On Mon, Dec 22, 2008 at 1:55 PM, Owen Densmore <[email protected]> wrote:

>
> Is anyone else pursuing a "The Network is the Computer" approach?  Any
> tales to tell?
>

Old habits die hard.  I did a demo of technology for a distributed file
system last week using Amazon EC2 for the network.

But back home, I've got the local network of retired laptops and
miscellaneous hardware fired up, again.  Just in time to turn them all off
for the holidays.

Starting up a half dozen c2d equivalents at EC2 makes me nervous -- what if
I forget and leave them running?  Like leaving the lights on at home, but
they're not at home and I can't walk around and look for them.  But if I
leave them running doing nothing at Amazon, they probably draw less power
than my forgotten light bulbs at home.

My scripts for starting EC2 clients just get cruftier.  The first twenty
times, waiting for the instance to report "running" was a good enough start
marker.  Then some instances needed a few more cycles to get their sshd's
working, so the script wraps the ssh in failure proofing.  Then some other
instances needed a few more cycles to get their DNS working, so the script
wraps the digs in more failure proofing.

I can see why this happens, even though all the machines are identical, all
the connections equivalent, and all the systems the same. There are paths
where everything goes just right, and paths where bad timing prevails.  So
my little ensemble of virtual blades expands to fill a little phase space of
linux boot variations, and my scripts get cruftier.

-- rec --
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org

Reply via email to