[ale] sshd resource intensive??

[Michael D. Hirsch]

On Sunday 16 May 2004 09:48 am, Geoffrey wrote:
> Drag0n wrote:
> > Considering that the sender has to encrypt the data on the fly and the
> > receiver has to decrypt it as it receives it, processor speed more than
> > bandwidth determines throughput on local networks. This is to be
> > expected unless you have dedicated ssl accelerators in each machine that
> > ssh has been configured to use.
> >
> > Drag0n
> > dragon at atlantacon.org
> >
> > On Sat, 2004-05-15 at 22:31, Geoffrey wrote:
> >>I have never put any thought to how resource intensive scp might be.
> >>Until now.  I was setting up a couple of boxes and testing file transfer
> >>rates.  I've noted that for a couple of older boxes (p200 and PII 350)
> >>it's much slower.  I then fired up top while the transfer was going
> >>along and noted sshd taking up quite a bit of cpu.
>
> I surely expected some degradation in transmission because of the ssh
> overhead, but certainly not 8000%.  11 seconds to transfer a 125M file
> from an athlon 2400+ to pentium IV 2.4G, verses 92 seconds to transfer
> the same file from the athlon to a PII 350.
>
> You learn something new every day..

Looks like you're on a 100Mbit network and your transfer has maxed out the 
network.  Since there is a factor of eight slowdown with ssh (800%, not 
8000%) you can solve the problem by switching to a 10MBit network.  Both 
transfers should take about the same time on that network.  :-)

Seriously, I found that over a T1, ssh  would take up a couple of percent of 
the CPU.  (This was a couple of years ago--it should be less with a modern 
CPU.)  That does make the numbers you're getting a little higher than I would 
expect.  At 1% of a CPU for T1 speed, I'd expect 10% at 10-baseT and 100% for 
100-baseT.  Seems like our numbers disagree by a factor of 10.

Michael