Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Jeff Tantsura]

+1 Dean.
I’m having this discussion on daily bases...
I do care about sustainability and long term growth though 

Regards,
Jeff

> On Jan 26, 2018, at 07:30, Dean Bogdanovic  wrote:
> 
> I will ask a different question
> 
> How many people have implemented the draft? And are they talking from 
> experience implementing the model? I have implemented LNE and NI and to be 
> honest, when customers ask about IETF compatibility, i reference a draft and 
> tell them it will take long time until IETF finalizes the RFC. When it does, 
> we will update the implementation if needed. Within WG are hearing very 
> little about implementation and operational experience and feedback during 
> the process. 
> If any company had to wait two or more years to release software, they would 
> find themselves out of customers.
> 
> Dean
> 
>> On Jan 26, 2018, at 10:22 AM, Juergen Schoenwaelder 
>>  wrote:
>> 
>> OK, I accept that you do not care. Please also accept that others do
>> care. And these people believe YANG library bis is needed.
>> 
>> Since you do not want to read emails and involve yourself in
>> discussions of technical details, I assume this is where our
>> conversation stops.
>> 
>> I tought you wanted to start a constructive conversation towards a
>> resolution of the problem but it seems I misunderstood.
>> 
>> /js
>> 
>>> On Fri, Jan 26, 2018 at 10:06:06AM -0500, Christian Hopps wrote:
>>> 
>>> In the context of holding up this work, I don't care one iota about YANG
>>> library bis, and it works just fine with NMDA AFAICT.
>>> 
>>> We need models to get work done.
>>> 
>>> Thanks,
>>> Chris.
>>> 
>>> Juergen Schoenwaelder  writes:
>>> 
> On Fri, Jan 26, 2018 at 09:18:55AM -0500, Christian Hopps wrote:
> 
> Now it seems we are supposed to wait a bunch longer on yet other works
> in progress for as near as I can tell (could be wrong here as I just
> don't have time to read the very long email threads that netmod
> generates) capturing meta-data in a cleaner way than another. This does
> *not* seem like a reason to stall this work any further.
> 
 
 What is your interpretation of 'a bunch longer'? Or said differently,
 how much time do you think it will take to get the current schema
 mount approved (which has pending WG last call issues) and how much
 time would you find acceptable for a solution that also complies with
 NMDA and YANG library bis? I believe people are willing to give the
 later high priority.
 
 /js
>> 
>> -- 
>> Juergen Schoenwaelder   Jacobs University Bremen gGmbH
>> Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany
>> Fax:   +49 421 200 3103 
>> 
>> ___
>> netmod mailing list
>> netmod@ietf.org
>> https://www.ietf.org/mailman/listinfo/netmod
> 
> ___
> netmod mailing list
> netmod@ietf.org
> https://www.ietf.org/mailman/listinfo/netmod
___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Juergen Schoenwaelder]

Chris,

nobody (I think) is having an issue with the lne and ni models. And it
is great to hear that you have implemented them. However, please also
understand that schema mount is a rather fundamental extension of the
YANG technology and that people maintaining that technology and
writing generic tools need to have this extension work together with
other extensions we have been working on. For you, this may be
irrelevant or not important and it may be 'divorced from your "get
shit done" reality'. But then, please accept that there are other
communities with different priorities.

I believe we will be more effective if we find a way forward that can
work for everybody involved. Trying to convince other communities that
their priorities are plain wrong is neither constructive nor does it
lead to faster progress I think.

