Re: [tsvwg] L4S drafts: Next Steps

[Bob Briscoe <[e-mail]>]

David,

On 10/03/2021 18:01, Black, David wrote:
>
> Bob,
>
> Reproducing text from the slide used in today's meeting:
>
> ------------------------
>
> _L4S Drafts: What Needs To Be Done_
>
> __
>
> 1. Transport-protocol-independent requirements to use L4S service
>
> – Goal: Interoperability – multiple transport protocols able to meet 
> requirements
>
> • Specifications and/or implementations – WG decides what is sufficient
>
> – Location: draft-ietf-tsvwg-ecn-l4s-id
>
> 2. Safety of Internet-wide L4S experiment
>
> – Goal: L4S experiment unlikely to cause significant damage to the 
> Internet, and
>
> • Any damage that results is expected to be tolerable
>
> – Location: draft-white-tsvwg-l4s-ops (candidate for WG adoption)
>
> • WG Last Call on l4s-ops draft not required to meet safety goal
>
> • Criteria: WG “rough consensus” that each goal has been reached
>
> – WG chairs will run initial consensus calls before next IETF meeting week
>
> ------------------------
>
> This discussion is about the first goal: 
> Transport-protocol-independent requirements to use L4S service
>
> > I accept that we're messing with core protocols here, so the bar 
> should be higher.
>
> > But, 2 independent full implementations for an EXP would be way 
> beyond any interpretation of IETF process.
>
> Two independent full implementations of protocols that meet the 
> requirements ought to suffice, but may not be necessary.
>
> As stated on the slide (and quoted above): " Specifications and/or 
> implementations – WG decides what is sufficient."
>

[BB] I think the number 'two' comes from wanting to be sure that 
developers other than the author-team can understand the spec. 
sufficiently to implement.

Pls confirm that could also be satisfied by /each/ requirement having 
been implemented by /at least one other/ team than the authors (or 
reasonably expected to implement).

The reason this is important is that there is a heavy performance 
evaluation burden attached to each requirement, because they are nearly 
all about impact on the queuing delay of other hosts (or impact on 
throughput of others in the case of the Classic ECN AQM one).

For a second CC developer to invest the time and effort in the 
performance testing necessary for every requirement, they would have to 
be very confident that the IETF was going to follow through with 
codepoint assignment. A great job has been done in undermining that 
confidence in the last couple of years.

To be clear, there are many implementations of the network part, and a 
load of network operators ready to deploy the network part. There's no 
problem there (other than all those projects have gone cold waiting for 
the IETF). The focus here is CC development, I assume.

> > This part of your original email (which was as a chair) asked us to 
> /restructure/ the drafts. I said they already have the structure you want.
>
> > No reply to that point. Now what you talk about above (albeit as an 
> individual now) are wording changes not restructuring.
>
> > We've been doing wording changes with the survey of implementers. 
> Fine with that.
> >
> > But I'd like confirmation (as a chair) that restructuring is not needed.
>
> Confirmed.
>

Good. Thank you.


