Re: Speed of ssh for large transfers
From: Brian Pence (bpenceNOSPAM_at_mindspring.com)
Date: 05/14/03
- Next message: Brian Pence: "Re: putty and telnet proxy"
- Previous message: Joss: "Re: how to pop3 through openssh"
- In reply to: Ted Spradley: "Re: Speed of ssh for large transfers"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Wed, 14 May 2003 09:47:15 -0500
>
> > /dev/zero is probably not a good example of data that you'd
> > expect to find in real life, either, as it is just a sequence of
> > null characters. This would tend to be highly compressible.
> >
> > Real world results would vary depending on the type of data.
> > 100Meg of MP3 files would take longer than 100Meg of text, etc..
>
> Why? Sure, if there were compression then ssh with compression would be
> much faster than rsh without. But there is no compression, according to
> Mr. Dunham, and according to the actual results, where ssh got 60% of
> the throughput of rsh, copying /dev/zero.
>
> I thought I had seen something recently that suggested that ssh doesn't
> transfer at full TCP speed, but does some application-layer error
> checking and handshaking, and that causes it to be a bit slower. I was
> hoping someone here could confirm or deny.
There *is* overhead added by the SSH protocol. If compression is enabled,
this
can compensate somewhat, but it is dependant on a lot of things including
processor speed and the data itself.
- Next message: Brian Pence: "Re: putty and telnet proxy"
- Previous message: Joss: "Re: how to pop3 through openssh"
- In reply to: Ted Spradley: "Re: Speed of ssh for large transfers"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
