IT Essentials: A warning, not a catastrophe – this time

Europe’s internet has a single point of failure. It’s called America

A fault at a single AWS datacentre took down a large swathe of the global internet this week. We can’t afford to treat that warning as a temporary inconvenience.

AWS was able to recover quickly, but essential services like banking were still down for longer than was comfortable.

Service owners suffered, and not always because they were at fault; even if they built their infrastructure without AWS, their own supply chain almost certainly contains dependencies. That’s what caused the cascade of errors that took down platforms like HMRC, Snapchat and Halifax on Monday.

The problem was isolated to one region (US-East-1), AWS' largest and oldest. But that datacentre is a routing point for a huge amount of global internet traffic; the epitome of a single point of failure.

The aftermath should be a wake-up call to examine your own infrastructure. Find those failure points, build in resilience and redundancies, and make sure there’s a recovery plan in place. Even failing over to a different AWS region would have made the difference in this case.

On a wider scale, this incident only boosts the argument for a sovereign European or UK cloud provider.

We’ve spoken about data sovereignty at length this year as trust in US institutions falls, but it's time to go beyond that: we need infrastructure sovereignty, too. Even existing European cloud providers like SAP run their infrastructure on American tin.

This is much easier said than done. Any European provider will be playing a long and expensive game of catch-up to compete against the US hyperscalers, and their offerings will invariably be more expensive and develop more slowly. Is that a price worth paying?

I would argue yes, depending on your risk profile. The internet is simply too important to the modern world to rely on a single point of failure at a single supplier. We need more suppliers, offering true independence from the current American hegemony.

Even Nick Clegg (whom I disparaged last week) thinks that’s a good idea. He told The Times this week, “We are wholly dependent on every level of the stack for technology from a country where the geostrategic interests are no longer aligned in the same way they have been for the last 30 years.”

The solution ultimately has to be a political and technical one. For all the talk about small government and independence, Microsoft, Amazon and Google didn’t get where they are today without hefty support from Washington (even if some of it was unintentional). It simply isn’t possible for any company to bootstrap themselves to hyperscale without help.

The AWS disruption was a warning, not (this time) a catastrophe. The question now is whether policymakers and businesses treat it as a fleeting inconvenience or the catalyst for building a more resilient internet for the decades ahead.

Penny Horwood has collated industry reaction to the AWS outage – and I’m far from the only market watcher to raise concerns about sovereignty.

On the diversity side, Penny attended the closing of Women in Tech Week at the House of Commons, where experts called for structural change to widen the pipeline of disabled tech talent.

In our latest Ctrl Alt Lead podcast episode, Queen Mary University CIO Rachel Bence explained how sustainable computing, cloud and AI are reshaping higher education.

And John Leonard talked to Satpal Chana, deputy director of data analytics at VisitBritain, about how AI is changing how the UK presents itself to the world.