[33207] in North American Network Operators' Group
Re: Further Internet Metrics
daemon@ATHENA.MIT.EDU (Randy Bush)
Wed Jan 3 13:07:51 2001
From: Randy Bush <randy@psg.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: "dave o'leary" <doleary@juniper.net>
Cc: nanog@nanog.org
Message-Id: <E14DsII-000H2H-00@rip.psg.com>
Date: Wed, 03 Jan 2001 10:05:50 -0800
Errors-To: owner-nanog-outgoing@merit.edu
>>> e.g. how do you assign the traffic share of the traffic going to a
>>> smallish network which is multi-homed behind sprint, uunet, and which is
>>> also appears at mae-west? hint: if you look from a uunet or sprint
>>> perspective, then you assign it to uunet or sprint depending on
>>> perspective. but if you also peer openly at mae-west, then you don't
>>> assign it to either.
> If nothing else, a couple of well defined measures of market share would
> be quite useful, particularly if they were specified in such a way that
> they could be measured in a deterministic manner. However, as you
> point out this is essentially impossible, but I wouldn't characterize these
> as political, but rather as market issues.
i chose my example with care. it's a real problem. it is *extremely* hard
to get a point of perspective where you can get a meaningful measurement.
even the view from a widely-peered large provider (not a lot of those, eh?)
is biased by the type of traffic that provider has.
randy
