On 09/10/2007, Gregory Maxwell <gmaxwell(a)gmail.com> wrote:
Wikimedia is reaching a point where increases in
bandwidth usage *may*
actually result in reductions the total bandwidth costs. I'm not too
concerned there. Certainly, Commons itself isn't a major bandwidth
user.
I'm surprised, given images are large. I suppose thumbnails would be
counted against the wiki they're generated on.
Nor is capacity... Common's growth over this year
looks mostly linear,
with an average growth rate of 27960 bytes per second.
[...]
redundancy (write every image twice, OK), even assume
a doubling of
the growth.. As long as it doesn't go exponential it's not scary at
all.
You mean like en:wp's popularity? Imagine a popular Commons. (Assuming
we can work out how to deal with the floods of crappy uploads that
will result in.)
I currently mirror commons (and all WMF images) at
home. At the
current rates have space for a couple of years. Perhaps we'll have a
nice growth spike? that would be good: I'd rather outgrow my storage
before it starts failing on it.
Heh. You realise you have one of the few mirrors of Commons, then. You
may wish to put some tarballs up on
download.wikimedia.org.
Video should only up the rate by a constant factor. ..
none of this is
hard. To make it hard we need the increasing returns that can only
come from increased adoption. Commons storage isn't hard but the
mission of the commoners should be to make it hard.
Precisely :-)
- d.