[OSM-dev] Lite OSM backend

Nicola Ranaldo ranaldo at unina.it
Thu Sep 7 09:26:26 BST 2006


On Thursday 07 September 2006 10:04, Nigel Magnay wrote:
> We're talking about several different things here

:)

> * An 'plug-n-play' OSM server. This is a good idea, which is partially why
> I wrote OSM2Go. But, distributing a database with the server as well would
> make the download (IMO) intractable at over 1Gb. I could ship it with
> planet.osm.bz2 instead, and have a 'please wait for 2-3 hrs' when you
> start. NB: The majority of the time is spent in the database doing bulk
> upload, not parsing the source data.

Yes it's a very good idea, and as it's a server, users are supposed to be 
skilled and choose the best data set they want, so it could be distributed 
alone without any planet.osm.

> * Cutting down datasets to alleviate the above. IMO also good idea, as day
> to day I'm unlikely to start needing data from other continents.

Yes, this should be applied to the xml files and to the binary files.

> * Generating a binary data format. Not sure about this - parsing of the XML
> isn't particularly slow, bz2 does a fairly substantial job of removing
> redundancy, and you still have to put it into some kind of database. Unless
> you're writing your own, but that's a different kettle of fish - storage
> for the nodes alone is ~144Mb, and that's before you add any kind of
> indexing.

Parsing the xml is fast, howewer if you want to do some filtering on data you 
need a lot of memory, or you have to do lookups in the database, slowing down 
the import. If you have a pocketpc in your car and want to update to last UK 
maps you need a binary file!!!!

	Niko




More information about the dev mailing list