[Mapserver-users] Map broker?

[Krzysztof Chodak]

Why do you meen by having "tiled" data? Could you write little more 
about it?

>There are many of us that have many GBs of data online under a single 
>mapserver. The whole US in Tiger/Line or GDT data requires about 8-
>20GB of shapefiles. I am working on a project to bring 19-20 TB of 
>image data online using mapserver.
>
>The performance has little to do with the total amount of data but 
>rather how well you have you data indexed and tiled so mapserver only 
>need to look at a small amount of the total data for any given page 
>view.
>
>Try running shptree on all your shapefiles.
>
>-Steve W.
>
>On 17 Mar 2003 at 13:55, Krzysztof Chodak wrote:
>
>  
>
>>Has anyone of You any experience with MapServer and large amount of
>>data. By large amount of data I take approximately 300MB of
>>shapefiles. I think that MapServer's current architecture is not
>>suitable for this use (when taking into account couple hits per
>>second).
>>
>>I got some knowledge of MapInfo MapXtreme - there is a mechanism of
>>map objects broker which is preinitializing map objects and "renting"
>>them on request. Don't you think that it would be great to have such
>>"map-buffering" mechanism? I'm inspecting MapServer's code to find
>>some clues wether it is hard to do it or not...
>>
>>Krzysztof Chodak
>>
>>_______________________________________________
>>Mapserver-users mailing list
>>Mapserver-users at lists.gis.umn.edu
>>http://lists.gis.umn.edu/mailman/listinfo/mapserver-users
>>
>>    
>>
>
>
>  
>