Hi you are getting into details out of the scope of this article. I want to address your points but in an article, then we can link it on HN and discuss it there.
The CNCF is an entirely different beast... which I have already started writing about already. It is the Mos Eisley of Open source. I am just kidding, it is not that bad.
You asked why I was writing this, and I told you why.
"We also collect the content you create, upload, or receive from others when using our services. This includes things like email you write and receive, photos and videos you save, docs and spreadsheets you create, and comments you make on YouTube videos."
Their words not mine.
I am sorry you feel that way. The point of the article was not really attempting to address the feature sets of Kubernetes (which I also have issues with), it was really about the sales pitch being delivered by Google.
I have A LOT of issues with the things you posted above, and I hope to address them in future articles. Stay tuned for more, and thanks for reading.
But you still have to do that in Kubernetes, unless you are running Fargate. Someone has to provision and maintain that machine, and in the process introduce a ton of administrative overhead.
Hi danpalmer, I kind of hit on this in the article. How do you think abstracting away the server adds any value in "public cloud"? Where you can get bespoke VMs, with no concern for the underlying hardware? Can you elaborate?
Thanks for the response. I knew of this talk, and had delved into the code base years ago. I did not want to deal with that specific issue in the article because it diverted from the main discussion. My assumption this was a known aspect of the code base, from that talk to the security review Kubernetes received. Obviously it not a known, when I roll out an update to the webpage I will link to that fosdem talk.
I think the design behind Kubernetes served one purpose, normalize people's workloads. Not to reach Uber or Netflix scale. That is important in the context of the history, and I care because I don't think it is a very good product, because of this very reason.
It was definitely rushed to market, they say it in the video, and all the major players in the ecosystem, IBM, Redhat, Google, and VmWare, all wanted a piece of that pie, because all of them were barely treading water.
I don't think that is how market share is measured.
I am not sure how you decrease the power of the big 3 like this. Kubernetes does not solve that issue. I agree it is an admirable ambition, but I think you would have better luck trying to regulate them like utilities.
Hi streetcat1 thanks for responding in such a courteous manner.
To your points:
1. I am not going to get into the ethics of AWS rebranding OSS. That could be an entirely different topic. Unlike almost all other OSS products, K8s was not designed to solve OUR problems. It was developed to solve GCP's lagging marketshare. GCP needed to normalize our workloads in the hopes people could easily migrate to their cloud.
2. I am sorry that is not true. The documentary clearly spells out Kubernetes was not community first. I point out that fact several times. They even quote Urs as saying that. I believed what you believed at one point as well, I don't anymore.
3. I am not understanding your point, I am familiar with the history of Openstack.
your next points:
1. That is portability not scalability, and I agree a clear bonus brought by
Docker, not Kubernetes.
2. I am not sure the point here. If you mean I can fork it, I addressed that. AWS has way more workloads then 5%, in just straight market share and as a slice of the K8s pie.
3. I am completely familiar with the "control loop", I hope the best for your business, I guess.
So I have been playing with NanoVMs and I was able to migrate the Google Cloud Microservices Demo to AWS EC2 unikernels. In the process I also created some simple service discovery with HashiCorp's Consul as a DNS server. This was very educational.