It all goes back to the question WHAT FOR ?
Unless there is a SUSTAINABLE PURPOSE there can be no chance of meaningfull
success.
On the server side: I can only quote my example. We have a server on
sponsored badwidth which has been idling for the past year because 2 disks
having failed. Easier to copy our data on a friendly server than to take a
trip across the globe (to Seattle) to replace the disks (and upgrade the
machine). Obviously they will not get fixed up if there isn't any meaningful
purpose.
I know of two more identical examples.
Still if you have 2 Tb of tiles to serve and if it is for some meaningfull
cause I can try my best to make arrangements within a few hours and I am
positive I will succes. And I'm certain that Chris, as long as a meaningfull
purpose is really there, can do 100 times better than me.
I am probably a nut case and being in a third world country (South Africa)
my problems are probably very different from yours.
Still I have 2 tiling servers (12 x 1Tb disks in total) next to me and I'm
building a third tiling server with 24 x 1Tb storage (have been waiting 2.5
months for the right raid card to arrive). Just on top of my desk I count 13
more 1Tb USB disks and I have a few more on a shelf. In my little world I
"do" identify with the "too much data" problem Chris had previously
mentioned.
Still, I havent been running any of the tilers for thepast week. Data has to
be checked first.
Some will need fixing georeferencing / additional coordinate system infos,
some other will need white or black no-data areas to be masked, some will
cry for colour balancing. And then, like yesterday, one dataset with stupid
sun reflections on water will come along and waste me one full day on
nothing.
Sure, one can just tile everything as it comes along, quick quick, after all
it's just running the same script over and over.
But believe me: output will be very disappointing and not worth the disk
space which has been wasted on.
... and the same question comes up over and over: WHAT FOR ?
In my little world I have my own local motivation: the 2010 Soccer World Cup
... hopefully some local cellular operator will come along and make all my
work worthwhile.
And, more importantly, who is going to pick up the tab?
Since the are many "tabs" to be picked up you'll need to specify which tab
you wish an answer on.
Ciao
Maurizio
----- Original Message -----
From: "Jeffrey Johnson" <[email protected]>
To: "Maurizio" <[email protected]>
Cc: "Christopher Schmidt" <[email protected]>;
<[email protected]>
Sent: Wednesday, August 12, 2009 3:44 AM
Subject: Re: [OAM-talk] Open Aerial Map Restart Meetup
And, more importantly, who is going to pick up the tab?
On Tue, Aug 11, 2009 at 6:43 PM, Jeffrey Johnson<[email protected]>
wrote:
I believe tile serving is the smallest of the challenges: if data is
available for serving a server will miracously become available
"somewhere"
and if that is not enough one more server will follow.
I'm not sure this is the case. I have the tooling setup on AWS to
process and tile as much imagery as you can throw at me ... I could
make quick work of the 4TB of NAIP I described earlier ... but what do
you intend to do with the few dozen TB that result from this process?
Jeff
_______________________________________________
talk mailing list
[email protected]
http://openaerialmap.org/mailman/listinfo/talk_openaerialmap.org
