[Tilecache] TMS Server

[Bruce Foster]

All,

Thanks for the interesting thoughts on TMS server.

Basic idea is to weight the value that TMS gives over either WMS or
other protocols. Recently, while looking around I saw ArcGIS server
has modified the TMS and called it ArcGIS server TMS which is sort of
locked (no wonder). Then on the erdas side which always claims the
fastest, has some thing called otdf for tiles. I couldn't find any
performance number of these guys, they just claim without any number
evidence.

Now that we are moving away from WMS, which is more processor
intensive as paul pointed out, does anyone know what kind of
performance increase we get on TMS (if its cached). I'm looking for
numbers but none so far.

Be it arcserver or erdas, they talk much but not concrete test numbers
are revealed for one to make a decision. I saw the last year foss4g
presentation on performance, but the numbers dont look correct to me.
I was not sure if the test was done purly on random bbox, because the
throughput is around

recently i had access to a erdas server and while doing performance
test, it came out shocking low figures. Far less than what geoserver
and mapserver as in the foss4g presentation.  (well, even though its
not apple to apple). quite confusing..

assuming on a entry level server (intel xenon duel core, 2 gb ram,
500gb  7200rpm disk), one get a throughput of 12/sec with WMS, what
will the performance be with TMS?, will that be 10 fold or 50 fold?

On a curiosity note, how does google serve? I know they have massive
brute force with respect to server power and bandwidth.. but does it
serve TMS like tiles?

the other subject of performance, the known bottle necks are;

1. Disk read speed, 7200rpm disk vs 15000rpm disk vs SSD disk ...
2. Bandwidth 100 Mbps (100 mega bit per second or 12.6 mega byte per
second), this can carry only a certain amount of traffic only.
3. Network card and the routers, switches to which they are attached
4. CPU of the server, xenon was amd etc...

have you got any results or any other known bottleneck for image serving?


I beleive, if we can put them on a presentation, it will be nice.


Thanks

Burce







On Fri, Sep 18, 2009 at 9:16 PM, Christopher Schmidt
<crschmidt at metacarta.com> wrote:
> On Fri, Sep 18, 2009 at 07:28:32PM +0800, Bruce Foster wrote:
>> List
>>
>> I'm trying to find if there are any TMS server in the market now.
>>
>> I know TileCache can act as TMS but it uses WMS source and caches all
>> the tiles and then serves them. What I need is a server that can serve
>> TILES directly from images in real time.
>
> The "In real time" aspect is the issue here. The reason that tiles are
> cached is because, unless you have a perfectly generated image, you can't
> serve tiles 'in real time' out of a source image as fast as you can out
> of a static file on disk.
>
> However, if you don't want TileCache to cache, you can write a trivial
> Cache. Actually, you don't even need to: It's called 'Test', and it doesn't
> do any caching at all, so you'll always be reading the images live.
>
> As soon as you do this -- unless you're far more experienced with
> image processing than I am, which is a possibility, I suppose -- you'll
> find that you want the cache back.
>
> As for reading images from disk: For small images, use an Image layer.
> For big images, use a GDAL layer.
>
>> GeoWebCache is also similar to TileCache I believe.
>>
>> Only issue is, if we make Tile Cache and then serve, the overhead on
>> the harddisk space is high. Just wonder if this is a good idea.
>
> Yes.
>
> Regards,
> --
> Christopher Schmidt
> MetaCarta
>