OSM's future. Was: Re: [Openstreetmap] OSM Performance is terrible

[SteveC]

* @ 02/02/06 07:12:22 AM preben at chin.dk wrote:
> > Even at 5 OK-spec machines with excellent JANET hosting, it's not that
> > fast. So we probably need another 3-5 excellent machines. And these
> > machies, for the most part, need to be physically close to each other
> > because of the interplay of the database, tile creation and the applet.
> >
> > The type of bandwidth we're using is huge. The tile server ships a lot
> > of data around, as does www. Commercial hosting would probably run in
> > to hundreds of pounds a month.
> 
> Just a question: Did you analyse in which step the most
> bandwidth/processing power/waiting time is being used? For example, if the
> landsat image fetching is the slowest part, could the editing be speed up

The applet doesn't rely on landsat, but I see yout point. Mikel will
know a bit more about this than me.

If you accept that whatever the implementation, OSM is a high bandwidth
and processing proposition then the current bottleneck is memory. All
the machiness are swapping (apart from dev) due to low memory. Why don't
I put more memory in the computers? They're not my computers. :-)

www runs more that one website and also has many modules there, like
php, svn etc that it needn't for just running OSM. There are many things
that need analysis (please dig in!), I think it was imi that pointed out
how slow REXML was for example.

> by disabling this? How many hits for individual map locations are there
> currently?

See other emails :-)

have fun,

SteveC steve at asklater.com http://www.asklater.com/steve/