Chances are, at one point in your life, you’ve been required to read Samuel Beckett’s magnum opus, Waiting for Godot. Often described as a “tragicomedy” by those in the hallowed halls of the humanities, the play focuses on two characters who spend their time on stage waiting for the eponymous Godot. The play has been the subject of a broad range of interpretations, ranging from the existential to the extremely popular, “I’ll never get those two hours back,” since the day it was published. For those within the less ethereal data center industry, the story can be viewed as a metaphor for the prolonged period of waiting for rack densities to reach the lofty levels that pundits have been forecasting for the past score of years.
For those of you still reeling from my decision to use a play — a play whose meaning has vexed high school and college students for years — as a metaphor for the state of rack density in the data center industry, let me just say that I always think it’s nice to class things up when you get the chance. Plus, if you ever took AP English, you might as well get some use out of it. As you may remember, many in the data center business have long predicted that rack density was poised for a growth spurt similar to those of seemingly harmless animals that have been exposed to radiation in 1950s sci-fi pictures. Unfortunately, for the purveyors of things like liquid cooling, so far no one has found the need to use a bazooka to take down a marauding 50kW rack.
Estimates of average rack densities vary, but there seems to be general agreement that in most instances, densities below 15kW continue to work just fine. Naturally, this begs the question: What is standing in the way of the mass scaling of the high-density rack threshold? I think it is pretty simple. People underestimate the refresh rate of IT.
You see, if you are in the advanced end of the IT bell curve, then you are likely refreshing your kit every few years. Amazingly, just like that razor thin, pane of glass television that hangs on your wall at home today and provides not just 4K Ultra HD picture, but can spy on your family and browse the internet — it still plugs into that same outlet that the 52-in. read projection TV that Uncle Saul gave you when he upgraded to plasma. That’s the joy of tech — most pundits and predictors have no clue for the rate of change in tech. Science relies up a control group with known conditions in order to perform a quality experiment. There are just some things in life that are really hard to create a control group for…black holes, anyone? — and innovation in tech is one of those elements.
Now, of course, some of you will point out that Blockchain and Bitcoin and rendering farms and the who’s-got-the-biggest-supercomputer wars negate my point. I say, whatever. You get my point.
As for as those of you who endured to the end of Beckett’s ode to nihilism may remember, Godot never showed up, and it appears those awaiting the fulfillment of the prediction of data centers filled with high density racks may find that their own waits become exercises in futility as well.