/js

On Fri, Jan 26, 2018 at 11:40:39AM -0500, Christian Hopps wrote:
> 
> Hi Juergen,
> 
> I want to be understood so I'll reply again. It's not that I don't want
> to involve myself in technical discussions, it's that I (and others)
> think that what's being discussed now no longer matters to getting work
> done. The work is good enough *now*. When we get to this point it
> doesn't make sense for me to participate anymore, the problem is solved,
> I need to work on other problems that aren't solved yet.
> 
> We need models for VRFs and VMs, people are now arguing about having
> totally different schema mounted at the same mount point based on the
> datastore (?!?) and where exactly the meta-data should reside. It's
> divorced from the "get shit done" reality.
> 
> Thanks,
> Chris.
> 
> Juergen Schoenwaelder  writes:
> 
> > OK, I accept that you do not care. Please also accept that others do
> > care. And these people believe YANG library bis is needed.
> >
> > Since you do not want to read emails and involve yourself in
> > discussions of technical details, I assume this is where our
> > conversation stops.
> >
> > I tought you wanted to start a constructive conversation towards a
> > resolution of the problem but it seems I misunderstood.
> 
> > /js
> >
> > On Fri, Jan 26, 2018 at 10:06:06AM -0500, Christian Hopps wrote:
> >>
> >> In the context of holding up this work, I don't care one iota about YANG
> >> library bis, and it works just fine with NMDA AFAICT.
> >>
> >> We need models to get work done.
> >>
> >> Thanks,
> >> Chris.
> >>
> >> Juergen Schoenwaelder  writes:
> >>
> >> > On Fri, Jan 26, 2018 at 09:18:55AM -0500, Christian Hopps wrote:
> >> >>
> >> >> Now it seems we are supposed to wait a bunch longer on yet other works
> >> >> in progress for as near as I can tell (could be wrong here as I just
> >> >> don't have time to read the very long email threads that netmod
> >> >> generates) capturing meta-data in a cleaner way than another. This does
> >> >> *not* seem like a reason to stall this work any further.
> >> >>
> >> >
> >> > What is your interpretation of 'a bunch longer'? Or said differently,
> >> > how much time do you think it will take to get the current schema
> >> > mount approved (which has pending WG last call issues) and how much
> >> > time would you find acceptable for a solution that also complies with
> >> > NMDA and YANG library bis? I believe people are willing to give the
> >> > later high priority.
> >> >
> >> > /js

-- 
Juergen Schoenwaelder   Jacobs University Bremen gGmbH
Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany
Fax:   +49 421 200 3103 

___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Christian Hopps]

Hi Juergen,

I want to be understood so I'll reply again. It's not that I don't want
to involve myself in technical discussions, it's that I (and others)
think that what's being discussed now no longer matters to getting work
done. The work is good enough *now*. When we get to this point it
doesn't make sense for me to participate anymore, the problem is solved,
I need to work on other problems that aren't solved yet.

We need models for VRFs and VMs, people are now arguing about having
totally different schema mounted at the same mount point based on the
datastore (?!?) and where exactly the meta-data should reside. It's
divorced from the "get shit done" reality.

Thanks,
Chris.

Juergen Schoenwaelder  writes:

> OK, I accept that you do not care. Please also accept that others do
> care. And these people believe YANG library bis is needed.
>
> Since you do not want to read emails and involve yourself in
> discussions of technical details, I assume this is where our
> conversation stops.
>
> I tought you wanted to start a constructive conversation towards a
> resolution of the problem but it seems I misunderstood.

> /js
>
> On Fri, Jan 26, 2018 at 10:06:06AM -0500, Christian Hopps wrote:
>>
>> In the context of holding up this work, I don't care one iota about YANG
>> library bis, and it works just fine with NMDA AFAICT.
>>
>> We need models to get work done.
>>
>> Thanks,
>> Chris.
>>
>> Juergen Schoenwaelder  writes:
>>
>> > On Fri, Jan 26, 2018 at 09:18:55AM -0500, Christian Hopps wrote:
>> >>
>> >> Now it seems we are supposed to wait a bunch longer on yet other works
>> >> in progress for as near as I can tell (could be wrong here as I just
>> >> don't have time to read the very long email threads that netmod
>> >> generates) capturing meta-data in a cleaner way than another. This does
>> >> *not* seem like a reason to stall this work any further.
>> >>
>> >
>> > What is your interpretation of 'a bunch longer'? Or said differently,
>> > how much time do you think it will take to get the current schema
>> > mount approved (which has pending WG last call issues) and how much
>> > time would you find acceptable for a solution that also complies with
>> > NMDA and YANG library bis? I believe people are willing to give the
>> > later high priority.
>> >
>> > /js

___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Dean Bogdanovic]

I will ask a different question

How many people have implemented the draft? And are they talking from 
experience implementing the model? I have implemented LNE and NI and to be 
honest, when customers ask about IETF compatibility, i reference a draft and 
tell them it will take long time until IETF finalizes the RFC. When it does, we 
will update the implementation if needed. Within WG are hearing very little 
about implementation and operational experience and feedback during the 
process. 
If any company had to wait two or more years to release software, they would 
find themselves out of customers.

