Flying in the clouds

When I was travelling back from holiday last week I had one moment of distinct unease. I realised that the sealed metal container I was travelling in was speeding through the air at many hundreds of miles per hour at a height of many tens of thousands of feet. To continue breathing and for my heart to carry on beating I was reliant on lots of very sophisticated systems and the skills of a couple of pilots to get me “safely” to the ground. My position seemed extremely tenuous; the situation felt very risky.

But, of course it isn’t. Commercial air transport is the safest way to travel by far and in comparison, the risks associated with driving my car to the airport are off the scale. If you take into account all the daily risks we are subject to, like road traffic accidents, falling down the stairs, food poisoning etc, then actually, being seated on an in-flight aircraft is probably not only the safest means of transport but the safest place to be in any circumstance. In comparison, being on the ground is distinctly dodgy!

If you think about it, it has to be so to make the air transport industry viable. You (or I) wouldn’t set foot on an aircraft if there was any real danger. Risk has to be taken out almost completely to make the whole mass airline market work.

So how do they do it? The answer of course is a massive attention to engineering detail and absolute adherence to process. Everything from the rivets that hold the plane together to the armrests in the cabin are engineered and tested to a degree that is mind-blowing. Everything critical is duplicated – a two engine jet can fly, take-off and land with just one and they are designed to not damage the rest of the aircraft in the event of a fire and so on and so forth. This all comes with a price of course; airliners cost hundreds of millions to buy and similarly eye-wateringly large sums of money to run.

Therefore, you would have thought that air transport would be prohibitively expensive but of course it’s not. Many years ago flying was the preserve of the rich and famous, it was an exclusive and luxury experience. But this has all been changed by the charter and so called “low cost” airlines. They cut costs but don’t cut corners on anything that might impact safety, rather they exploit the simple economies of scale. They make sure each flight is full up with lots of people and they turn them around in no time flat so their expensive plains are in the air as much as possible. This way, the individual cost of a seat is affordable, even for the likes of me!

Then, once I had convinced myself that I really was safer than houses and with nothing else productive to do on the journey, I started to think whether the IT industry could learn anything from this. In my experience, IT is notoriously unreliable and perceived to be expensive to use (in a corporate environment).  Could we run our industry to provide the sweet spot of reliability and low cost enjoyed by the airlines?

The obvious lesson I would draw out is that the way to reduce costs ought not to be to compromise on engineering and service but rather employ economies of scale.  You can already achieve these economies using cloud computing but you can’t expect the same costs by doing something different on a small scale in your in-house data centre.

Although air passengers get their fundamental requirements met – that is get safely from A to B for a reasonable price, they do have to make some sacrifices. Air travel is just, but only just acceptable in terms of comfort and convenience. Planes are noisy and cramped and I wouldn’t describe my recent flight to be a high point of my holiday!

Users who subscribe to a large-scale cloud service are likely to enjoy high levels of reliability at low cost. However as soon as they want something different from everyone else then they will inevitably incur higher costs and also may have to accept a poorer service.  This is a mistake that I think some organisations make; they take the mass market cloud service as the price point and then layer-on bespoke requirements and expect the same price. If you do this and insist on a custom solution, what you are likely to compromise is standards of service and engineering.

If the uptime of your IT was a matter of life or death, then you would accept sacrifices in other areas to make the price point. IT uptime might well be life or death for your business, so don’t compromise on the fundamentals but focus on what you really need.