How does Borg compare with Google's Kubernetes?
There's a whole section about it in the paper (lessons learned and future work).
Large-scale cluster management at Google with Borg
11–20 of 37 posts
Re: Large-scale cluster management at Google with Borg
#12Was always in awe of Borg and Omega while at Google. Really nice to see them finally publish a paper on this. Guess they've more far enough along now that it makes sense to do so. Omega will be a far superior beast to Borg and the open source Kubernetes but I have high hopes for the future of Kubernetes.
Re: Large-scale cluster management at Google with Borg
#13Apache Yarn[1] looks like same thing as borg\omega. A plus point with Yarn is we can get our hands on it.
[1] http://hadoop.apache.org/docs/current/hadoop-yarn/hadoop-yar...
Re: Large-scale cluster management at Google with Borg
#14Apache Yarn[1] looks like same thing as borg\omega. A plus point with Yarn is we can get our hands on it. [1] http://hadoop.apache.org/docs/current/hadoop-yarn/hadoop-yar...
Yarn tries to do the same thing, but is quite different. It's skewed toward running short running jobs, like batch jobs, it was meant for hadoop and co after all.
Borg/Omega seems more like a combination of Mesos at the scheduler layer and Marathon/Kubernetes on top. It's funny to see how many services they run on top of it though.
Re: Large-scale cluster management at Google with Borg
#15Is this similar in concept to Oracle Grid Engine (SGE)? How is it different, superior?
Re: Large-scale cluster management at Google with Borg
#16Is this similar in concept to Oracle Grid Engine (SGE)? How is it different, superior?
Yes, it's similar, but: it has a concept of services, which are jobs that run forever. It has a concept of allocs, which are permastorage that resides on a machine, where jobs can be rescheduled repeatedly that share the same storage. it scales to much larger size and is better at managing the fleet of machiens and scheduling work there.
Re: Large-scale cluster management at Google with Borg
#17> cc would be reachable via 50.jfoo.ubar.cc.borg.google.com. I've implemented similar at my current job, as that sort of naming is very convenient. http://www.boxever.com/using-google-apps-openid-connect-with... has a sketch of how to do this with Apache as a reverse proxy with Google Auth, though we're using a PAC file now going to a HTTPS forward proxy to avoid limitations of SSL wildcard certs.
Keep it up, Brian :-)
Re: Large-scale cluster management at Google with Borg
#18Is this similar in concept to Oracle Grid Engine (SGE)? How is it different, superior?
I started doing SGE stuff in ~2001...It still amazes me how well that thing worked and how often a lot of it's features are reinvented.
Re: Large-scale cluster management at Google with Borg
#19Is this similar in concept to Oracle Grid Engine (SGE)? How is it different, superior?
I started doing SGE stuff in ~2001...It still amazes me how well that thing worked and how often a lot of it's features are reinvented.
I agree. It's really simple to get started (short learning curve) and it's completely language agnostic.
I'm amazed how often it gets ignored. There's even StarCluster so you can automatically set up a cluster on EC2.
Re: Large-scale cluster management at Google with Borg
#20There was a great talk by John Wilkes (Google Cluster Management) re: Omega in 2011 at Google Faculty Summit [1]. Absolutely fascinating to see the scope of the problems they are dealing with. [1] https://www.youtube.com/watch?v=0ZFMlO98Jkc Edit: remove error in my comment re: borg/omega order.
If you want to see an up-to-date talk about this from John, he will be a speaker at http://dotscale.io on June 8!