Who Confirms The Accuracy Of ISP Usage Meters?

[Dave Farber <dave () farber net>]

Begin forwarded message:

> From: Jason Livingood <jason_livingood () cable comcast com>
> Date: January 13, 2010 3:26:52 PM EST
> To: Dave Farber <dave () farber net>, ip <ip () v2 listbox com>
> Subject: Re: [IP] Who Confirms The Accuracy Of ISP Usage Meters?

> This post on Broadband Reports came out last week.  The main concern  
> I had
> at the time was a specific claim about Comcast's transparency on the  
> matter,
> rather than the general question of meter accuracy (which seems a  
> logical
> question to ask, and which we anticipated).  Here is my response to  
> that,
> which was posted on their forum last week:
>
> <BEGIN QUOTE OF BBR POST>
> > Karl wrote: "Comcast recently unveiled their new meter in Portland,  
> > using a
> > firm of their choice to determine it's accuracy. The firm they  
> > hired, of
> > course, proclaimed proudly that the meter was accurate to within  
> > 0.5% each
> > month.
> >
> > Who'll confirm this? Nobody."
>
> While I can understand the generalities of your story*, the specific  
> ding
> against how Comcast released the meter seems unjustified. In an  
> environment
> where an ISP could simply release a meter, it seems an objectively  
> good
> thing that we used an independent party to audit the meter, that we  
> had them
> share a report on the matter, and that they fully described how the  
> system
> functioned and how they tested it.
>
> If you question their independence or background, which I think is
> unfounded, you can find out who they are here:
> http://netforecast.com/ContactInfoFrameset.htm
> - Scott Bradner - http://netforecast.com/biography%20Scott%20Bradner%20.htm
> - Rebecca Wetzel -
> http://netforecast.com/biography%20Rebecca%20Wetzel%20.htm
> - John Bartlett - http://netforecast.com/biography%20John%20Bartlett.htm
> - Peter Sevcik - http://netforecast.com/biography%20Peter%20S.htm
>
> You can see their client list here:
> http://netforecast.com/ClientFrameset.htm
>
> And I can tell you they were highly recommended by members of the  
> Internet
> community.
>
> Jason
> --
> JL
> Comcast
>
> * I think it is completely fair to ask how you can stand behind a  
> meter and
> is it accurate. And for that reason, I think the basics of your  
> story prompt
> a worthwhile discussion. When we planned the meter, I thought that  
> it'd be
> the case that the question of proving its accuracy would logically  
> be raised
> and so I suggested having a 3rd party audit it and - further - share  
> the
> results of that with our customers and the public.
> <END QUOTE OF BBR POST>
>
> http://www.dslreports.com/forum/r23607298-Hmmm
>
> Jason Livingood
> Comcast - Internet Systems Engineering
>
>
> On 1/13/10 2:19 PM, "Dave Farber" <dave () farber net> wrote:
>
> > Begin forwarded message:
> >
> > From: "George Ou" <george_ou () lanarchitect net>
> > Date: January 13, 2010 1:16:58 PM EST
> > To: <schlake+gmail () gmail com>, "'Bob Poortinga'" <nnsquad () k9sql us>
> > Cc: nnsquad () nnsquad org
> > Subject: [ NNSquad ] Re: Who Confirms The Accuracy Of ISP Usage  
> > Meters?
> >
> > This kind of paranoia is just getting silly.  It's basically  
> > assuming that
> > the ISP is committing fraud that is easily detectable.
> >
> > There are so many easy ways to measure this it's not even funny.   
> > If you're
> > running a firewall like IPCOP or for example, it has extensive  
> > usage reports
> > which will almost certainly match what the ISP is reporting.   
> > Moreover, you
> > can measure the internal Ethernet interface which does not include  
> > DoS
> > traffic because it gets blocked by the firewall.
> >
> > You can also easily enable PERFMON logging built into Windows and get
> > second-by-second usage data.  This only gives you measurements on a  
> > single
> > computer though but a lot of people only have one computer.  You  
> > could log
> > data from multiple computers and just merge the data though.
> >
> >
> > George
> >
> > -----Original Message-----
> > From: nnsquad-bounces+george_ou=lanarchitect.net () nnsquad org
> > [mailto:nnsquad-bounces+george_ou=lanarchitect.net () nnsquad org] On  
> > Behalf Of
> > Schlake
> > Sent: Wednesday, January 13, 2010 9:02 AM
> > To: Bob Poortinga
> > Cc: nnsquad () nnsquad org
> > Subject: [ NNSquad ] Re: Who Confirms The Accuracy Of ISP Usage  
> > Meters?
> >
> > So what we need is a independent device that can measure that network
> > usage.  It should be open source, to be verifiable, and cheap.  It
> > won't fix the problem, but it would put a solution in the hands of
> > many users.  The more users that have it the more accountability  
> > which
> > could be imposed on the service providers.
> >
> > A small box with two ethernet ports, one for the outside world, and
> > one for the inside world.
> >
> > A web server on the inside world that displays statistics.
> >
> > An NTP server that keeps the time up to date.
> >
> > A configurator page that lets users configure ip blocks to be tracked
> > separately (unmetered streaming video from the service provider for
> > instance).
> >
> > A NAT-like interface that can count packets at a variety of
> > resolutions.  Total packets through would be the main one, but what
> > about attacks on peoples homes where you send a large amount of
> > traffic at an IP address to raise their bill?  If that was detected
> > and listed as a separate line item it could be useful.  In general,
> > the more data a person can store about their own traffic the more
> > useful it would be (I'm sure law enforcement would agree on this
> > point).
> >
> > It would need a good robust data storage system internally, and a way
> > to automatically synchronize the data into a machine on the inside
> > world for backup purposes.
> >
> > Obviously if someone implements this you should send me some of the
> > money you make from my idea!
> >
> >
> > [ Actually, much of this functionality already exists in, or
> >   could be added to, somewhat higher-end commodity cable/DSL
> >   modem/routers.
> >
> >   But what happens if a subscriber's locally gathered usage
> >   statistics turn out to be at wide variance from the ISP's
> >   official figures?  Then the real fun begins.
> >
> >      -- Lauren Weinstein
> >         NNSquad Moderator ]
> >
> >
> >
> > On Wed, Jan 13, 2010 at 8:35 AM, Bob Poortinga <nnsquad () k9sql us>  
> > wrote:
> > > Here's a good article in DSL Reports which brings up some important
> > points:
> > >
> > <http://www.dslreports.com/shownews/Who-Confirms-The-Accuracy-Of-ISP-Usage-M
> > eters-106292>
> > > --
> > > Bob Poortinga  K9SQL
> > > Bloomington, IN  US



-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com