[Tilecache] WMS-T support

Brian Walawender Brian.Walawender at noaa.gov
Tue Jun 2 10:31:38 EDT 2009 

Guillaume,

We are using the tilecache to serve radar data ( http://radar.srh.noaa.gov/
).  Typically a radar image is updated every five minutes.  We have about
130 radars across the US with each radar offering about 10 different
products (layers).  On a busy day, we our static image pages will get nearly
100 million hits.  Typically those hits will be focused on the radars that
have storms.  For the most part an individual radar layer cache would be in
use for about an hour, so even though we are constantly rolling through the
cache directories, I think we are taking a tremendous load off our WMS
server with the tilecache.

The WMS-T is much more attractive than our current method which is very I/O
intensive because we are rotating the base images and (attempting) to rotate
the cache directories as a new image arrives.

bw

-----Original Message-----
From: tilecache-bounces at openlayers.org
[mailto:tilecache-bounces at openlayers.org] On Behalf Of Christopher Schmidt
Sent: Tuesday, June 02, 2009 6:33 AM
To: Guillaume Sueur
Cc: tilecache at openlayers.org
Subject: Re: [Tilecache] WMS-T support

On Tue, Jun 02, 2009 at 11:21:10AM +0200, Guillaume Sueur wrote:
> Hi,
> 
> I'm only thinking loud here, because the subject interests my own 
> projects. Do you think that a Tilecache approach is the best for dealing 
> with time data ? Doesn't this aim to build many cache dirs which will be
> rarely used and maybe never again ?

The Disk cache is not the only cache that TileCache has. memcached would
be an ideal candidate for this type of data, allowing you to set up a
host/number of hosts with memcached, and depend on memcache's LRU
behavior to clean up your cache for you.   

> More widely, when the parametric set of the request can have many 
> different values, is it wise to build a cache for each requested ? IMHO 
> a standard proxy (squid or whatever) could be more appropriate and 
> easier to maintain on a disk size or time basis. Am I wrong ?

Only if you're assuming that the cache isn't already cleaning up based
on LRU rules, which isn't neccesarily true.

> Best regards,
> 
> Guillaume
> 
> 
> Jim Groffen a écrit :
> > Hello,
> > 
> > I've finished the work on the WMTS implementation in
> > TileCache, but I need to:
> > 
> > - update it to the main trunk version
> > - improve the docs and maybe add more doctests
> > - incorporate any feedback to get it contributed
> > 
> > The work was done for OWS-6 and is a RESTful implementation,
> > I haven't implemented the KVP interface.
> > 
> > A major difference is The way you define the coverage of
> > your tiles, this needs some nice documentation to make it
> > as easy as possible.
> > 
> > There are only a couple of concerns I can see with my
> > solution and its impact on TileCache. One of which is how
> > I had to extend the config file reading. It's backwards
> > compatible, but adds complication. In short the normal
> > behavior of TileCache is to treat all sections as a layer
> > definition except for an options section. My change added
> > an option to the options section that allows you to list
> > config sections that are to be treated as TileMatrixSets
> > and not layers. This has no impact on existing config
> > files but the changes are in Server.py and not small. I
> > tried to have as few changes impact files besides the
> > wmts.py layer and service files as possible.
> > 
> > I'll hopefully have a patch prepared some-time next week,
> > so I can get some feedback.
> > 
> > Regards,
> > 
> > Jim Groffen.
> > 
> > -----Original Message-----
> > From: tilecache-bounces at openlayers.org
[mailto:tilecache-bounces at openlayers.org] On Behalf Of Christopher Schmidt
> > Sent: Tuesday, 2 June 2009 6:51 AM
> > To: Brian Walawender
> > Cc: tilecache at openlayers.org
> > Subject: Re: [Tilecache] WMS-T support
> > 
> > On Mon, Jun 01, 2009 at 03:38:48PM -0500, Brian Walawender wrote:
> >> Hello,
> >>
> >> Can I make WMS-T requests to tilecache?   My WMS requests look like:
> >>
> >> radar =  new OpenLayers.Layer.WMS( "TWX",
> >>         "http://
> >>
216.38.80.111/cgi-bin/mapserv?map=/usr/local/mapserver_config/TWX_time.map",
> >>         {
> >>                layers: 'TWX_N0R',
> >>                transparent: true,
> >>                time: '2009-06-01T15:55',
> >>                format: 'image/png'
> >>         }, {
> >>                singleTile: true,
> >>                isBaseLayer: false,
> >>                opacity: 0.9
> >>         }
> >> );
> >>
> >> My tilecache requests would look like:
> >>
> >> tc_url = ['http://radarcache.srh.noaa.gov/tilecache/tilecache.py'];
> >>
> >> radar = new OpenLayers.Layer.WMS( "State Borders", tc_url,
> >>         {
> >>                layers: 'TWX_N0R',
> >>                transparent: true,
> >>                format: 'image/png'
> >>        }, {
> >>                singleTile: false,
> >>                isBaseLayer: false,
> >>                opacity: 1.0
> >>        }
> >> );
> >>
> >> Is time a recognized param for tilecache?
> > 
> > No.
> > 
> > I feel like fredj wrote a patch that would allow you to create different
> > cache directories based on a specific parameter in the URL. However, I
> > don't remember ever integrating it, so there's a good chance that it got
> > lost in mailing list traffic. I would say that this would be the place
> > to look -- either in the code or in the mailing list for such a thing.
> > (Perhaps fredj can find it more easily, if he's listening.)
> > 
> > 
> > Regards,
> > --
> > Christopher Schmidt
> > MetaCarta
> > _______________________________________________
> > Tilecache mailing list
> > Tilecache at openlayers.org
> > http://openlayers.org/mailman/listinfo/tilecache
> > 
> > The contents of this email are confidential and may be subject to legal
or professional privilege and copyright. No representation is made that this
email is free of viruses or other defects. If you have received this
communication in error, you may not copy or distribute any part of it or
otherwise disclose its contents to anyone. Please advise the sender of your
incorrect receipt of this correspondence.
> > _______________________________________________
> > 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

More information about the Tilecache mailing list