[Tilecache] getting tilecache 1.4 working...

[Adrian Zhao]

Thanks Bill.  That makes a lot of sense to me.  Can I say that seeding
the cache is to generate some kind of lookup table for TileCache to use
when serve the client request.  If it finds it in this look up table it
will use cached image first.  If not it then generated on the fly.

If this is the case, do you have to refresh the cache seeding regularly?


I think we talked about what cache seeding is and if you can tell me
little bit more about how the cache seeding works that would be even
greater.

Thanks,
Adrian

-----Original Message-----
From: Bill Woodall [mailto:bwoodall at wardrobe.dhs.org] 
Sent: Thursday, February 22, 2007 11:13 PM
To: Adrian Zhao
Cc: Schuyler Erle; tilecache at openlayers.org
Subject: Re: [Tilecache] getting tilecache 1.4 working...

On Thu, 2007-02-22 at 14:19 -0500, Adrian Zhao wrote:
> I am assuming that you guys are writing a long reply for my previous
> message.
> 
> Here are two quick questions:
> 
> 1. What is the purpose of seeding the cache? 

Short answer: Responsiveness to the WebClient

Longer answer:
A purpose of TileCache is to serve up tiles that were previously
generated, so the WebClient does not have to sit and wait while the map
server generates the needed tiles.  If the Map that the WebClient has
several overlays active at the same time ( which is my case ) and using
the default buffer (WMS) at 2, then you would have something like 7x8
tiles x number of Layers. So in my case, 9 layers comes to ~504 tiles
for the map server to generate taking an average 6 seconds per tile
compared to the 0.01 second with TileCache.

> How important?

It depends, but since TileCache will build the cache as needed, so those
tiles that will get requested often will only have to be generated once.

Again in my case, I have 114 zipcodes that a WebClient could request, so
I seeded 9 layers in 9 zoom levels to have ~1448000 tiles in
about 6G of disk space.

I hope that made sense.

........Bill,