[171381] in North American Network Operators' Group
RE: What Net Neutrality should and should not cover
daemon@ATHENA.MIT.EDU (Hugo Slabbert)
Mon Apr 28 00:31:45 2014
X-Original-To: nanog@nanog.org
From: Hugo Slabbert <hslabbert@stargate.ca>
To: "NANOG Operators' Group" <nanog@nanog.org>
Date: Mon, 28 Apr 2014 04:31:37 +0000
In-Reply-To: <CAP-guGULA3y2J57hKr9ZoU3wAerzjKoWekJOn6d_+BPjrUz-SA@mail.gmail.com>
Errors-To: nanog-bounces@nanog.org
>> #4 On QoS (ie fast lane?):=0A=
>> In some of the articles I skimmed there was a lot of talk about fast lan=
e=0A=
>> traffic but what this sounds like today would be known as QoS and=0A=
>> classification marking that would really only become a factor under=0A=
>> instances of congestion. The tech bloggers and journalists all seems to =
be=0A=
>> unanimously opposed to this but I admit I am sort of scratching my head =
at=0A=
>> the outrage over something that has been in prevalent use on many major=
=0A=
>> networks for several years.=0A=
>=0A=
>It's prevalent on private work networks and users hate it. It=0A=
>generally disables activities the network owners don't approve of=0A=
>while engaging in doubletalk about how they're OK with it. Users don't=0A=
>want to see this migrate outward.=0A=
>=0A=
>Regards,=0A=
>Bill Herrin=0A=
=0A=
A couple of things come into play here, I think:=0A=
=0A=
1. Prevalence of congestion on shared-bandwidth media, e.g. cable.=0A=
2. Who controls the QoS?=0A=
=0A=
A thumbsuck seems to indicate that #1 is high or at least significant enoug=
h to cause user-visible impact in e.g. places where cable internet provider=
s in the US don't face any real competition. So, QoS measures can come int=
o play in those locales/situations.=0A=
=0A=
For #2: QoS is good. Deciding which traffic gets passed and which dropped=
in congestion is, in and of itself, a good ability to have and can be a gr=
eat value-added service. "You want to run VoIP on the same line as your re=
gular data but want to ensure your VoIP traffic gets through? No problem: =
Here's our QoS-Extraordinaire service!" The concern comes from the direct=
ion the rules seem to be taking on this in shifting control/input on how Qo=
S is applied from (a) just ensuring network-control doesn't get drowned out=
and (b) a value-add service where the customer picks their traffic priorit=
ization, to an external party paying for preferred access to the BB-provide=
r's customers. As a customer of BB-provider, this means that someone else =
now has control over how my packets get delivered based on a deal they cut =
with BB-provider. It's not about helping the end-user: It's about enrichin=
g BB-provider. It's another situation of opening up a two-sided market and=
fostering a situation where established players on the content side who ca=
n afford to pay BB-providers A through ZZ get beneficial treatment and ther=
e can be a larger barrier to entry to the markets occupied by those players=
.=0A=
=0A=
Yes, QoS should only come into play where congestion is involved. But, fro=
m experience we can see there are ways to let BE traffic degrade to affect =
e.g. latency-sensitive traffic without having to actively throttle it. Sur=
e: the "commercially unreasonable" clauses *should* protect against that to=
a degree, but that's a very vague definition that creates a lot more regul=
atory overhead. Rather than saying "you're not allowed to accept payment t=
o prioritize one content provider's traffic over another's," the FCC would =
now have to investigate this situations on a case by case basis to determin=
e if a specific situation is "commercially unreasonable". So; basically, h=
ow much confidence do we have in the FCC's capacity/competence in enforceme=
nt of those types of regulations? We could also tack on that this could cr=
eate a "barn door" situation, where lax or vague rules go into effect, "the=
market decides", and then we have a helluva time trying to stuff the cat b=
ack in the bag because at that point this type of preferential treatment wo=
uld already be an established/common practice.=0A=
=0A=
--=0A=
Hugo=0A=
Network Specialist=0A=
Phone: 604.606.4448=0A=
Email: hslabbert@stargate.ca=0A=
=0A=
Stargate Connections Inc.=0A=
http://www.stargate.ca=0A=
=0A=
________________________________________=0A=
From: NANOG <nanog-bounces@nanog.org> on behalf of William Herrin <bill@her=
rin.us>=0A=
Sent: Sunday, April 27, 2014 10:56 AM=0A=
To: Rick Astley=0A=
Cc: NANOG Operators' Group=0A=
Subject: Re: What Net Neutrality should and should not cover=0A=
=0A=
On Sun, Apr 27, 2014 at 2:05 AM, Rick Astley <jnanog@gmail.com> wrote:=0A=
> #3 On paid peering:=0A=
> I think this is where people start to disagree but I don't see what shoul=
d=0A=
> be criminal about paid peering agreements. More specifically, I see serio=
us=0A=
> problems once you outlaw paid peering and then look at the potential=0A=
> repercussions that would have.=0A=
=0A=
Double-billing Rick. It's just that simple. Paid peering means you're=0A=
deliberately billing two customers for the same byte -- the peer and=0A=
the downstream. And not merely incidental to ordinary service - the=0A=
peer specifically connects to gain access to customers who already pay=0A=
you and no one else. Where those two customers have divergent=0A=
interests, you have to pick which one you'll serve even as you=0A=
continue to bill both. That's a corrupt practice.=0A=
=0A=
What sort of corrupt practice? You might, for example, degrade your=0A=
residential customers' speed to the part of the Internet housing a=0A=
company you think should pay you for peering. Or permit the link to=0A=
deteriorate while energetically upgrading others to keep pace with the=0A=
times. Same difference.=0A=
=0A=
This doesn't have to be true. You could bill downstreams for=0A=
consumption and exclude the paid peering from that calculation. But=0A=
you don't do that. And you aren't planning to.=0A=
=0A=
=0A=
> #4 On QoS (ie fast lane?):=0A=
> In some of the articles I skimmed there was a lot of talk about fast lane=
=0A=
> traffic but what this sounds like today would be known as QoS and=0A=
> classification marking that would really only become a factor under=0A=
> instances of congestion. The tech bloggers and journalists all seems to b=
e=0A=
> unanimously opposed to this but I admit I am sort of scratching my head a=
t=0A=
> the outrage over something that has been in prevalent use on many major=
=0A=
> networks for several years.=0A=
=0A=
It's prevalent on private work networks and users hate it. It=0A=
generally disables activities the network owners don't approve of=0A=
while engaging in doubletalk about how they're OK with it. Users don't=0A=
want to see this migrate outward.=0A=
=0A=
Regards,=0A=
Bill Herrin=0A=
=0A=
=0A=
=0A=
--=0A=
William D. Herrin ................ herrin@dirtside.com bill@herrin.us=0A=
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>=0A=
Falls Church, VA 22042-3004=0A=