Dean

> On Jan 26, 2018, at 10:22 AM, Juergen Schoenwaelder 
>  wrote:
> 
> OK, I accept that you do not care. Please also accept that others do
> care. And these people believe YANG library bis is needed.
> 
> Since you do not want to read emails and involve yourself in
> discussions of technical details, I assume this is where our
> conversation stops.
> 
> I tought you wanted to start a constructive conversation towards a
> resolution of the problem but it seems I misunderstood.
> 
> /js
> 
> On Fri, Jan 26, 2018 at 10:06:06AM -0500, Christian Hopps wrote:
>> 
>> In the context of holding up this work, I don't care one iota about YANG
>> library bis, and it works just fine with NMDA AFAICT.
>> 
>> We need models to get work done.
>> 
>> Thanks,
>> Chris.
>> 
>> Juergen Schoenwaelder  writes:
>> 
>>> On Fri, Jan 26, 2018 at 09:18:55AM -0500, Christian Hopps wrote:
 
 Now it seems we are supposed to wait a bunch longer on yet other works
 in progress for as near as I can tell (could be wrong here as I just
 don't have time to read the very long email threads that netmod
 generates) capturing meta-data in a cleaner way than another. This does
 *not* seem like a reason to stall this work any further.
 
>>> 
>>> What is your interpretation of 'a bunch longer'? Or said differently,
>>> how much time do you think it will take to get the current schema
>>> mount approved (which has pending WG last call issues) and how much
>>> time would you find acceptable for a solution that also complies with
>>> NMDA and YANG library bis? I believe people are willing to give the
>>> later high priority.
>>> 
>>> /js
> 
> -- 
> Juergen Schoenwaelder   Jacobs University Bremen gGmbH
> Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany
> Fax:   +49 421 200 3103 
> 
> ___
> netmod mailing list
> netmod@ietf.org
> https://www.ietf.org/mailman/listinfo/netmod

___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Juergen Schoenwaelder]

OK, I accept that you do not care. Please also accept that others do
care. And these people believe YANG library bis is needed.

Since you do not want to read emails and involve yourself in
discussions of technical details, I assume this is where our
conversation stops.

I tought you wanted to start a constructive conversation towards a
resolution of the problem but it seems I misunderstood.

/js

On Fri, Jan 26, 2018 at 10:06:06AM -0500, Christian Hopps wrote:
> 
> In the context of holding up this work, I don't care one iota about YANG
> library bis, and it works just fine with NMDA AFAICT.
> 
> We need models to get work done.
> 
> Thanks,
> Chris.
> 
> Juergen Schoenwaelder  writes:
> 
> > On Fri, Jan 26, 2018 at 09:18:55AM -0500, Christian Hopps wrote:
> >>
> >> Now it seems we are supposed to wait a bunch longer on yet other works
> >> in progress for as near as I can tell (could be wrong here as I just
> >> don't have time to read the very long email threads that netmod
> >> generates) capturing meta-data in a cleaner way than another. This does
> >> *not* seem like a reason to stall this work any further.
> >>
> >
> > What is your interpretation of 'a bunch longer'? Or said differently,
> > how much time do you think it will take to get the current schema
> > mount approved (which has pending WG last call issues) and how much
> > time would you find acceptable for a solution that also complies with
> > NMDA and YANG library bis? I believe people are willing to give the
> > later high priority.
> >
> > /js

