Subject: Re: Some discussion of NewReno implementation issues
To: Jason Thorpe <>
From: Charles M. Hannum <>
List: tech-net
Date: 01/27/2005 01:33:16
On Thursday 27 January 2005 01:05, Jason Thorpe wrote:
> On Jan 26, 2005, at 4:29 PM, Charles M. Hannum wrote:
> > As such, given that the performance appears to have been brought back
> > in line
> > with Reno, I am not inclined to touch it again unless I see hard
> > numbers to
> > demonstrate that another variant is better.
> What about enabling the fixed NewReno as the default?

As it turns out:

1.164        (tls      22-Apr-04): int  tcp_do_newreno = 1;     /* Use the New 
Reno algorithms */

This was apparently the reason for a 16% drop in my wireless throughput a 
while back, which I had misattributed to switching wireless drivers.

I am rather shocked that it was enabled when it was known to be broken.