Re: [Webpush] Benoit Claise's No Objection on draft-ietf-webpush-protocol-11: (with COMMENT)

[Dan Romascanu <[e-mail]>]

Hi,

I agree with Benoit that the document in its current phase does not seem to
include any show stopper. As the OPS-DIR reviews are written to help the
OPS ADs, its up to them if the lack of some of the operational and
manageability considerations is critical for such documents.

I do with however to express my reservations about Martin's statement that
the would not be 'an awful lot of actionable material in
this review'. The review clearly points to the lack of considerations
related to a number of operational and manageability aspects like
scalability, coexistence, initial setup, fault management, etc. One can
decide and be explicit or not about dealing with these, but how to fix the
lack of documentation about these aspects is actionable in my opinion.

I also do not see why RFC 5706 would not apply for such a document that
describes a standards track IETF protocol extension. OPS-DIR reviewers are
referred to RFC 5706 for such documents.

Also, I disagree with the statement that '[HTTP operational practions] ...
this was
assumed to be obvious enough to omit.' The Security Consideration section
starts with the sentence:

> This protocol MUST use HTTP over TLS [RFC2818] following the

   recommendations in [RFC7525].

Security considerations are based on previous documents and this is
explicitly stated. There is no reason not to state explicitly the same
for the manageability considerations

Regards,

Dan


On Fri, Oct 14, 2016 at 1:38 PM, Benoit Claise <[e-mail]> wrote:

> Martin,
>
> Here is the high level message: no show stopper here from my perspective,
> but from an OPS point of view it could be better documented.
>
> Regards, Benoit
>
>> On 14 October 2016 at 00:36, Benoit Claise <[e-mail]> wrote:
>>
>>> Benoit Claise has entered the following ballot position for
>>> draft-ietf-webpush-protocol-11: No Objection
>>>
>> [...]
>>
>>> Here is Dan Romascanu's OPS DIR review:
>>>
>> I have confess, I don't see an awful lot of actionable material in
>> this review.  I worry that RFC 5706 isn't an entirely appropriate
>> template for a review of this sort of thing.  I will concede that the
>> level of consistency across the industry when it comes to managing and
>> operating HTTP servers and services is almost uniformly inconsistent.
>>
>> Dan does observe that the draft doesn't mention that typical HTTP
>> operational practices apply; it was my understanding that this was
>> assumed to be obvious enough to omit.
>>
>> (I'm happy to have a more involved conversation about operational and
>> management concerns with push services, but it's probably something
>> that is better done with a different CC list.)
>> .
>>
>>
>