Re: [tcpm] TCP tuning

[Jerry Chu <[e-mail]>]

On Wed, Feb 3, 2010 at 12:51 PM, John Heffner <[e-mail]> wrote:
> On Wed, Feb 3, 2010 at 3:44 PM, Jerry Chu <[e-mail]> wrote:
>> Hi John,
>>
>> On Wed, Feb 3, 2010 at 12:32 PM, John Heffner <[e-mail]> wrote:
>>> A couple thoughts:
>>> - A large initial window should probably only be used by mutual
>>> agreement.  Maybe via a TCP option during the handshake?
>>
>> This may not be needed. A receiver can effectively constrain the
>> sender's window by advertising a smaller initial receive window.
>>
>> Obviously this works for only the initial window, not the restart window.
>> (Oh, on a second thought this can be made to work for the restart
>> window too.)
>
> A good point, though most receivers (Linux 2.4 and later being the
> exception) advertise all available window space, not expecting a large
> initial burst from the sender.  I think you need a negotiation for
> opt-in rather than opt-out.

Yes, Linux 2.4 and later constrain its initial receive window per RFC3390,
making it harder for any future change to initcwnd to be deployed (i.e., both
sides' stack need to change). We're trying to fix this. See
http://www.spinics.net/lists/netdev/msg115912.html.

The change will also be useful for those hosts behind slow links for purposes
mentioned before.

Jerry

>
>  -John
>