[OSM-dev] scaling

[Matt Amos]

On Tue, Jan 11, 2011 at 1:28 PM, Kai Krueger <kakrueger at gmail.com> wrote:
> What this shift of pattern means for scaling is imho not entirely clear. On
> the one hand more and large imports would likely put more burden on the
> write paths of the API, as imports don't need to download data first to edit
> it. On the other hand with the growing density of map data, each map call is
> likely becoming much more complex than it used to be, shifting the burden
> back towards read loads.

however, those imports probably *should* be doing some sort of reading
in order to ensure they're not duplicating or stomping on existing
data. the fact that they don't generally at the moment is (probably)
not a reflection on the speed of the API ;-)

cheers,

matt