required frequent maintenance, hence the benefit of a withdrawable chassis. Today most medium voltage breakers use vacuum bottles to interrupt the arc and don’t require anywhere close to the maintenance interventions of oil-filled breakers. Given the significant cost savings, fixed circuit breaker switch gear should be considered for data center applications.
I’m sure there are many more of these questions, but my hope is that this blog will at least get us to stop and ask ourselves, why a procedure is written the way it is, or why you need to specify a raised floor for your new data center, or why you need to route your cables under the raised floor, or… If you have any other examples, please share them with the community in the comments field below.