This article is more than 1 year old

How do we train the next generation of data centre wranglers?

First, train the current generation

Define urgent?

Urgent support doesn't work in a hosted world, though. If you're a 500-person company using Office365 and the service starts behaving oddly, can you phone Microsoft and beg them to rush through a fix because you're trying to get a sale over the line? Of course not: if the service has a problem, all you can do is live with it and communicate with your users about workarounds and the provider's estimated fix time.

So the less you do in the data centre, the more you're relying on third parties' service level agreements. Which are less flexible than your own. This means you need to forewarn your users and communicate effectively to manage their expectations in the event of a problem.

New areas?

The stuff I just mentioned as “new” is as relevant now as it will be in the future. We're already working hideously inefficiently, and have been for years by writing shoddy code, by not doing analysis on our underlying databases, by failing to monitor network traffic.

We're already in the habit of running up a new virtual server on our on-premise virtualised estate because it costs us next to nothing to do and only uses CPU and memory when it's doing work that needs it. And we already have to manage user expectation, even if all our systems are on-premise and there's no hosted stuff in there.

So if you're not doing any of these things now, you need to address this now, not with the next generation.

First principles

And the way to address it is by proper education and training of your staff, and by ensuring that new hires are suitably qualified or at the very least suitably experienced.

If you know how Ethernet works, you have the basis to extrapolate how 100Base-T works, and how Gigabit works, and how 10GbE works, and how the next generation will work.

If you know how to analyse the complexity of an algorithm today, you'll know how to do the same for an algorithm in 15 years' time, because the theory still applies.

If you know how to design a relational database and optimise your queries now then you'll know how to in 10 years, because it's not changed radically in 20 years and it isn't going to change in the next 10.

There's a world of difference between knowing how to configure a system and understanding how it works. And a decent IT techie needs a grounding in both.

So …

… to train the next generation of data centre wranglers correctly, we must ensure they have a firm grounding in technology and how it works. We need to make sure they are able to communicate with their customers, because over time we'll have less and less ability to do quick fixes on essential systems.

We need them to be conscious of cost and able to monitor it closely, because it's getting way easier to run up a new entity that costs money than it was in the past. We need them to know how to use CPU cycles and memory effectively in order that they don't need to run up more and more costly virtual servers to run inefficient code.

Same as today's generation, in fact. Because all of the above applies right now – all that will happen over time is that the potential to do stuff wrong increases – the actual underlying technology won't be all that different. So let's start by sorting out the current generation – then they can pass it on to their successors.

More about

TIP US OFF

Send us news


Other stories you might like