[mark_l_watson]

I am a little surprised that Google is join this route. Making Google Cloud into a huge business must be a high priority. I used Borg for a while in 2013, and it was amazingly, specially the logging for tracking down runtime problems.

[fweespeech]

Why? Google hands out all sorts of papers and such a few years after the fact.

This is just another example, an open source version of Borg that will always be a few years behind.

[thockingoog]

While it is true that Borg is 10 years ahead of Kubernetes, we have the benefit of knowing where Borg arrived rather than following the same winding course.

[fweespeech]

Of course, but unless there is a massive maintainability flaw in Borg, its unlikely you can duplicate 10 years of effort AND maintain pace with Borg's/Borg's successor(s) progress.

Well, without a huge influx of developers to the open source side.

[thockingoog]

First, there are a LOT of people working on Kubernetes right now.

Second, of course it will be asymptotic for a long time. But keep in mind that we don't want or need everything that Borg has, and Borg does not have everything that Kubernetes does. Kubernetes is not a clone of Borg - it is inspired by the lessons and experiences we got by doing Borg. Often that means we learned how NOT to do something.

If we can get 75% as functional as Borg in 1/3 the time, we'll be doing pretty darn well. I think we're on track for that.

[fweespeech]

> If we can get 75% as functional as Borg in 1/3 the time, we'll be doing pretty darn well. I think we're on track for that.

I'm sure you can do that. That doesn't invalidate my statement that it isn't surprising Google would do this precisely because it'll only be 75% as functional as Borg at any given point in time.