Bob
>
> Thanks, --David (as a TSVWG chair)
>
> *From:* Bob Briscoe <[e-mail]>
> *Sent:* Wednesday, March 10, 2021 7:53 AM
> *To:* Black, David; [e-mail]
> *Subject:* Re: [tsvwg] L4S drafts: Next Steps
>
> [EXTERNAL EMAIL]
>
> David,
>
> As you've just pointed this old thread out, I'll continue on this 
> thread. See [BB] inline...
>
> On 09/12/2020 22:28, Black, David wrote:
>
>     Bob,
>
>     Reminder of the goal for [1] (Rework the “Prague Requirements”
>     into the following two entities ...):
>
>          >> Success of this exercise will be judged by the WG reaching
>     “rough consensus” that multiple transport protocol implementations
>          >> currently meet and/or are reasonably expected to be able
>     to meet the normative requirements (a).
>
>     If the L4S folks believe that this has already been done, then
>     it’s time to start bringing forward the multiple transport
>     protocols that demonstrate success of that exercise.
>
>
> [BB] This is understood and fine. This is where our focus is.
>
> I assume readers are aware that the IETF's requirement for 2 
> independent implementations is for progressions from Proposed Standard 
> to full Internet Standard, and few RFCs have ever taken that step 
> (only 118 RFCs are STDs or parts of STDs) (e.g. RFC3168 is not).
> https://www.rfc-editor.org/search/rfc_search_detail.php?sortkey=Number&sorting=DESC&page=All&pubstatus%5B%5D=Standards%20Track&std_trk=Internet%20Standard 
> <https://www.rfc-editor.org/search/rfc_search_detail.php?sortkey=Number&sorting=DESC&page=All&pubstatus%5B%5D=Standards%20Track&std_trk=Internet%20Standard>
>
> I accept that we're messing with core protocols here, so the bar 
> should be higher. But, 2 independent full implementations for an EXP 
> would be way beyond any interpretation of IETF process.
>
>
>
>     >Nonetheless, if there are specific items in S.4 that someone
>     believes should or shouldn't be there, or should be worded
>     differently, let's have those discussions now on the list - I'm
>     all ears
>
>     Speaking only for myself as an individual beyond this point, these
>     two normative requirements struck me as ones that might be better
>     stated as design and implementation guidelines:
>
>       
>
>         o  A scalable congestion control MUST eliminate RTT bias as much as
>
>            possible in the range between the minimum likely RTT and typical
>
>            RTTs expected in the intended deployment scenario (see
>
>            Appendix A.1.5  <https://datatracker.ietf.org/doc/html/draft-ietf-tsvwg-ecn-l4s-id-12#appendix-A.1.5>  for rationale).
>
>       
>
>         o  A scalable congestion control SHOULD remain responsive to
>
>            congestion when typical RTTs over the public Internet are
>
>            significantly smaller because they are no longer inflated by
>
>            queuing delay (seeAppendix A.1.6  <https://datatracker.ietf.org/doc/html/draft-ietf-tsvwg-ecn-l4s-id-12#appendix-A.1.6>  for rationale).
>
>     The use of “as much as possible” in the first bullet makes it hard
>     to figure out whether a specific congestion control meets that
>     mandatory requirement for participation in the L4S experiment. 
>     The second bullet appears to involve a prediction of the future as
>     currently written.
>
>
> [BB] I'll refrain from getting into text-specifics in this thread, 
> 'cos I'm trying to be absolutely sure I understand what the chairs 
> want. We've felt subjected to a "No, bring me a different rock" 
> process so far, so we want to be absolutely clear what sort of rock 
> you want.
>
> This part of your original email (which was as a chair) asked us to 
> /restructure/ the drafts. I said they already have the structure you 
> want. No reply to that point. Now what you talk about above (albeit as 
> an individual now) are wording changes not restructuring. We've been 
> doing wording changes with the survey of implementers. Fine with that.
>
> But I'd like confirmation (as a chair) that restructuring is not needed.
>
> ==============================================================================================
> FYI, here's the structure of the transport requirements (I already 
> explained this in the previous email in this thread):
>
> ecn-l4s-id
> * Sec.4. (Normative) Transport Layer Behaviour
>     All the factors that involve potential harm to others (containing 
> MUSTs and now some SHOULDs)
>     With a pointer from each one to further guidance in Appx A.1
> * Sec 5 (Normative) Network Node Behaviour
> ...
> * Appx A.1 (Informative) Requirements for Scalable Transport Protocols
>     Guidance on features that involve potential harm to others
> * Appx A.2 (Informative) Scalable Transport Protocol Optimizations
>    Guidance on features that would benefit the flow itself
>
> draft-briscoe-iccrg-prague-congestion-control (just posted)
> * Detailed guidance, specification, implementation notes, etc.
>
>
>
> Bob
>
>
>     Thanks, --David
>
>     *From:* Bob Briscoe <[e-mail]>
>     <mailto:[e-mail]>
>     *Sent:* Tuesday, December 8, 2020 6:01 PM
>     *To:* Black, David; [e-mail] <mailto:[e-mail]>
>     *Subject:* Re: [tsvwg] L4S drafts: Next Steps
>
>     [EXTERNAL EMAIL]
>
>     David and chairs,
>
>     On 04/12/2020 15:54, Black, David wrote:
>
>         The WG chairs have consulted among ourselves and with our AD
