[Tilecache] S3 storage and TileCache

[Steven M. Ottens]

On Nov 13, 2008, at 9:41 PM, Christopher Schmidt wrote:

> On Thu, Nov 13, 2008 at 12:28:28PM -0800, Josh Livni wrote:
>> Well, to be fair, S3 can give you bandwidth, as long as you  
>> request directly
>> from S3.  So if you have a situation where you know you can  
>> pregenerate all
>> tiles, for example stored in a TMS or Gmaps structure on S3, then  
>> you can
>> just have your client request tiles directly - S3 supplies tiles and
>> bandwidth, and TileCache just pre-renders the tiles for you.   Or  
>> perhaps
>> you could have TileCache (on a non-ec2 server) check local tile  
>> metadata
>>  (not currently supported) and redirect to S3 if it's there (and  
>> put it
>> there if it's not).
>
> Right, I'd say this is the right solution: take the hits on the
> TileCache server, and if the tile exists, redirect to S3. (You can
> always do the redirect, if you do things right, because by the time
> they've followed the redirect, you've finished your .set() onto the
> server.)

I'm not sure the .set() will be quick enough when a massive amount of  
non-generated tiles are hit, but it is a start.
I do like the local tile-metadata option and will look into that.
I'm wondering about the tilecache-metadata performance:  Has anyone  
used tilecache to manage 200+ layers with 15+ zoomlevels? I'm  
guessing that just checking if a tile exists in such a large amount  
of tiles isn't optimal. Would something like sqlite be helpful?

Steven


>
> -- Chris
>
>> But anyway, assuming you can pregenerate all tiles,
>> the first option might make lots of sense...
>>   -Josh
>>
>> On Thu, Nov 13, 2008 at 12:21 PM, Attila Csipa  
>> <plists at prometheus.org.yu>wrote:
>>
>>> On Thursday 13 November 2008 16:00:51 Steven M. Ottens wrote:
>>>> usage of the server running TileCache. My reason to use S3 is  
>>>> because
>>>> we run out of bandwidth at our setup and we assume Amazon has
>>>> slightly more available :)
>>>
>>> I think there is a misconception here... S3 gives you storage  
>>> space, not
>>> bandwidth. What you want sounds more like multiple EC2 instances  
>>> that serve
>>> static content fetched from S3 (and that would be close to a  
>>> setup that I
>>> actually use at the moment :)
>>> _______________________________________________
>>> Tilecache mailing list
>>> Tilecache at openlayers.org
>>> http://openlayers.org/mailman/listinfo/tilecache
>>>
>
>> _______________________________________________
>> Tilecache mailing list
>> Tilecache at openlayers.org
>> http://openlayers.org/mailman/listinfo/tilecache
>
>
> -- 
> Christopher Schmidt
> MetaCarta
> _______________________________________________
> Tilecache mailing list
> Tilecache at openlayers.org
> http://openlayers.org/mailman/listinfo/tilecache