-- 
Juergen Schoenwaelder   Jacobs University Bremen gGmbH
Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany
Fax:   +49 421 200 3103 

___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Juergen Schoenwaelder]

On Fri, Jan 26, 2018 at 09:18:55AM -0500, Christian Hopps wrote:
> 
> Now it seems we are supposed to wait a bunch longer on yet other works
> in progress for as near as I can tell (could be wrong here as I just
> don't have time to read the very long email threads that netmod
> generates) capturing meta-data in a cleaner way than another. This does
> *not* seem like a reason to stall this work any further.
>

What is your interpretation of 'a bunch longer'? Or said differently,
how much time do you think it will take to get the current schema
mount approved (which has pending WG last call issues) and how much
time would you find acceptable for a solution that also complies with
NMDA and YANG library bis? I believe people are willing to give the
later high priority.

/js

-- 
Juergen Schoenwaelder   Jacobs University Bremen gGmbH
Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany
Fax:   +49 421 200 3103 

___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

Re: [netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[David Bannister]

Chris +1 on the taking too long in pursuit of the perfect model.  No
service provider or enterprise wants to put their network evolution on hold
waiting for the IETF.  Instead they will seek what they need from other
SDOs as you point out.  The IETF needs to modernize their process and
perhaps adopt some of the methods used by OpenConfig.  Read that as getting
away from a long drawn out consensus process that turns into a multi-year
effort.

On Fri, Jan 26, 2018 at 9:18 AM, Christian Hopps  wrote:

>
> Maybe a meeting at this point is useful? It would consolidate things and
> get away from the endless email threads.
>
> If this isn't already known to everyone. There are many people for whom
> the length of time to market from IETF simple doesn't work in particular
> with models. That's one big reason that openconfig exists. Sitting on
> working solutions waiting for them to be perfect is just getting us
> ignored by industry.
>
> In particular when I, Lou, et al. realized we needed a way to "mount
> schema" for a clean VRF and VM solution, we thought this was a simple
> thing and we could do it rather quickly -- the concept is just not that
> complex. The idea was picked up by Martin and Lada who produced drafts,
> and there were in fact some devil in the details and those got worked
> out over longer than anyone wanted, but it is what it is.
>
> Now it seems we are supposed to wait a bunch longer on yet other works
> in progress for as near as I can tell (could be wrong here as I just
> don't have time to read the very long email threads that netmod
> generates) capturing meta-data in a cleaner way than another. This does
> *not* seem like a reason to stall this work any further.
>
> Thanks,
> Chris.
>
> Ladislav Lhotka  writes:
>
> > Hi,
> >
> > Kent Watsen  writes:
> >
> >> Thank you all for the important discussion since the completion of WGLC
> on Nov 6th.
> >>
> >> Per normal process, drafts typically progress once LC comments are
> >> address unless significant faults are found.  Post LC comments have
> >> been made, which needed consideration, notably the relationship with
> >> NMDA and rfc7895bis and an alternate representation of inline schema.
> >> These have been considered respecting their impact on the last call
> >> consensus and it is the position of the chairs that it is best to
> >> advance the existing schema-mount document at this time.
> >
> > I guess I have no chance but strongly object to this. Is it normal to
> > proceed this way without reaching WG consensus and against the will of
> > *both* document authors?
> >
> >>
> >> Given that there are significant concerns for how the solution
> >> proposed in this draft operates with NMDA, we do think it reasonable
> >> to add an applicability statement to the draft that covers its
> >> operation in NMDA implementations. We do not believe that such a
> >> statement substantively alters the draft nor would it impact drafts
> >> that normatively reference the current draft.
> >>
> >> In addition to resolving the remaining open thread [1],
> >
> > Hmm, who resolved this thread? Lou proposed some text and nobody
> > expressed any agreement with it. In fact, I believe it is nothing more
> > than hand-waving.
> >
> > I must say that the work on this draft was very frustrating for me. Even
> > more than on RFC 8022, and this tells you something.
> >
> > Lada
> >
> >> we also agree
> >> with the recently made comment that the schema mount draft should
> >> allow the use of rfc7895bis (i.e., not reference /modules-state),
> >> thereby enabling the draft's use (though not ideal) on servers
> >> supporting rfc7895bis.
> >>
> >> The chairs will propose specific text for the updates mentioned in this
> message to be reviewed by the WG for correctness before final submission
> and advancement.
> >>
> >> [1] https://www.ietf.org/mail-archive/web/netmod/current/msg20049.html
> >>
> >> Thanks,
> >> Kent, Lou, and Joel
> >>
> >>
> >>
> >> ___
> >> netmod mailing list
> >> netmod@ietf.org
> >> https://www.ietf.org/mailman/listinfo/netmod
>
> ___
> netmod mailing list
> netmod@ietf.org
> https://www.ietf.org/mailman/listinfo/netmod
>
___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod

