[115624] in North American Network Operators' Group
RE: Fire, Power loss at Fisher Plaza in Seattle
daemon@ATHENA.MIT.EDU (Michael K. Smith - Adhost)
Fri Jul 3 15:14:13 2009
Date: Fri, 3 Jul 2009 10:57:56 -0700
From: "Michael K. Smith - Adhost" <mksmith@adhost.com>
To: "Tomas L. Byrnes" <tomb@byrneit.net>,
"David Hubbard" <dhubbard@dino.hostasaurus.com>,
"NANOG list" <nanog@nanog.org>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
-----Original Message-----
From: Tomas L. Byrnes [mailto:tomb@byrneit.net]
Sent: Fri 7/3/2009 10:20 AM
To: David Hubbard; NANOG list
Subject: RE: Fire, Power loss at Fisher Plaza in Seattle
=20
This begs the question of what basic parameters should be for a "carrier
hotel" or co-lo.
Given that we're getting designated "Critical Infrastructure", we'd
getter start coming up with some, or we'll have them defined for us.
----
I think the more important question is, "what do you consider =
redundancy?" We have facilities in Plaza East (no down) and Plaza West =
(unaffected). If you are critical infrastructure there is no amount of =
redundancy that you should offload onto a colo provider. Instead, you =
build your redundancy across different data centers, different =
providers, different everything. If you rely on a single provider for =
any of the aforementioned then you have built in at least one single =
point of failure, regardless of the resiliency of the underlying =
provider.
My .02, worth almost every penny.
Mike
