TightVNC compression vs. SSH compression

[David Brodbeck]

> From: John Roland Elliott [mailto:John_Roland_Elliott "at" Hotmail.com]
> Sent: Thursday, November 08, 2001 12:05 PM
> To: vnc-list "at" uk.research.att.com
> Subject: TightVNC compression vs. SSH compression

> Can an SSH data stream even benefit from compression? Might it be the case
> that the SSH compression algorithms introduce so much "randomness" into
the
> data stream that it no longer compresses effectively? (This is probably an
> academic question since the compression is surely done before the
encryption
> in all cases anyway, right?)

Encrypted data compresses poorly, so ssh compresses the data before
encrypting it.  (This also makes the encryption stronger, by making the 
encrypted data more random.)

> Is there reason to believe that the compression embodied it TightVNC is
> dramatically more effective than the compression implemented in SSH?

Yes.  TightVNC has an advantage over SSH because it knows what's being
compressed, and can optimize things a bit more.  In particular, the
"local cursor" feature of TightVNC makes a huge difference in usability
on slow connections.  Also, TightVNC can use JPEG compression...as a lossy
algorithm this can be a bit more effective than ssh's compression, which has
to be non-lossy.

> I wonder where the breakpoint is ... would a 166MHz Pentium with a cable
> modem connection be better off with or without compression? How 'bout a
500
> MHz Pentium III with an ISDN connection?

All I can tell you is "it depends."  My boss has a cable modem connection
with a good path to our office, and usually just uses Hextile encoding.
My DSL provider, on the other hand, is many hops away from our office
network-wise, and there tend to be bottlenecks and latency problems,
so I find Tight encoding much more usable.

One thing I've found is that if you're already using Tight encoding,
adding SSH compression on top of it mostly just increases CPU load
and latency.  The Tight-encoded stream seems to be too compressed
already for SSH to do much with.
---------------------------------------------------------------------
To unsubscribe, mail majordomo "at" uk.research.att.com with the line:
'unsubscribe vnc-list' in the message BODY
See also: http://www.uk.research.att.com/vnc/intouch.html
---------------------------------------------------------------------