It’s a beautiful day in the year 2047. Your drone-delivered breakfast was on your doorstep on time. There was no turbulence as your self-flying car navigated the city. You even got a ton of work done on the way to the office, thanks to the conferencing system with built-in video, audio, and data sharing. Somewhere on the way into the office, you flew over a number of self-driving taxis in the downtown area and were glad you weren’t stuck in ground traffic today. You even had time to scroll through the latest space travel catalog and started thinking of saving up to take that trip to Mars. The moon trips seem to have become a bit commercialized for your taste. Yes, it’s just another day in the near future.
But now you’ve gotten to the office and, suddenly, you’re in total darkness. You peek through the window and you see that everything around you has crashed into a simliar darkness. The power to your building, your flying car, its data systems, the taxis, traffic controls, service distributions, and this entire future world you’ve become accustomed to is completely stalled out. Even worse, people are seriously injured by self-flying cars falling from the sky and autonomous taxis failing to stop and communicate with each other.
A Cloud Armageddon?
What happened? Turns out, you’re living in a repeat of the 2017 AWS outage that happened just over 30 years ago. Back then, a critical lack of redundancy and widespread dependence on a single cloud provider caused headaches for hundreds of thousands of websites across the United States. That’s when major sites like these went down (as well as tens of thousands of other smaller sites):
Fortunately for all of us, no one dies if Netflix is down (even if your teenager thinks so). The outage was so far reaching that even Amazon itself wasn’t able to update the AWS service health dashboard because it was hosted on, you guessed it… AWS. This is due to a phrase many of us know as “putting all your eggs in one basket,” which is never a great choice. Plenty of business was lost over the AWS S3 outage. At the end of the story, it turns out that the root cause was a human error. It was a billing update gone awry. Being able to access videos, images, social posts, and productivity tools—services that people have come to expect to be available in this day and age—caused plenty of frustration, but thankfully again, it didn’t physically harm anyone.
Back to the Future
Fast forward again to 2047, where we’re all in these flying machines, autonomous cars, and where there are integrated systems in every aspect of our lives, including healthcare, maybe even bionics. What happens in that world when just about everyone and everything might be stacked into an AWS-heavy future? The impact of this one-cloud dependency is much, much worse. What should have been learned the first time around? A world that depends on one monolithic cloud service provider, and has put its critical infrastructure (and yes, lives) into an infinitely stacked monster cloud, has to deal with the reality that one misstep will mean big consequences. If a similar scale outage were to occur in this future state, human lives would be affected significantly, the economic impact would be massive, and the impact on civilization could be much more than a loss of entertainment and work tools.
Redundancy in a Better Multiverse
This, among many other reasons, is why the AWS cloud outage has many people thinking about a multi-cloud hybrid strategy. The value of a hybrid multi-cloud, mixing multi-regional with multi-level computing resources, is incredibly compelling. It’s the key to a reliable, but flexible future. Businesses today shouldn’t be pushed to “go the the cloud” (only to then maybe later free fall from it), but should be focused on integrating combinations of computing resources that incorporate the services of multiple clouds. Organizations should strategically look into a hybrid cloud strategy that includes service levels, support, and price points that cover a wide range, all in the primary pursuit of redundancy, but also efficiency and long-term scalability.
Spanning multiple data centers, and carrying data and services across an enterprise and across the world, this multi-cloud, hybrid-based approach to cloud computing is the future—and the time to build it is now, before human lives are impacted. Imagine a scenario where internet of things (IoT) devices relying on the cloud are keeping all of our mission-critical healthcare devices functional? While a strategy that integrates AWS in a multi-cloud configuration is possible, it isn’t easy. That’s because building across data centers and cloud services requires single-point visibility, so that automation, reporting, and deployment are universal. When you lay out all of these considerations, it’s clear that a commoditized, entry-level public cloud product has its place in the big picture. The world will always use hyperscale clouds like AWS, but in the future, their business and enterprise picture gets smaller as we realize the risk of a non-diversified cloud. To AWS we say, thanks for the wake up call.