RE: The first octet

["Roni Even (A)" <[e-mail]>]

Hi,
I meant Montreal not London

From: Ian Swett [mailto:[e-mail]]
Sent: Wednesday, August 08, 2018 4:57 PM
To: Roni Even (A)
Cc: [e-mail]; IETF QUIC WG; Martin Thomson; Kazuho Oku
Subject: Re: The first octet

Thanks Roni, agreed, let's focus on the issues Martin raised.

On Wed, Aug 8, 2018 at 9:55 AM Roni Even (A) <[e-mail]<mailto:[e-mail]>> wrote:
Hi,
In London the spin bit and VEC were not discussed in the session, my take from the offline discussions we had was that currently we experimented with good results only the spin bit. It also seemed that the current VEC as proposed in Brians's draft is not good enough and we intend to look at other algorithms that will provide better results for packet loss using the 2 bits.
I think that this discussion is pre mature and that it should be discussed in the future
Roni

-----Original Message-----
From: QUIC [mailto:[e-mail]<mailto:[e-mail]>] On Behalf Of [e-mail]<mailto:[e-mail]>
Sent: Wednesday, August 08, 2018 4:44 PM
To: Ian Swett
Cc: IETF QUIC WG; Martin Thomson; Kazuho Oku
Subject: Re: The first octet

Sure:

  - spin+VEC are documented in draft-trammel-spin-03

  - 7 bits for packet loss were presented at
      http://snaggletooth.akam.ai/IETF-101-HotRFC/10-Ferrieux.pdf

Re "I'm curious why you'd even want them": please clarify which applies:

  (a) you recognize the criticality of loss location, but you have a cleverer algorithm to do the same with fewer than 7 bits

  (b) you don't believe network operators need to locate losses


On 08/08/18 15:29, Ian Swett wrote:
> I'm having a hard time understanding why you'd need 7 bits to measure
> packet
> loss(1 seems sufficient in most cases), can you send a link to the
> spin + VEC and 7 bit proposals you mentioned?
>
> I'll also note I think the chances the WG will agree to 7 bits is
> extremely close to 0, but I'm curious why you'd even want them.
>
> On Wed, Aug 8, 2018 at 9:22 AM <[e-mail]<mailto:[e-mail]>
> <mailto:[e-mail]<mailto:[e-mail]>>> wrote:
>
>     On 08/08/18 12:05, Martin Thomson wrote:
>      >
>      > I know you discussed taking more bits and even bytes, but the case
>      > hasn't been made in this working group to my knowledge.  I certainly
>      > don't think that's justified; it improved fidelity of measurements,
>      > but the spin bit appears to be pretty good at that without the extra
>      > overheads.
>
>     Fidelity is indeed the criterion to decide 1 (spin bit) vs 3 (spin bit + VEC)
>     for RTT measurements, right.
>
>     However, measurements also include packet loss, which is typically more
>     widespread than delay in big pipes (which don't have enough RAM to keep a
>     packet
>     for too long at line rate). And spin+VEC do not address packet loss at all,
>     hence our 7-bit spin counter proposal shown in London.
>
>     I admit that 7 is hefty if you're starting from Christian's songle spin bit ;
>     however:
>
>        - I assume nobody will emit any doubt about the criticality of
>     non-delay-related (big pipe) loss location. Unless everybody here has perfect
>     networks and ideal reporting, that is.
>
>        - Stick the spin bit on top of these 7, make that an extra easurement
>     byte and
>     you've got your three spare bits again in the first byte.
>
>
>
> ______________________________________________________________________
> ___________________________________________________
>
>     Ce message et ses pieces jointes peuvent contenir des informations
>     confidentielles ou privilegiees et ne doivent donc
>     pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu
>     ce message par erreur, veuillez le signaler
>     a l'expediteur et le detruire ainsi que les pieces jointes. Les messages
>     electroniques etant susceptibles d'alteration,
>     Orange decline toute responsabilite si ce message a ete altere, deforme ou
>     falsifie. Merci.
>
>     This message and its attachments may contain confidential or privileged
>     information that may be protected by law;
>     they should not be distributed, used or copied without authorisation.
>     If you have received this email in error, please notify the sender and
>     delete this message and its attachments.
>     As emails may be altered, Orange is not liable for messages that have been
>     modified, changed or falsified.
>     Thank you.
>



_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration, Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law; they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.