Re: [tsvwg] L4S operational guidance draft

[Sebastian Moeller <[e-mail]>]

Hi Ingemar,



On 18 November 2020 07:44:04 CET, Ingemar Johansson S <ingemar.s.johansson=[e-mail]> wrote:
>Hi Jake and Pete + others
>
>I try to understand this input, thanks for raining this BTW. 
>
>I interpret this as: 
>1) Flow queue AQMs don't work properly across tunnels, as flow
>isolation based om 5-tuple does not work


         Another, less negative, interpretation is that tunnels, that multiplex multiple flows into one externally visible flow are treated and recognized only as that flow. Which for things like encrypted VPNs is exactly the behavior that seems appropriate, no?

Best Regards
          Sebastian

>2) L4S flows can potentially (or will) get an unfair share as a
>consequence of #1 above because the bootleneck only implement a RFC3168
>ECN marking. 
>
>Is this understanding correct ?
>
>/Ingemar
>
>> -----Original Message-----
>> From: tsvwg <[e-mail]> On Behalf Of Holland, Jake
>> Sent: den 18 november 2020 03:21
>> To: Pete Heist <[e-mail]>; Wesley Eddy <[e-mail]>
>> Cc: [e-mail]
>> Subject: Re: [tsvwg] L4S operational guidance draft
>> 
>> Hi Pete,
>> 
>> On 11/17/20, 12:04 PM, "Pete Heist" <[e-mail]> wrote:
>> > One addition to this is the fact that when tunneled traffic
>traverses
>> > fq_codel, since the 5-tuple for all of the tunnel's inner flows is
>the
>> > same, that places them in one fq_codel queue, leading to the same
>> > safety problem when L4S and non-L4S flows meet. So, while single
>queue
>> > RFC3168 AQMs or hash collisions are one way for that to happen,
>> > tunneled traffic may be a more likely way. We recently added a test
>> > scenario using Wireguard through fq_codel to cover that case, which
>> > for some reason I didn't think to try earlier:
>> >
>> > https://protect2.fireeye.com/v1/url?k=aaf3845d-f568bd70-aaf3c4c6-
>> 8692d
>> > c8284cb-99cc9ade2897f933&q=1&e=b25cd448-5648-4fc4-8858-
>> a85d6c2df9b0&u=
>> > https%3A%2F%2Fgithub.com%2Fheistp%2Fl4s-tests%2F%23unsafety-in-
>> tunnels
>> > -through-rfc3168-bottlenecks
>> >
>> > That should probably be covered in the guidance in relation to FQ
>> > classic ECN as well...
>> 
>> Thanks Pete, I 100% agree.  I knew I was forgetting something.
>> 
>> I meant to include exactly that sentiment and link, and I thought it
>was a very
>> helpful point when you posted it a few weeks ago, but several hundred
>> messages later it slipped my mind while writing this up.  Thanks for
>getting it in
>> there.
>> 
>> Best regards,
>> Jake
>> 

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.