From Rack and Stack to Serverless

With the entire world enveloped in COVID-19, I just had to take a break and talk about something else. I have been spending part of my time during this lockdown understanding some of the newer technologies out there like serverless infrastructure. Before I jump into the new tech, I thought it might be a good idea to go down memory lane of how we got to serverless.

Phase 1: Enterprise Server Rooms

When I started my career 25 years ago, I was spent some time in server rooms where we used to rack and stack physical servers along with the Cisco networking gear. These were server rooms that a company would house all of their computer, networking and telecommunications equipment so they could function. This meant the company had to hire a team to manage the infrastructure along with the cooling, ventilation, electricity and other things that were not core to their company. This is what I call enterprise server rooms, of course some of these rooms were massive but they contained only one companies technology infrastructure.

Phase 2: Data Centers

Towards of the end of the 1990’s and during the dot.com bubble, the rise of data centers took place with companies like Exodus Communications. Do you remember Exodus? Oh, I do. I bought a bunch of Exodus stock thinking it was a great investment, only to see it go bankrupt. The idea of the data center was to move all of your companies servers to a company that specialized in managing servers, the telecommunication links, the cooling and ventilation. This meant an enterprise/company no longer had to deal with that headache. If the company wanted to add more servers they would email or call Exodus and within a couple days the servers would be available.

Phase 3: Cloud Providers

In the mid 2000’s, Amazon.com was growing very fast and their infrastructure team had to keep up and so they came up with innovative ways to provision new servers and data storage. They soon realized many companies were in the same boat so why not offer these solutions to the outside world. In 2005, Amazon Web Services (AWS) was born and in 2006 they launched a service called EC2 which was an acronym for Elastic Compute Cloud. The brainchild of this service was Chris Pinkham. It was a server you could rent by the hour and rent as many as you would like on demand. That was a sea change for the industry and really kicked started the cloud provider market.

Before we get to Phase 4, let’s recap the first 3 phases of this transition. In Phase 1, a company owned everything – servers plus infrastructure. Phase 2, they only owned the servers. Phase 3, they rented the servers.

Phase 4: Serverless

Around 2016, AWS launched another service that caught everyone’s attention – Lambda. Lambda allows you to take your computer code and have AWS run it for you. This means you don’t need to rent a server from them, then configure that server and finally put your code on that server. Instead, you hand them a .zip file with your code and they will run the code for you. This is also known as Function as a Service (FaaS).

There is another technology that is considered serverless and that is containers. I see large enterprises and large organizations go with containers because it streamlines their internal processes and gives the technology teams some familiarity of what they are used to. FaaS is a completely new way of thinking about technology architecture.

Summary

When I look at the shift we have experienced over the past 25 years, it almost looks like we have come full circle with the mainframe computer. Where you have a large computer (AWS), punch cards (FaaS) and dumb terminals (Web browsers).