>         (Martin Duke).
>
>         Our guidance to the WG and the L4S draft authors is that at
>         least the following two things need to be done before WGLC
>         will become possible on the L4S drafts:
>
>         [1] Rework the “Prague Requirements” into the following two
>         entities, which will overlap:
>
>         1.Transport-protocol-agnostic normative requirements for all
>         congestion-controlled traffic that uses the L4S low latency queue.
>
>         1.The DualQ AQM design relies upon traffic adhering to these
>         requirements.
>
>         2.Design and implementation guidelines for new congestion
>         controls that meet the normative requirements (a).
>
>         1.Ultimately, the choice of whether a particular aspect is a
>         guideline (b) or requirement (a) is a WG “rough consensus”
>         decision.
>
>
>     [BB] I believe draft-ietf-tsvwg-ecn-l4s-id is already structured
>     for this division between normative requirements and design and
>     implementation guidelines.
>     * [1]a) is Section 4 "Prerequisite Transport Layer Behaviour".
>     Nearly every paragraph also refers off to subsections of Appendix
>     A.1 for extra non-normative info, headed "Requirements for
>     Scalable Transport Protocols". [Actually, it might be useful to
>     swap these two headings]
>     * [1]b) is Appendix A.2: "Scalable Transport Protocol Optimizations"
>
>     The rule that determined which aspect went in which was:
>     [1]a) if it's about an aspect of one transport's behaviour that
>     has potential to impact/harm others
>     [1]b) if it's solely about how a transport can improve it's own
>     performance.
>     They have been divided like that from the very first ad hoc
>     meeting in Prague that formulated them (in 2015)
>
>     Code to address all the requirements and most of the optimizations
>     has been implemented. Over the last year, 3 or 4 of the later
>     requirements in a) have become SHOULDs not MUSTs. Personally I
>     would have left most as MUSTs, but I've tried to reflect what the
>     WG wanted. Those demoted to SHOULD have generally been considered
>     a lower risk of harm (either low likelihood of occuring or minor
>     harm if they do) - taking into account the complexity of
>     implementing them.
>
>     We could certainly flag the point where the transition from MUSTs
>     to SHOULDs occurs within section 4. But I think we should still
>     group all the items with potential for harm together in the same
>     section. Because assessment of risk will change as the Internet
>     landscape changes (possibly over the duration of the experiment).
>
>     Now I've explained, I hope the chairs will all agree that "rework
>     the Prague Requirements into two entities" (normative requirements
>     and design and implementation guidelines) is already done and dusted.
>
>     Nonetheless, if there are specific items in S.4 that someone
>     believes should or shouldn't be there, or should be worded
>     differently, let's have those discussions now on the list - I'm
>     all ears.
>
>
>
>     Bob
>
>
>
>              Success of this exercise will be judged by the WG
>         reaching “rough consensus” that multiple transport protocol
>         implementations
>              currently meet and/or are reasonably expected to be able
>         to meet the normative requirements (a).
>
>         [2] Build the safety case for L4S experimental Internet-wide
>         deployment in the L4S Ops draft.
>
>         1.This safety case does not rely on endpoints running TCP
>         Prague, though it can assume endpoints are meeting the
>         reworked Prague requirements
>
>         2.Ops draft will provide guidance on how to detect L4S
>         problems on their RFC 3168 network, and how to mitigate them.
>
>         3.Ops draft must consider these scenarios:
>
>         1. An unsophisticated user purchases an L4S endpoint and runs
>         it on a service provider's single-queue RFC 3168 network.
>
>         2. L4S traffic from another domain enters an RFC 3168
>         single-queue network (e.g. a peer-to-peer application)
>
>              Success of this exercise will be judged by the WG
>         reaching “rough consensus” that deployment of the L4S experiment
>              is unlikely to cause significant damage to the Internet,
>         and that any damage that results is expected to be tolerable.
>
>         Thanks, --David [as TSVWG WG co-chair]
>
>         ----------------------------------------------------------------
>
>         David L. Black, Senior Distinguished Engineer
>
>         Dell Technologies, Infrastructure Systems Group
>
>         176 South St., Hopkinton, MA  01748
>
>         [phone] <tel:[phone]> Mobile: [phone]
>         <tel:[phone]>
>
>         [e-mail] <mailto:[e-mail]>
>
>         ----------------------------------------------------------------
>
>
>
>
>     -- 
>
>     ________________________________________________________________
>
>     Bob Briscoehttp://bobbriscoe.net/  <http://bobbriscoe.net/>
>
>
>
> -- 
> ________________________________________________________________
> Bob Briscoehttp://bobbriscoe.net/  <http://bobbriscoe.net/>

-- 
________________________________________________________________
Bob Briscoe                               http://bobbriscoe.net/