Re: [tcpm] increasing max window size of TCP

[Yoshifumi Nishida <[e-mail]>]

Hi Richard,

On Fri, Mar 13, 2015 at 3:16 AM, Scheffenegger, Richard <[e-mail]>
wrote:

>  Hi Yoshi,
>
>
>
> multiple times
>
> s/archive/achieve/
>

Ugh.. Thanks for pointing this out. We'll fix them.

>
>
> One comment: A receiver that sees WS > 14 is expected to clamp WS at 14;
> (see the paragraph in RFC7323 right after the one that you quoted).
>
>
>
> So, if you signal to a regular RFC7323 receiver a WS of 15, if will behave
> as if WS 14 was sent, limiting its maximum sending windowsize, right? No
> harm done except that session getting less bandwidth.
>

Right.
But, I personally would like to not discuss clamping cases in the draft if
possible. We would like to avoid this situation rather than allowing it.

>
>
> A receiver also implementing your proposal could, however, make use of the
> nearly twice as large windowsize. Monitoring of the evolution of the
> receive window signaled vs. data transmitted could allow a sender to
> determine passively, if the receiver actually supports the larger
> windowscale option. But I’m not sure if this is really necessary… (the
> signaled window would shrink half as fast for unprocessed data in the
> receive buffer, compared to WS14; however, the step change in the window
> would only happen when there are 16 kB (or 32kB-1) bytes of data
> unprocessed.
>

Yes. I'm not very sure if this is necessary, either. (as I would like to
avoid this)
But, I think it is possible.


>
> But the sender must stop sending when the receive window is signaled to be
> zero (except for retransmissions) either way.
>
>
>
> Also, I guess you want to clamp every other received value than 15 down to
> 14 still… Not that you start scaling by (2^255 – 1).
>
>
>
> Since you are effectively doing a special case handling with WS==15, you
> may want to describe the necessary operations to an implementer alike
> section 2.3 of RFC7323, that is, instead of only a simple left shift by 15
> bits(multiply by 2^15), you need to multiply by (2^15-1) if I understand
> your proposal correctly. (Or course, a multiplication by 2^15-1 is a left
> shift by 15, followed by a subtraction of the original value; perhaps gcc
> even optimizes this multiplication nowadays J )
>

Hmm. I've checked 7323, but I couldn't find this formula. I think I missed
something. Could you point it out?


> BTW, I think the formula you give in the last sentence of sec  1 is off;
> the signaled window spans 2^16, but the maximum value of the window
> signaled is 2^16-1; times (2^15-1) gives a maximum window of 2147385345
>

>
> I don’t think an additional option is necessary for this change, and
> provided I understood your intention and calculations correctly (as
> described above), I would support this going forward as experimental.
>

Great! Thank for the feedback!

Regards,
--
Yoshi

>
>
>
>
>
> *From:* tcpm [mailto:[e-mail]] *On Behalf Of *Yoshifumi
> Nishida
> *Sent:* Mittwoch, 11. März 2015 07:35
> *To:* [e-mail]
> *Cc:* [e-mail]
> *Subject:* [tcpm] increasing max window size of TCP
>
>
>
> Hi,
>
>
>
> We have submitted a short draft which proposes to increase max window size
> of TCP.
>
> If you're interested, please take a look at the following info.
>
> Some may say this is a minor thing, but we still believe we need to
> consider this for future extensions of TCP.
>
>
>
> We will appreciate your feedback!
>
> --
>
> Yoshi & Hiro
>
>
>
>
>
>
>
> -------------------------------------------------------------------------
>
> From: <[e-mail]>
> Date: Mon, Mar 9, 2015 at 1:09 PM
> Subject: I-D Action: draft-nishida-tcpm-maxwin-00.txt
> To: [e-mail]
>
>
>
> A New Internet-Draft is available from the on-line Internet-Drafts
> directories.
>
>
>         Title           : Increasing Maximum Window Size of TCP
>         Authors         : Yoshifumi Nishida
>                           Hirochika Asai
>         Filename        : draft-nishida-tcpm-maxwin-00.txt
>         Pages           : 6
>         Date            : 2015-03-09
>
> Abstract:
>    This document proposes to increase the current max window size
>    allowed in TCP.  It describes the current logic that limits the max
>    window size and provides a rationale to relax the limitation as well
>    as the negotiation mechanism to enable this feature safely.
>
>
> The IETF datatracker status page for this draft is:
> https://datatracker.ietf.org/doc/draft-nishida-tcpm-maxwin/
>
> There's also a htmlized version available at:
> http://tools.ietf.org/html/draft-nishida-tcpm-maxwin-00
>
>
> Please note that it may take a couple of minutes from the time of
> submission
> until the htmlized version and diff are available at tools.ietf.org.
>
> Internet-Drafts are also available by anonymous FTP at:
> ftp://ftp.ietf.org/internet-drafts/
>
> _______________________________________________
> I-D-Announce mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/i-d-announce
> Internet-Draft
> <https://www.ietf.org/mailman/listinfo/i-d-announceInternet-Draft>
> directories: http://www.ietf.org/shadow.html
> or ftp://ftp.ietf.org/ietf/1shadow-sites.txt
>
>
>