The website is down. The cloud is up.

It has been well over a decade since I embarked on my journey with the cloud. I spent the first five years of my career as a customer of Amazon Web Services (AWS). Now I’ve spent seven years as an employee working on the cloud at AWS.

Whenever I think back on my career, the first and most immediate thing I notice is how much the culture of the people who keep the internet online has changed over time. For someone starting out with cloud computing today there just isn’t the same lived experience of what things were like before the cloud. This is a good thing. The culture we have now is arguably far better than the culture we had back then.

But let’s look back at a snapshot of how things were before the cloud:

The year is 2008. Amazon EC2 launched two years ago in 2006. The vast majority of websites are still served by computers that live inside corporate offices and collocating facilities. The tech industry is slowly recovering from the Dot-com bubble burst. IT professionals are a niche but growing community with a unique internet first culture. Josh Weinberg launches a video series called “The Website is Down”.

Back then getting a couple tens of thousands of views on YouTube was an incredible feat. “The Website is Down” acheived that and more. It was a viral hit among IT professionals of the time. We saw ourselves or our coworkers reflected in the characters of the series, and we laughed and shared it with each other until nearly everyone in the IT community of the day had seen the series.

“The Website is Down” is now mostly forgotten. After 14 years on YouTube, it has not even managed to accumulate two million views. Meanwhile Mr Beast videos get millions of views within the first 15 minutes. But “The Website is Down” remains as a snapshot of early 2000’s IT history.

Content Warning: The following YouTube video contains crass language, immature humor, and depictions of behavior that would not be appropriate in a modern work environment.

The series features Derrick, a brilliant but lazy IT professional. Derrick plays video games during work hours, and attempts to avoid work as much as possible. He blatantly lies to coworkers, and tries to solve problems in the easiest way possible. In episode #1 of the series, Derrick’s laziness catches up to him when he reboots a webserver at the wrong time, taking down the website.

What follows is an IT action sequence in which Derrick frantically juggles multiple Skype calls, while using his computer skills and elevated system access levels to successfully gaslight his manager and coworkers in order to avoid all personal responsibility for the downtime. Of course, more mistakes follow, leading to further outages, and complete havoc. Somehow this depiction of irresponsible behavior becomes hilarious, in part due to the characters’ chipmunk voices, which were a common side effect of speeding up the video to fit within YouTube’s video size limits of the time.

This was the state of IT in the early 2000’s, before widespread adoption of cloud computing. The Derricks of the world were not good at uptime. The websites that they managed were down frequently. But website downtime is much more rare now. In fact, as of a couple months ago Google officially retired the “Cache” link that used to appear on the search engine results page. The stated reason:

Yes, it’s been removed. I know, it’s sad. I’m sad too. It’s one of our oldest features. But it was meant for helping people access pages when way back, you often couldn’t depend on a page loading. These days, things have greatly improved. So, it was decided to retire it. @searchliaison

In recent times, some in the tech community call for a return to on-premise or self managed computing. Typically this is justified by a naive cost to cost comparison of running your own hardware compared to running similar hardware in the cloud. A deeper level of analysis may factor in the cost of the supporting infrastructure and the cost of the humans that will maintain the hardware and ensure that the software continues running on that hardware, so that your website stays online.

But what these comparisons have missed is the main factor that drove the tech industry trend of moving from on-premise to cloud. The biggest factor by far was the desire to externalize responsibility and outsource professionalism. In short, the cloud succeeded because companies wanted to fire Derrick.

There are a lot of Derricks out there. It is easy to hire a Derrick. When you hire a Derrick it is hard to prevent that Derrick from causing havoc. It takes a lot of structure, access control, and bureaucracy to keep your company free of Derricks and to keep any Derricks that do slip in from causing havoc. Unfortunately, that level of control and bureaucracy tends to be antithetical to fast, agile development. Startup companies need to build fast, but also build well, with minimal downtime.

Modern startups can pay a cloud company to help keep the Derricks of the world from taking down the website. The cloud company takes responsibility for how cloud services contribute to overall uptime. For example, AWS services come with a Service Level Agreement (SLA). The SLA defines a monthly uptime percentage, and if an AWS service fails to meet that uptime percentage, then you get a service credit. This gives AWS a tremendous financial motivation to keep AWS cloud services online and to have high uptime on all services. AWS filters out the Derricks, and AWS builds the structure that allows continuous delivery of cloud services that are resilient to the havoc that a Derrick might try to create.

Everyone building a business wants to hire brilliant people who will do great things if given a broad level of access and responsibility. Maybe they can be lucky and avoid all the Derricks. Maybe they’ll get unlucky and a Derrick will be the one managing the website. But if they are building with cloud services then the website will have a lot less exposure to the mistakes that a Derrick can make.

We have come a long way in the last couple decades of the internet. But by far, the biggest change that has happened is that we went from a world where “the website is down” to a world where the cloud is (almost) always up.