Re: Call for Adoption: draft-pauly-quic-datagram

[Kazuho Oku <[e-mail]>]

+1 for adoption.

I agree that there's concern that the DATAGRAM draft is too generic, but
the flip side of that is that it is good for running experiments / private
protocols that only require datagrams. And we can rely on ALPN for what is
being transmitted using the DATAGRAM frames.

To paraphrase, I consider DATAGRAM to be as useful as generic UDP is.


2019年12月12日(木) 10:07 Lucas Pardue <[e-mail]>:

> I support adoption of this draft but will note that in the abstract sense
> it is hard to reason about the design of DATAGRAM and hard to interop
> without an application of it. Since the round of call for adoption does not
> incorporate an application, I have concern about the scope.. I wrote up
> SiDUCK [1] as means to an end, there are probably better ways to make us
> comfortable as a WG.
>
> [1] - https://tools.ietf.org/html/draft-pardue-quic-siduck-0
>
> On Wed, Dec 11, 2019 at 11:03 PM Jana Iyengar <[e-mail]> wrote:
>
>> I support adoption of this document.
>>
>> This is a natural extension to QUIC and there are multiple
>> implementations and side-conversations happening already, and as I
>> understand it, there is some deployment experience.
>>
>> One note however. Per my read, this is the first extension that extends
>> the functionality of QUIC beyond the charter's original conception of it
>> for HTTP. Many things are possible with QUIC, but the things we want to
>> standardize are the things that are likely to be used. There's no better
>> argument for useful extensions than a discussion of actual deployment where
>> it exists... It enables others to use it similarly, and importantly, it
>> informs why the mechanism is designed so. I would have hoped for more
>> active discussions around current uses of this extension where it exists,
>> including deployment experience where it exists. Perhaps the authors can
>> take this as feedback going forward.
>>
>> - jana
>>
>> On Wed, Dec 11, 2019 at 2:00 PM Mikkel Fahnøe Jørgensen <
>> [e-mail]> wrote:
>>
>>> Yes, datagrams are important for custom protocols.
>>>
>>>
>>> On 11 December 2019 at 22.49.23, Victor Vasiliev (
>>> vasilvv=[e-mail]) wrote:
>>>
>>> I support adoption of this draft.  We already support a version of this
>>> in gQUIC and I am willing to work on making sure it's interoperable with
>>> other implementations.
>>>
>>> On Wed, Dec 11, 2019, 13:36 Mark Nottingham <mnot@mnot...net
>>> <[e-mail]>> wrote:
>>>
>>>> This is a Call for Adoption of the following document:
>>>>   https://tools.ietf.org/html/draft-pauly-quic-datagram-05
>>>>
>>>> The Working Group has discussed this document for some time, most
>>>> recently in Singapore.
>>>>
>>>> Please state whether you support adoption of the document by the
>>>> Working Group, and any additional information that you feel may be helpful.
>>>>
>>>> This call will end on 15 January 2020.
>>>>
>>>> Regards,
>>>>
>>>> --
>>>> Mark Nottingham   https://www.mnot.net/
>>>>
>>>>

-- 
Kazuho Oku