Re: [patch] delayed acks decrease performances in congestioned networks

Andi Kleen (ak@muc.de)
Tue, 17 Nov 1998 18:38:00 +0100

On Tue, Nov 17, 1998 at 05:48:55PM +0100, Olaf Titz wrote:
> > setsockopt(sk, SOL_TCP, TCP_NODELAY, ((int *){1}), sizeof int);
>
> I think this is Nagle not delayed ack.

You are right, sorry for the confusion. Reasons against it:

- People will enable it which will lead to more congestion because of
more packets.
- A global sysctl is not fine grained enough, if you need it make it at
least a per route attribute (like mss or window)

Because of the first reason I would prefer to not implement it.

-A.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/