[Mapserver-users] Large Raster files

[Martin Weinelt]

On Monday 17 May 2004 17:09, Jan Hartmann wrote:
> Martin Weinelt wrote:
> > I am not Ed  ;-) but handling DEM-hillshades in a current project which
> > can also get quite big. What I am doing reads almost exactly like the
> > procedure decribed by Jan. I stay away from any on-the-fly projection and
> > do as much 'preprocessing' as can be done. I also change RGB to indexed
> > colors, which gives almost the same impression IMHO.
>
> Yes, both preprocessing tasks (projecting and converting RGB to indexed
> colors) are almost mandatory for production servers
>
> > Everything is done by GDAL and the utilities. I decided to have tiles
> > about the size of the map, without knowing exactly why this would be
> > better then 2000px (Jan).
>
> Nothing holy about the number. It's just that multiplying and dividing
> by 2 is so easy with a computer :-). I chose 2000, because a single
> request should be as much confined to one tile, but it should not have
> to resample the image too much. Most map images are about 500 by 500
> pixels, so there is a good chance that it is completely within one tile.

Sometimes I am facing white stripes (1px) at the margings of the tiles. I 
guess I have to tune the tiling in respect to the geocoded extent /pixel size 
ot the image

> > I also think that compressing the rasters is even contraproductive,
> > because 'in memory' there is no compression and you add an extra job
> > (decoding) to the application.
>
> That's what I thought too, but serving compressed TIF files is not much
> faster than uncompressed ones, at least in my experience. Some of my
> rasters have a very high rate of compression (5 to 20 times), so I would
> run into disk problems with this procedure. Reason is (I guess) that
> regions can be extracted from compressed tifs without decompressing the
> whole file, so the actual time lost to decompressing the region is
> marginal.

That is a point I never thought of. Must check.

> > I keeping the shapefiles in different resolutions aswell, making my map
> > file a pain (number of MAXSCALE-MINSCALE-layers for the same data). I
> > will look deeper into the new DATAPATTERN and variable substitution
> > mechanisms soon.
>
> Additionally, I have been thinking about generalizing shapefiles to
> different levels with PostGIS. IMO generalizing is one of the
> fundamental problems of WebMapping.

Absolutely. And in the end that is what causes much work in setting up a nice 
and fast server.

Martin

-- 
 --- Martin Weinelt 
 --- kk+w - digital cartography 
 --- Kiel, Germany
 --- Tel: +49.431.5791165
 --- http://www.planiglobe.com