[108889] in North American Network Operators' Group
Re: Peering - Benefits?
daemon@ATHENA.MIT.EDU (Paul Vixie)
Fri Oct 31 12:53:36 2008
To: "Paul Stewart" <pstewart@nexicomgroup.net>
From: Paul Vixie <vixie@isc.org>
Date: Fri, 31 Oct 2008 16:53:16 +0000
In-Reply-To: <89D27DE3375BB6428DDCC2927489826A01B4782E@nexus.nexicomgroup.net>
(Paul Stewart's message of "Fri\, 31 Oct 2008 06\:14\:37 -0400")
X-Vix-MailScanner-From: vixie@isc.org
Cc: nanog@merit.edu
Errors-To: nanog-bounces@nanog.org
"Paul Stewart" <pstewart@nexicomgroup.net> writes:
> ...
>
> My question was meant at a much higher level - a level where costs are
> equal for peering/transit and all the "technical" and the "financial"
> homework has been done already.... now I'm the stage of one last meeting
> with top level management to explain "peering" and it's magic. These are
> mainly non-technical people - so my question to NANOG was for viewpoints
> on peering of which hopefully I could reinforce some of my own thoughts
> with. Whether or not someone operating at scale isn't the discussion -
> and it's funny how many people involved with companies (that are
> "operating at scale") have emailed me offline since this discussion
> started a few days ago with questions/thoughts and strategy.
if the financials and technicals are similar enough to be factored out,
then what you have to look at is possible variance between tactical and
strategic cost/benefit ratios. basically this boils down to the cost of
lock-in. if you're going to avoid lock-in then you have get your own
address space and build out to at least one IXP and then, buy diverse
transit. once you have done all that, the cost of also peering is in the
noise, whereas the advantage of also peering is noticeable if not always
easily measureable. if you're not going to avoid lock-in, then everything
you'd need to spend to avoid it can be avoided, and you won't be peering
unless it's for purely strategic reasons.
--
Paul Vixie
