|
|
toggle quoted messageShow quoted text
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
|
|
Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
toggle quoted messageShow quoted text
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
|
|
Of the major benefits highlighted, #3 - developer velocity (and the inherent increase in consistency of infrastructure configuration) - is not to be overshadowed by the the first major benefit of performance improvements. Each are substantial.
The goal of the CNF Testbed's apples-to-apples comparison resonates with the goal of the Meshery project with aims for apples-to-apples performance comparison between prominent service mesh projects. Meshery is an alpha state project slated for demo in an upcoming Networking WG call.
toggle quoted messageShow quoted text
On Feb 25, 2019, at 7:49 AM, Liz Rice < liz@...> wrote:
Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan? On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson < alexis@...> wrote:
--
|
|
Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project.
toggle quoted messageShow quoted text
On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
|
|
If so then I assume this article is all a set of misquotes by the journalist:
toggle quoted messageShow quoted text
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|
Chris, can we put this on the agenda for the next TOC meeting please.
I'd like the TOC to have a clear understanding of what's happening with the Cloud-native Network Functions effort, as it's very unclear at the moment. See also the thread below.
Thanks
Q
From: Dan Kohn [dan@...]
Sent: Wednesday, January 30, 2019 1:00 PM
To: Quinton Hoole
Cc: Chris Aniszczyk; Alexis Richardson
Subject: Re: CNFS Cloud-native Network Functions
That's a mistake, then. I will make clear the problem to them and ask them to get my signoff first on future submissions.
Thanks for clarifying Dan
There are several proposed Kubecon talks (e.g. for Barcelona in May) on Cloud-native Network Functions, some by CNCF-paid contractors, which do not portray it in that light.
Q
Here's background on CNFs. In short, it's a demonstration and marketing term for how to use Kubernetes as a substrate for the telco use case. There's no project associated with it, and a key consideration is how to support telcos without "breaking"
Kubernetes's success for enterprise users. There are no actual upstream changes or new projects that we're proposing.
On Wed, Jan 30, 2019 at 2:16 AM Chris Aniszczyk < caniszczyk@...> wrote:
+Dan Kohn
The nomenclature is a bit off but this is a CNCF funded effort / demo, not an official CNCF project per se.
Hi Alexis, Chris
I recently noticed that CNFS claim to be “in the CNCF”.
I think this is the first I’ve heard of this. Am I asleep at the wheel? Was this ever passed by the TOC? Where exactly does it fit in?
There is a CNFS project in the CNCF GitHub:
But it’s not listed as a sandbox (or any other) CNCF project?
What am I missing here?
Thanks for any guidance.
Q
--
Chris Aniszczyk (@cra) | +1-512-961-6719
From: cncf-toc@... [cncf-toc@...] on behalf of alexis richardson [alexis@...]
Sent: Thursday, February 28, 2019 10:09 AM
To: Dan Kohn
Cc: Alexis Richardson via cncf-toc; Liz Rice
Subject: Re: [cncf-toc] Anyone know what this is?
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote:
Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project.
On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote:
Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|

Chris Aniszczyk
sure I’ll slot a bit of time
toggle quoted messageShow quoted text
On Mar 1, 2019, at 7:56 AM, Quinton Hoole < quinton.hoole@...> wrote:
Chris, can we put this on the agenda for the next TOC meeting please.
I'd like the TOC to have a clear understanding of what's happening with the Cloud-native Network Functions effort, as it's very unclear at the moment. See also the thread below.
Thanks
Q
From: Dan Kohn [dan@...]
Sent: Wednesday, January 30, 2019 1:00 PM
To: Quinton Hoole
Cc: Chris Aniszczyk; Alexis Richardson
Subject: Re: CNFS Cloud-native Network Functions
That's a mistake, then. I will make clear the problem to them and ask them to get my signoff first on future submissions.
Thanks for clarifying Dan
There are several proposed Kubecon talks (e.g. for Barcelona in May) on Cloud-native Network Functions, some by CNCF-paid contractors, which do not portray it in that light.
Q
Here's background on CNFs. In short, it's a demonstration and marketing term for how to use Kubernetes as a substrate for the telco use case. There's no project associated with it, and a key consideration is how to support telcos without "breaking"
Kubernetes's success for enterprise users. There are no actual upstream changes or new projects that we're proposing.
