Re: [Webtransport] Confirming Consensus on WebTransport protocols

["Morten V. Pedersen" <[e-mail]>]

I'm sorry for being uninformed here. But does this preclude unreliable 
datagram support from an application point of view. I.e. will users of 
Webtransport be able to send unreliable datagrams when using 
Webtransport over HTTP/3?

All the best,
Morten

On 12/01/2021 19.28, Luke Curley wrote:
> Yeah, I want to reaffirm that my vote for HTTP/3 was tied to the 
> removal of pooling. The current Http3Transport draft is more complex 
> than it needs to be in order to support pooling, which is a fringe 
> optimization at best.
>
> On Tue, Jan 12, 2021 at 10:15 AM Eric Kinnear 
> <ekinnear=[e-mail] 
> <mailto:[e-mail]>> wrote:
>
>     I support this consensus.
>
>     To Alan’s point, I’d expect that we’ll continue to discuss
>     pooling, but that we’re not committing to delivering something
>     that supports pooling without further identifying the complexities
>     involved and designing something which sufficiently resolves those
>     concerns. (As would be reasonable for any protocol feature.)
>
>     Thanks,
>     Eric
>
>
>>     On Jan 12, 2021, at 10:10 AM, Alan Frindell
>>     <afrind=[e-mail]
>>     <mailto:afrind=[e-mail]>> wrote:
>>
>>     I support the consensus on Question 1 and Question2.
>>
>>     > Additionally, we discussed pooling and decided that we would
>>     not allow pooling in WebTransport over HTTP/3 for now, due
>>     to the complexities of pooling.
>>     I think this statement is worded too strongly.  We discussed
>>     pooling and made it clear that choosing H3 does not mandate that
>>     we will support pooling. My understanding is that it will
>>     continue to be discussed in the working group.
>>     -Alan
>>     *From:*Webtransport <[e-mail]
>>     <mailto:[e-mail]>> on behalf of David
>>     Schinazi <[e-mail] <mailto:[e-mail]>>
>>     *Date:*Tuesday, January 12, 2021 at 9:26 AM
>>     *To:*WebTransport <[e-mail]
>>     <mailto:[e-mail]>>
>>     *Subject:*[Webtransport] Confirming Consensus on WebTransport
>>     protocols
>>     Hi WebTransport enthusiasts,
>>     Today we had an interim of the IETF WEBTRANS WG,
>>     where we reached consensus in the (virtual) room on
>>     the following questions:
>>     Question 1: number of protocols
>>       Should the working group adopt only one UDP-based transport?
>>         1A: only one transport (QUIC or HTTP/3)
>>         1B: multiple transports (QUIC and HTTP/3)
>>     The consensus was option 1A.
>>     Additionally, we noted that this does not preclude us from
>>     building a second protocol at a later date if new information
>>     emerges.
>>     Question 2: UDP-based protocols
>>       Which UDP-based option should we adopt as a starting
>>       point for WebTransport protocol?
>>         2A: WebTransport over HTTP/3
>>         2B: WebTransport over QUIC directly (separate ALPN)
>>     The consensus was option 2A.
>>     Additionally, we discussed pooling and decided that we would
>>     not allow pooling in WebTransport over HTTP/3 for now, due
>>     to the complexities of pooling.
>>     We'd like to confirm this consensus on the list. If you disagree
>>     with these points, please reply on the list. If you do so, please
>>     state how strongly you feel - whether you are expressing a
>>     preference, or an imperative. Please also provide as much
>>     detail as possible to explain your position. Please also respond
>>     if you support this consensus - a simple "I support this consensus"
>>     email to the list would be helpful.
>>     We'll be running this consensus call for two weeks, please reply
>>     before 2021-01-26, the chairs will determine consensus on that date.
>>     Thanks,
>>     David
>>     --
>>     Webtransport mailing list
>>     [e-mail] <mailto:[e-mail]>
>>     https://www.ietf.org/mailman/listinfo/webtransport
>>     <https://www.ietf.org/mailman/listinfo/webtransport>
>
>     -- 
>     Webtransport mailing list
>     [e-mail] <mailto:[e-mail]>
>     https://www.ietf.org/mailman/listinfo/webtransport
>     <https://www.ietf.org/mailman/listinfo/webtransport>
>
>