VNC optimized to almost 30% bandwidth raw (40% zlib)

Const Kaplinsky const "at" ce.cctpu.edu.ru
Mon, 28 Aug 2000 21:59:52 +0000


>>>>> "RDV" == Rudi De Vos <Rudi.De.Vos "at" netbuilding.simac.be> writes:

RDV> The function GetChangedRegoin (vncbuffer) looks for a changed
RDV> line in a rect and then copy the full rectangle to the update
RDV> buffer. I changed the code so that only the changed rectangle is
RDV> passed for update and not the full.

RDV> I put the exucutables and source on
RDV> http://services.simac.be/vnc/single.html vncexe28aug.zip
RDV> vncsrc28aug.zip

RDV> ++ Is there a reason to use 32x32 rectangles and not a variable
RDV>     (8X32, 17x32 ..)

RDV> ++ I tested the code using zlib in debug mode on a single
RDV>     application (internet explorer showing a running text in java)

How precisely did you evaluate the bandwidth drop? You should be able
to run original and modified VNC servers for the same VNC session. Not
for the similar ones, but for the same. How did you do this?

And also note that such optimizations are interesting only on
low-bandwidth connections, but you seem to test things locally.
However, the situation differs on slow connections: there are much
fewer screen updates and rectangle sizes are usually larger.

And one more thing I do not understand: what are those "40% zlib"?
Zlib compresses data, and more similar data means higher compression
ratios. You get less data and at the same time it compresses much
better? It's strange (at least).

-- 
With Best Wishes,
Constantin
---------------------------------------------------------------------
To unsubscribe, send a message with the line: unsubscribe vnc-list
to majordomo "at" uk.research.att.com
See also: http://www.uk.research.att.com/vnc/intouch.html
---------------------------------------------------------------------