On Wed, Jan 30, 2019 at 2:16 AM Chris Aniszczyk < caniszczyk@...> wrote:
+Dan Kohn
The nomenclature is a bit off but this is a CNCF funded effort / demo, not an official CNCF project per se.
Hi Alexis, Chris
I recently noticed that CNFS claim to be “in the CNCF”.
I think this is the first I’ve heard of this. Am I asleep at the wheel? Was this ever passed by the TOC? Where exactly does it fit in?
There is a CNFS project in the CNCF GitHub:
But it’s not listed as a sandbox (or any other) CNCF project?
What am I missing here?
Thanks for any guidance.
Q
--
Chris Aniszczyk (@cra) | +1-512-961-6719
From: cncf-toc@... [cncf-toc@...] on behalf of alexis richardson [alexis@...]
Sent: Thursday, February 28, 2019 10:09 AM
To: Dan Kohn
Cc: Alexis Richardson via cncf-toc; Liz Rice
Subject: Re: [cncf-toc] Anyone know what this is?
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote:
Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project.
On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote:
Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson < alexis@...> wrote:
--
--
|
|
The statement about competing with openstack is especially unfortunate.
And all solutions have noisy neighbor challenges. It's workload dependent.
toggle quoted messageShow quoted text
On Thu, Feb 28, 2019, 10:09 AM alexis richardson <alexis@...> wrote:
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|
The statement about competing with openstack is especially unfortunate.
I agree, I'm very disappointed by that statement.
And all solutions have noisy neighbor challenges. It's workload dependent.
Indeed. Similarly all solutions have security challenges. And then pointing to microvms as interesting after describing Kata as FUD is minimally confusing and maximally reads as just bad mouthing another project.
Having said the above, that's mostly about messaging concerns. I do see real value in Kubernetes on bare metal for a number of usecases including network functions.
So I'm curious to see who is interested in Kubernetes on bare metal? For those interested, we're working on a project called metalkube:
Feel free to jump on slack and learn what's going on there.
thanks, -chris
On Thu, Feb 28, 2019, 10:09 AM alexis richardson <alexis@...> wrote:
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|
Xuan Jia <jason.jiaxuan@...>
Hi Chris, i am very interested in this project. It may have some value in edge computing. We are finding if community have any solution to solve the bare metal machine management problem . MetalKube is the one. How can we do together ?
From my point of view, the resource in Edge Computing Data Center is limited. CNF is a very good chose in Edge.
Xuan Jia Edge Computing Architect China Mobile Research Institute
toggle quoted messageShow quoted text
The statement about competing with openstack is especially unfortunate.
I agree, I'm very disappointed by that statement.
And all solutions have noisy neighbor challenges. It's workload dependent.
Indeed. Similarly all solutions have security challenges. And then pointing to microvms as interesting after describing Kata as FUD is minimally confusing and maximally reads as just bad mouthing another project.
Having said the above, that's mostly about messaging concerns. I do see real value in Kubernetes on bare metal for a number of usecases including network functions.
So I'm curious to see who is interested in Kubernetes on bare metal? For those interested, we're working on a project called metalkube:
Feel free to jump on slack and learn what's going on there.
thanks, -chris
On Thu, Feb 28, 2019, 10:09 AM alexis richardson <alexis@...> wrote:
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|
That's great Xuan! I agree, Kube is very promising for a variety of edge and other usecases in the network.
Jump on #wg-onprem in Kubernetes Slack, visit http://metalkube.org/, or ping Russell (Cc'd) if you need any other pointers. thanks, -chris
toggle quoted messageShow quoted text
Hi Chris, i am very interested in this project. It may have some value in edge computing. We are finding if community have any solution to solve the bare metal machine management problem . MetalKube is the one. How can we do together ?
From my point of view, the resource in Edge Computing Data Center is limited. CNF is a very good chose in Edge.
Xuan Jia Edge Computing Architect China Mobile Research Institute
The statement about competing with openstack is especially unfortunate.
I agree, I'm very disappointed by that statement.
And all solutions have noisy neighbor challenges. It's workload dependent.