[netmod] Live meeting? and my opinion. [Re: moving forward with schema mount]

[Christian Hopps]

Maybe a meeting at this point is useful? It would consolidate things and
get away from the endless email threads.

If this isn't already known to everyone. There are many people for whom
the length of time to market from IETF simple doesn't work in particular
with models. That's one big reason that openconfig exists. Sitting on
working solutions waiting for them to be perfect is just getting us
ignored by industry.

In particular when I, Lou, et al. realized we needed a way to "mount
schema" for a clean VRF and VM solution, we thought this was a simple
thing and we could do it rather quickly -- the concept is just not that
complex. The idea was picked up by Martin and Lada who produced drafts,
and there were in fact some devil in the details and those got worked
out over longer than anyone wanted, but it is what it is.

Now it seems we are supposed to wait a bunch longer on yet other works
in progress for as near as I can tell (could be wrong here as I just
don't have time to read the very long email threads that netmod
generates) capturing meta-data in a cleaner way than another. This does
*not* seem like a reason to stall this work any further.

Thanks,
Chris.

Ladislav Lhotka  writes:

> Hi,
>
> Kent Watsen  writes:
>
>> Thank you all for the important discussion since the completion of WGLC on 
>> Nov 6th.
>>
>> Per normal process, drafts typically progress once LC comments are
>> address unless significant faults are found.  Post LC comments have
>> been made, which needed consideration, notably the relationship with
>> NMDA and rfc7895bis and an alternate representation of inline schema.
>> These have been considered respecting their impact on the last call
>> consensus and it is the position of the chairs that it is best to
>> advance the existing schema-mount document at this time.
>
> I guess I have no chance but strongly object to this. Is it normal to
> proceed this way without reaching WG consensus and against the will of
> *both* document authors?
>
>>
>> Given that there are significant concerns for how the solution
>> proposed in this draft operates with NMDA, we do think it reasonable
>> to add an applicability statement to the draft that covers its
>> operation in NMDA implementations. We do not believe that such a
>> statement substantively alters the draft nor would it impact drafts
>> that normatively reference the current draft.
>>
>> In addition to resolving the remaining open thread [1],
>
> Hmm, who resolved this thread? Lou proposed some text and nobody
> expressed any agreement with it. In fact, I believe it is nothing more
> than hand-waving.
>
> I must say that the work on this draft was very frustrating for me. Even
> more than on RFC 8022, and this tells you something.
>
> Lada
>
>> we also agree
>> with the recently made comment that the schema mount draft should
>> allow the use of rfc7895bis (i.e., not reference /modules-state),
>> thereby enabling the draft's use (though not ideal) on servers
>> supporting rfc7895bis.
>>
>> The chairs will propose specific text for the updates mentioned in this 
>> message to be reviewed by the WG for correctness before final submission and 
>> advancement.
>>
>> [1] https://www.ietf.org/mail-archive/web/netmod/current/msg20049.html
>>
>> Thanks,
>> Kent, Lou, and Joel
>>
>>
>>
>> ___
>> netmod mailing list
>> netmod@ietf.org
>> https://www.ietf.org/mailman/listinfo/netmod

___
netmod mailing list
netmod@ietf.org
https://www.ietf.org/mailman/listinfo/netmod