To: cygwin AT cygwin DOT com
From: "Cary Jamison" <Cary_Jamison AT Symantec DOT com>
Subject:  Re: How to improve scp speed?
Date:  Thu, 1 Dec 2005 13:40:47 -0700
Lines: 28
Message-ID: <dmnn4i$ogg$1@sea.gmane.org>
References:  <61f6f4390511281238o1e512ef8i14647828fcf32c AT mail DOT gmail DOT com> <BAY105-F2693514CAA56EEF86ECAC3E9480 AT phx DOT gbl> <dmii0m$s50$1 AT sea DOT gmane DOT org> <438D6EE0 DOT 1000406 AT gmx DOT de> <dml03k$6gq$1 AT sea DOT gmane DOT org> <438E30CB DOT 4327D914 AT dessent DOT net>
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com

In news:438E30CB DOT 4327D914 AT dessent DOT net,
Brian Dessent typed:
> Cary Jamison wrote:
>
>> I think we all know that encryption is a factor.  The site still
>> states :
>
> No, I wouldn't say that we all know that.  I just ran "openssl speed"
> and on my very modest Athlon XP 1700 machine and both the aes-128 and
> blowfish ciphers clocked in at approximately 60-65 megabytes/sec
> throughput.  That's about 50 times faster than the theoretical maximum
> throughput of 10Base-T ethernet.  Even this modest system can encrypt
> 5 times faster than the absolute maximum rate of 100 megabit ethernet.
>
> I think you are vastly overestimating the CPU requirement of
> encryption.  And it is also why the original poster should not waste
> his time trying to find a nonexistent, useless, and insecure option to
> disable encryption.

Ok, I probably am.  But, you wouldn't expect the buffer tuning to make such 
a big difference, either, since a fast processor can surely move data around 
in memory several times faster than it can move it over the net.  All these 
things add up, though....


Cary




--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/