The Kubernetes history- the project which almost never happened
"It's 2013. The environment in silicon vally was on the heels of the devops movement, automation tools were all the rage"
- Kelsey Hightower, Principal Engineer, Google.
The speakers in the 2022 Kubernetes documentary by Honeypot includes a fantastic list of speakers reflecting on the early days of Kubernetes and the wider ecosystem changes at the time of its inception.
What follows are interesting quotes from the speakers in the Kubernetes documentary, highliting the historical significance of the decisions made and inception of the project.
"If we didn't have docker we couldn't have done it- and it lit this (great) fire, but we had all this experience about where it should go..."
- Brendan Burns, What is Kubernetes? (origional Kubernetes co-author), Corporate Vice President, Microsoft, Prev Google
See also:
- The original Borg paper : Large-scale cluster management at Google with Borg
- Understanding Kubernetes
The documentary reminded me of Andrew Clay Shafer's obvervation at his 2013 Velocity conference talk: "There is no talent shortage" in 2013 where he re-states:
"You can either easily manage complex systems at scale... or you cant"
- Andrew Clay Shafer, in which he was referning to his 2008 slides.
"Google was looking for ways to apply it's internal infrastucute expertise (Borg) to the cloud ... and leverage that expertise and acelelrate our move into the cloudspace.
- Brian Grant, Distinguished engineer, Google
The dotscale conference was a signifant event- "It was a big deal, the concept and the implementation behind it.. will affact how we shift software in the future".
"What the hell is a container, why should I care?" where common questions at the time.
- Joe Beda - Principle Engineer, VMware Prev Google
"Docker developed an eco system which was really streamlined"
- Tim Hockin, Principal SW Engineer, Google
What is Kubernetes again?
Let's imaginge: "Let's imagine a PostOffice- it doesn't want individual tiny things- you have to package it. The PostOffice abstracts all that away from you- the stamp, the address- planes can break down, but nobody at the Post Office calls you when these things happen- how they do it is not a concerns. Kubernetes in based on this thing called 'promise theory' - just like the post office's job is to make sure your application keeps running- if you can afford the postage, we'll run it for them".
- Kelsey Hightower, Principal Engineer, Google.
"Google was answering that question- we have this machine, we have borg, what's next? We knew there was a better way to offer developer platforms. The idea was can we build a borg like system in a way that applies to cloud. Do we use borg itself, or play with something out there already like mesos? Building an open source community was the best way to build a defacto standard".
- Joe Beda, Principle Engineer, VMware, Prev Google
"There was an inital prototype; Joe, me and Brenden ; the moutain view team. We needed to make sure we wern't enabling the competition.".
- Brian Grant, Distinguished engineer, Google
- Brendan Burns, Corporate Vice President , Microsoft
How did Kubernetes nearly not happen?
The initial response from the (Google) exec team was; I don't think we neeed to do thing."
- Tim Hockin, Principal SW Engineer, Google
"It was stessful because it was obvious to us what needed to be done".
- Joe Beda - Principle Engineer, VMware Prev Google
"Google had real concerns about funding our competitors through doing this".
- Tim Hockin, Principal SW Engineer, Google
"Eric brewer and I were discussing on a bus- the target position was : No. this was too precious not to open source. Google is not big enough to do that by itself. To pull that off you need a lot of fellow travellers- provided you can actually get people to participate".
- Craig McLuckie, vice president R&D
"We needed approval pretty quickly to launch at docker con"
- Brian Grant, Distinguished engineer, Google
"With the VP team, one VP was very much against us open sourcing it- but there was a turning point in the meeting that it was clear we were going to do it. We were surpised, we were able to get it out there, and then thought ok great now what do we do since we've got the go ahead?"
- Joe Beda - Principle Engineer, VMware Prev Google
"(since) It took months to get approval, and we wanted to present at dockercon- we may have started before we got formal approavl"
- Craig McLuckie, vice president R&D
"Then it was ready- we did a rewrite into go, created a declatitive companent ontop of docker written by Tim".
- Joe Beda - Principle Engineer, VMware Prev Google
Where did that name & logo come from?
Driving in a car, the steering weel- also it passed the Google test (not many search results for the name).
The logo? Since the internal design team wasn't very fussed about a small open source project, it was left to the team to make their own.
Early on Tim created the first logo- he actually has a degree in graphic design"
- Joe Beda - Principle Engineer, VMware Prev Google
The missing paragraph
The focus is on delivering systems which enable developer velocity to production in a managed way. Forget Kubernetes- the focus is on tooling. Don't get consumed by the buzzword of today- todays hype is tomorrows legacy. But do watch out, if you hold your head in the sand and ignore the better tools available to your organisatoin, project or business- you'll be too late to find out other individuals are able to produce comparible deliverables to what your entire organisation is only delivering yearly with large teams.
See also: It's not about scalability it's about developer experience which makes experimentation fast.