[OpenLayers-Users] Performance of own (vector) data in Open Layers

[Christopher Schmidt]

On Thu, Feb 05, 2009 at 02:24:43PM +0100, Kai Behncke wrote:
> Dear Mika,
> 
> well "my server " is a normal root server in the Internet.
> "My data" in that case are Vector-Point-Data, I have them in a PostGIS-Database or can have them as a "typical" shape-file.
> I`m quite new to Open Layers and definetely quite impressed.
> 
> I wonder in general now which way might be the most performant way to show "own" geodata (in Combination with the google-api-layer,osm-api-ayer,virtual earth-api-layer) to the public in Open Layers.
> 
> But, well, I will test beside from that question also how to improve the speed of the WFS, I`m pretty sure there will be some possibilities (indizes and so on).

It seems you are trying to address multiple performance concerns at the
same time.
 * Server side: the time it takes to return a set of features in
   response to a query
 * Client side: How long it takes to parse and draw them.

In general, the latter is based entirely on the type of features: there
is usually not a *significant* difference i parsing time between
different formats of vector serialization once you've whittled
the result set down to the IE6/IE7 limit of ~100 features.
Once you've optimized the server side, investigating different options
on the client side should be relatively simple, but pick one at a time
and optimize.

Regards,
-- 
Christopher Schmidt
MetaCarta