[Chapter-delegates] [Itu2012chapters] The Internet Days Conference in Stockholm Oct 23-24 - net neutrality session Oct 24

Tue Oct 9 23:55:51 PDT 2012

Patrik,

all great points. Plus, you don't get "QoS like a circuit" over more than a single link. No way it can work in networks outside the highly-connected Northern Telcoland and even there, doubtful, right?

"A connection with high priority" means "a connection for which the telco/carrier/etno-member wants to extort a higher fee." How neat that ETNO comes with a crisp dictionary.

Yours,

Alejandro Pisanty

! !! !!! !!!!
NEW PHONE NUMBER - NUEVO NÚMERO DE TELÉFONO

+52-1-5541444475 FROM ABROAD

+525541444475 DESDE MÉXICO

SMS +525541444475
     Dr. Alejandro Pisanty
UNAM, Av. Universidad 3000, 04510 Mexico DF Mexico

Blog: http://pisanty.blogspot.com
LinkedIn: http://www.linkedin.com/in/pisanty
Unete al grupo UNAM en LinkedIn, http://www.linkedin.com/e/gis/22285/4A106C0C8614
Twitter: http://twitter.com/apisanty
---->> Unete a ISOC Mexico, http://www.isoc.org
.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .

________________________________
Desde: itu2012chapters-bounces at elists.isoc.org [itu2012chapters-bounces at elists.isoc.org] en nombre de Patrik Fältström [patrik at frobbit.se]
Enviado el: miércoles, 10 de octubre de 2012 01:43
Hasta: Vint Cerf
CC: Chapter Delegates; itu2012chapters at elists.isoc.org
Asunto: Re: [Itu2012chapters] The Internet Days Conference in Stockholm Oct 23-24 - net neutrality session Oct 24

On 10 okt 2012, at 02:17, Vint Cerf <vint at google.com<mailto:vint at google.com>> wrote:

I understand that ETNO is now amending its "sender pays" to a version of "sender pays for better service" that leads one to imagine that the service for which both sender and receiver pay is "not very good"... incentives are all wrong here.

It is not only "sender pays"(*) but also a theory that "addition of QoS will make things better", i.e. addition of QoS will increase the ability to send a higher number of ones and zeroes over the same link.

This is a fundamental flaw in the thinking by telco-minded people. Packet based networks do not behave that way(**).

Priority is _only_ effective if a link is full, and what priority does is to influence the decision making process of what packets to delay and throw away.

Now, to be able to do what telco-minded-people are interested in doing, they have to introduce _artificial_ situations where links are full. I.e. the goal is to have invested in bandwidth that is not used.

Artificial scarcity to ensure the prices on the market stays high.

Think about what the average politician thinks about that in a world of market economy.

Specifically as the number of bits to send in the tubes do double every year(***).

The rest of us do believe it is better to add bandwidth as adding any kind of priority might make the experience of using _a_full_link_ better, but only for a very small amount of "too much traffic". Two weeks later (due to increase of traffic) the link is so saturated that it must be upgraded anyways.

   Patrik

(*) I do not know what "a connection with high priority" is. See this document by myself and Gordon Lennox:

http://stupid.domain.name/stuff/circuits.pdf

(**) See this presentation from last RIPE meeting by Geoff Houston:

<https://ripe65.ripe.net/presentations/67-2012-09-25-qos.pdf>https://ripe65.ripe.net/presentations/67-2012-09-25-qos.pdf

(***) Netnod yesterday announced we have 100G interfaces at the IX in Stockholm, and also the first customer. Bahnhof that is an access network provider. But not one of those telco-bell-heads that are members of ETNO.

https://www.netnod.se/bahnhof-becomes-first-netnod-100-gbps-customer

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://elists.isoc.org/mailman/private/chapter-delegates/attachments/20121010/e021a53c/attachment.htm>