I need to look more into this, probably after I get 0.7 out (not the 0.7 still advertised on the website).
Note that, unlike JTileCache, GeoWebCache stores the tiles as compressed images in the format you want to return, so most of the work is figuring out what the correct cache key is for a given request and then how to get that from the backend. I had to pull JAI out right before 0.6 because I had some problems with it. But it's only three lines in the code (slicing up tiles), and wasn't that important since it saves milliseconds after a WMS request that typically takes several seconds anyway. I wrote a few lines on how GeoWebCache (currently) works, short and subject to change, but hopefully a starting point for further inquiries: http://geowebcache.org/trac/wiki/developers/overview -Arne ps. If anyone knows of a good and simple persistent hashtable, suitable for storing PNGs and the likes to a file, please let me know :) Simone Giannecchini wrote: > Here you have a few thoughts that I have been giving to three topics > which may be seems different but which I would like to reconcile at > least at the tilecache level > > 1>geoweb cache > 2>geotools tile > 3>geotools internal overviews management (yeah I am talking about > wither reusing overviews when you have them or about generating them > on the fly using JAI and persisting them to disk when you don't have > them) > > The goal in my mind would be unifying the tile cache management for > the three different things into only one. I have done some work in > this direction in the past but for medical imaging and for speed limit > enforcement cameras (yeah, don't spread the word out there :-) ). I > have played while at foss4g with small parts of this code and tried to > integrate them in uDig and the difference in performances is pretty > big > > I already talked a little bit to arne just to try and understand how > geowebcache works (even because we plan on using it). I think it would > be great to have some documentation on how geowebcache works (I don't > know if it already exists :-) ). > > > Simone. > > On Sat, Mar 1, 2008 at 12:41 PM, Chris Holmes <[EMAIL PROTECTED]> wrote: > >> Interesting, I didn't realize that org.geotools.tile exists. >> >> It might be really good to reconcile the implementation in geotools with >> what Arne's done in GeoWebCache, so we can reuse code. All our work we >> like to get in geotools if there's the potential for wider reuse, but I >> wasn't aware there was already work done on this. I'm not sure if it's >> appropriate to reuse everything, but it'd be nice to at least use the >> same code to access alternate tile servers. Also note that we've got >> 'metaTiling' code in GeoWebCache, leveraging JAI, which can make for >> some nicer looking tiles. Arne should be able to tell you more and >> figure out how best to collaborate. >> >> best regards, >> >> Chris >> >> >> Jody Garnett wrote: >> >>> Sounds good, I am the module maintainer for unsupported/tile if you have >>> any questions please ask. >>> >>> You should find the code base organized to allow: >>> - alternate TileCache implementations to be substituted (although >>> perhaps we need to make a plug-in mechanism?) >>> - additional "tile" servers >>> >>> We may find it worthwhile to talk to the JAI Gurus on this list at a >>> later stage - they could hook us up with the JAI Tile Cache etc... >>> >>> In the broader scheme of things we should make a specific MapLayer >>> implementation for the MapContext data structure and allow >>> the GeoTools renderer to make use of these services. >>> >>> For my part I would like to make a implementation that talks to MapGuide >>> Open Source; so GeoServer could hang out the front end and offer a good >>> WMS implementations on top of it. >>> >>> Jody >>> >>>> Hi everyone, >>>> >>>> the module org.geotools.tile currently has an implementation for >>>> accessing the NASA World Wind server and an implementation of a >>>> SimpleTileCache, which caches only the result of the last query. >>>> >>>> We would like to add support for WMS-C [1]. Server implementations are >>>> available from MetaCarta [2] or from GeoWebCache [3]. >>>> >>>> Additionally we want to create at least two tile cache implementations >>>> (MemoryTileCache and DiskTileCache). It should be possible to configure >>>> the size of the cache (number of tiles to be cached) and to plug-in a >>>> cache algorithm like Least Recently Used, Least Frequently Used, or >>>> maybe Adaptive Replacement Cache. Caches can be used in a cache >>>> hierarchy. >>>> >>>> Are there currently and other activities in that area that we should be >>>> aware of? Are there any ideas oder suggestions from you? >>>> >>>> Regards, >>>> Holger >>>> >>>> >>>> [1] http://wiki.osgeo.org/wiki/WMS_Tiling_Client_Recommendation >>>> [2] http://www.tilecache.org/ >>>> [3] http://geowebcache.org/ >>>> >>>> --- >>>> Holger Jaekel >>>> phone: +49 (89) 121528-75 mailto:[EMAIL PROTECTED] >>>> fax: +49 (89) 121528-79 http://www.gaf.de >>>> GAF AG Arnulfstr. 197 80634 Muenchen Germany >>>> >>>> Vorstand: Dr. Peter Volk, Aufsichtsratsvorsitzender: Marcello Maranesi >>>> >>>> Amtsgericht Muenchen HRB 140 509, Firmensitz: Muenchen >>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------------------- >>>> This SF.net email is sponsored by: Microsoft >>>> Defy all challenges. Microsoft(R) Visual Studio 2008. >>>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ >>>> _______________________________________________ >>>> Geotools-devel mailing list >>>> Geotools-devel@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/geotools-devel >>>> >>>> >>> >>> ------------------------------------------------------------------------- >>> This SF.net email is sponsored by: Microsoft >>> Defy all challenges. Microsoft(R) Visual Studio 2008. >>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ >>> _______________________________________________ >>> Geotools-devel mailing list >>> Geotools-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/geotools-devel >>> >>> >>> >>> >> ------------------------------------------------------------------------- >> This SF.net email is sponsored by: Microsoft >> Defy all challenges. Microsoft(R) Visual Studio 2008. >> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ >> _______________________________________________ >> Geotools-devel mailing list >> Geotools-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/geotools-devel >> >> >> > > > > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Geotools-devel mailing list Geotools-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geotools-devel