Indeed. Similarly all solutions have security challenges. And then pointing to microvms as interesting after describing Kata as FUD is minimally confusing and maximally reads as just bad mouthing another project.
Having said the above, that's mostly about messaging concerns. I do see real value in Kubernetes on bare metal for a number of usecases including network functions.
So I'm curious to see who is interested in Kubernetes on bare metal? For those interested, we're working on a project called metalkube:
Feel free to jump on slack and learn what's going on there.
thanks, -chris
On Thu, Feb 28, 2019, 10:09 AM alexis richardson <alexis@...> wrote:
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|
> "So I'm curious to see who is interested in Kubernetes on bare metal? "
If the TOC is interested in Kube on bare metal... there's also KRIB on Digital Rebar which uses KubeAdm and has a wide range of options including HA.

toggle quoted messageShow quoted text
On Fri, Mar 1, 2019 at 10:01 AM Chris Wright < chrisw@...> wrote: The statement about competing with openstack is especially unfortunate.
I agree, I'm very disappointed by that statement.
And all solutions have noisy neighbor challenges. It's workload dependent.
Indeed. Similarly all solutions have security challenges. And then pointing to microvms as interesting after describing Kata as FUD is minimally confusing and maximally reads as just bad mouthing another project.
Having said the above, that's mostly about messaging concerns. I do see real value in Kubernetes on bare metal for a number of usecases including network functions.
So I'm curious to see who is interested in Kubernetes on bare metal? For those interested, we're working on a project called metalkube:
Feel free to jump on slack and learn what's going on there.
thanks, -chris
On Thu, Feb 28, 2019, 10:09 AM alexis richardson <alexis@...> wrote:
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|
Russell Bryant <rbryant@...>
Thanks, Chris! That channel just got archived on slack, unfortunately. We are in #cluster-api as well, though.
we can set up a mailing list for metalkube too if that helps.
Russell Bryant
toggle quoted messageShow quoted text
On Sat, Mar 2, 2019 at 1:24 PM Chris Wright < chrisw@...> wrote: That's great Xuan! I agree, Kube is very promising for a variety of edge and other usecases in the network.
Jump on #wg-onprem in Kubernetes Slack, visit http://metalkube.org/, or ping Russell (Cc'd) if you need any other pointers. thanks, -chris Hi Chris, i am very interested in this project. It may have some value in edge computing. We are finding if community have any solution to solve the bare metal machine management problem . MetalKube is the one. How can we do together ?
From my point of view, the resource in Edge Computing Data Center is limited. CNF is a very good chose in Edge.
Xuan Jia Edge Computing Architect China Mobile Research Institute
The statement about competing with openstack is especially unfortunate.
I agree, I'm very disappointed by that statement.
And all solutions have noisy neighbor challenges. It's workload dependent.
Indeed. Similarly all solutions have security challenges. And then pointing to microvms as interesting after describing Kata as FUD is minimally confusing and maximally reads as just bad mouthing another project.
Having said the above, that's mostly about messaging concerns. I do see real value in Kubernetes on bare metal for a number of usecases including network functions.
So I'm curious to see who is interested in Kubernetes on bare metal? For those interested, we're working on a project called metalkube:
Feel free to jump on slack and learn what's going on there.
thanks, -chris
On Thu, Feb 28, 2019, 10:09 AM alexis richardson <alexis@...> wrote:
If so then I assume this article is all a set of misquotes by the journalist:
On Mon, 25 Feb 2019, 15:58 Dan Kohn, < dan@...> wrote: Definitely, there's a lot more to come. We're also hoping to get a bunch of telcos and their vendors engaged and making improvement to the Testbed so that it becomes a real, collaborative open source project. On Mon, Feb 25, 2019 at 8:49 AM Liz Rice < liz@...> wrote: Fwiw I think this is pretty cool, but I suspect a lot of parties interested in this comparison would also want to see latency & jitter info added to slide 9. Is that in the pipeline Dan?
On Mon, 25 Feb 2019 at 10:18, Dan Kohn < dan@...> wrote:
On Mon, Feb 25, 2019 at 4:13 AM alexis richardson <alexis@...> wrote:
--
--
|
|