Amazon – Too Big to Fail?

The following is a guest contributed post by Andreas Hassellöf, founder of Ombori

On February 28th, we were in the process of launching Ombori’s new app conversion tool, Bruce. Everything was going perfectly, and we were briefing our clients on the results. Then, without warning, it stopped working. Our clients were, understandably, concerned. As you can imagine, it was an awkward moment.

We weren’t alone.

Over 148,000 other sites and services who depend on Amazon Web Services (AWS) were also experiencing something similar.  As we learned later, an Amazon engineer had made a simple typo when taking a server offline. That tiny mistake took several other servers offline, which then took down two critical S3 subsystems. One of those subsystems manages the location of all S3 objects, and businesses across the world lost vital services. It wasn’t just small businesses like ours who were hit: massive companies including Spotify, Pinterest, Netflix and Buzzfeed were all affected.

Fortunately, there was little or no lasting damage. The service was restored in a few hours, and it doesn’t appear that any data was irretrievably lost. A brief outage may not seem like a big deal. But consider the hidden costs to Amazon’s clients. Hundreds of thousands of man-hours were wasted while IT staff tried to figure out what was wrong and create work-arounds, diverting attention from other projects. Millions of dollars of revenue were lost when customers couldn’t access sites belonging to AWS-dependent businesses and went elsewhere.

Single point of failure

Amazon has become an alarming single point of failure. A significant part of the world’s economy now depends on Amazon for their sales and distribution, for their customer communication, and for their day to day operations. They effectively own the cloud business. With a market share of 42%, they have almost three times more than their closest rival Microsoft, and six times more than Google. They provide the back end for hundreds of thousands of businesses worldwide. In addition, Amazon’s retail portal is the main sales channel for hundreds of thousands of small and medium businesses around the world, as well as millions of third party Marketplace sellers. When they go down, even for a few hours, the impact is instantaneous and massive.

As a recent report from the Institute for Local Self-Reliance noted in November last year, “Amazon is far more than a big, aggressive retailer… Amazon increasingly controls the underlying infrastructure of the economy. By controlling this critical infrastructure, Amazon both competes with other companies and sets the terms by which these same rivals can reach the market. Amazon is using its market power to eliminate competition and take control of one industry after another, leaving us with an economy that is less diverse and innovative, and which affords fewer opportunities for businesses to start and grow.”

Amazon are unquestionably a great company. They’re hugely innovative, and they have a great focus on customer experience, but it’s dangerous to keep relying on them to underpin every aspect of our businesses.

February 28th was only a brief outage: imagine if AWS were to go down for a day or two. We simply cannot afford for Amazon to fail.

What’s your Plan B?

If you’re one of the millions of businesses who are reliant on Amazon, ask yourself this. Do you have a backup plan, or are you completely dependent on them? If they go down again – or more accurately, when – what would be the cost to your business?

Now is the time to investigate alternative services, or perhaps consider developing your own technology. The more Amazon continues to grow, the greater the risks we all face.