Back in 2011, Marc Andreessen quipped, “Software is eating the world”, and I’ve since found this phrase coming up again and again. An industry friend of mine had used it a while ago in a conversation with me when questioning whether my shareholders really understood the value of what my software business was capable of.
The wonderful thing about software is its near perfect scalability. When I was educated as an electrical and electronic engineer it was during the transition from analogue to digital electronics so I learnt how to bring electronics alive with software.
In the early part of my professional career I got into control electronics. It was a period I enjoyed because I was in industrial process automation and robotics; to me the perfect meeting place between hardware and software.
I was reminded of Marc Andreessen’s phrase again more recently when attending a data center conference in London. I was presenting some interesting findings from an analysis I’d done comparing the data center industry ten years ago compared to today.
For those of you that don’t follow the data center industry, it has been transforming itself over the past 6-7 years, and while there is still a long way to go, the industry has come a long way forward in terms of minimising waste and being more efficient.
Anyway, it was during a conversation with another industry colleague about the presentation I’d just given, I realised just how powerful software was and indeed it was ‘eating the world’.
We were discussing and debating what type of data center cooling design would be the best (most energy efficient) for a given particular climate region in the US. The two types we were discussing are called Direct and Indirect air side cooling.
Like many situations where there is more than one obvious choice for something, understanding and analysing, which if any given option, is best or which would best meet the desired outcomes is not a simple or straightforward task.
In this case, to answer the question empirically you’d need some mechanical engineering and mathematic skills. There are plenty of these skills in the data center industry, yet despite this, there continues to be discussion and debate because there are many other variables at play in deciding the outcome.
So we had a problem that was complex, contained main variables but was finite and ultimately deterministic. In other words, an ideal problem for software to get its tiny teeth into!
Luckily for me my business is in the game of software and specifically software that allows me to feed such a question into it, ask it to resolve the problem and give me results.
I went back to the office that day and asked one of the guys in the development team to test which of three different cooling designs (chilled water, direct and indirect air) would give me the optimum efficiency and optimum operating cost. I knew having just presented a tiny subset of data that those two things (optimum cost vs optimum efficiency) may not be achievable in the same cooling design when placed within a particular cost and climate region.
He ran all three situations through our software across all the US cities we had energy cost and climate data for, which amounted to some 208 different combinational analyses that needed to be run. Doing this by hand is a task that would be so time consuming (and costly) that nobody in his or her right mind would do it, which is why I guess that nobody did it before!
Four hours later I received an email with a spreadsheet of results attached to it. I have to say I was rather excited when I opened it and started to look at the data. It wasn’t long before I had our CTO sat next to me and we were graphing the results together and looking at what the data revealed.
Now if I had to estimate how long that would have taken to do without our software I’d put the figure at something close to 2,496 hours or 312 days assuming an eight hour day (I don’t know many humans that can work 24/7 continuously until the job is done, do you?!).
This got me thinking further. What if I just fed into my software all the questions that every person had ever asked every design engineer in the world? Wow! could you imagine it! The point would not be to eliminate the human but to make him super-human!
Isn’t that ultimately what software does for us, make us super humans? Any amount of data available at our fingertips. Think about just how well informed you are these days just by asking your phone a question and getting the answer spoken back to you.
Yet software isn’t perfect; it’s imperfect while designed and programmed by humans, yet even with imperfections it can completely change your world, whether that’s knowing exactly what turn to take on the way to your destination or which cooling design to use in which climate for your data center.
For those data center geeks out there that are actually interested in the results of the analysis I described in this blog, you can see the results here.
I hope whatever industry you are in, you are making use of the software that’s eating your world because if you’re not at the table, then you’re on the menu and will yourself be eaten!
Finally, for those of you who think this story is just about data centers, I’d urge you to think again!