[Tsvwg] Re: terminology issues in draft-floyd-tsvwg-besteffort - flat monthly pricing

[Bob Briscoe <[e-mail]>]

Sally,

I just realised that I should have added a very important point about "who 
controls the choice" to the rather philosophical response below.

I'm trying to make sure we at the IETF don't have to decide whether the 
Internet more favours rich people, malicious people or selfish people. I'm 
trying to make it so what you think or what I think about this doesn't 
matter (at least any more than what everyone else in the world thinks).

Firstly, what I'm proposing would sit alongside the current Internet 
(re-ECN packets are distinguishable from non-re-ECN packets, so they 
effectively partition all traffic orthogonally to Diffserv).

Secondly, one partition would be the current Internet, with all its 
vulnerabilites to selfish and malicious users as well as all its strengths. 
The other partition would be this new proposal where the market and social 
governance can control the way the Internet shares out resources.

At a grander level, the choice of how much resource each of the two 
partitions gets would be under the direct control of ISPs, but in turn 
their choices are also affected by the market and perhaps by social control 
through government regulation.

I know this fits your preferred model of how new resource allocation 
strategies should be added to the Internet. The question is whether it is 
sufficiently important to use IP header space to provide this choice. We 
don't need to decide that now, but we do need to debate it widely - more 
widely than tsvwg.


Bob

At 18:16 24/07/2007, Sally Floyd wrote:
>>3/ User pricing: Flat monthly (in our simplest effective proposal).
>>
>>Please don't imply we're advocating user congestion pricing, given the 
>>whole reason we brought re-ECN to the IETF back in 2005 was because we 
>>found a way to limit congestion without constraining how operators price 
>>their services to their customers. Specifically, re-ECN is designed to 
>>work with flat monthly pricing.
>
>This draft is not about re-ECN.  It is about the usefulness of
>"simple best-effort traffic", and the usefulness of flow-rate fairness
>for simple best-effort traffic.
>
>This draft is not about the presence or absence of flat monthly
>pricing for users.  One could have flat monthly pricing with no
>other limitations, one could have flat monthly pricing with volume
>limits, or one could have flat monthly pricing with some form of
>"you get what you pay for".  The first two forms don't result in
>"rich" users getting all of the bandwidth in times of high congestion
>(e.g., after the earthquake).  The third form easily could.  E.g.,
>if all of the traffic was "you get what you pay for", even with
>flat monthly fees, then it could be really expensive to send packets
>in the time interval after the earthquake, or some other highly-congested
>period, and only the rich would be able to send packets.  I would
>be opposed to setting such a "you get what you pay for" goal as the
>overriding goal for the global Internet, with or without flat monthly
>fees.

I know. It's distasteful to me to (I am a life-long socialist). But the 
alternatives are worse. I have thought about this long and hard.

On the demand side: If the rich don't get to be able to buy the right to 
use capacity, the alternative is that the malicious and selfish get to 
drive out those trying to be sociable in their use of the Internet. And 
that's for always, not just during earthquakes. Also, the lack of any 
resource sharing arbitration will lead to a vacuum that will be filled by 
the rich and powerful (which is what deep packet inspection represents - 
those with power and influence controlling what can and can't be done on 
the Internet).

On the supply side: The malicious and selfish will use the Internet 
excessively, but not pay their share, so there won't be sufficient 
incentives for anyone to invest in more capacity for those of us who want 
to use the Internet responsibly.

In summary, a knee-jerk reaction against a market mechanism needs to be 
tempered by the thought of what will otherwise fill the vacuum - the lesson 
of the Soviet Union?

However, you will notice in my draft on fairness (and in my design of 
re-ECN) that I have been careful to enable more equitable resource 
allocation schemes locally. This can give _true_ equality, whereas flow 
rate equality is just a license for the selfish and malicious to take what 
they want.

Overall, I put my hands up and admit it: we don't have a better system for 
allocating resources on planet scale than a market. I wish we did. We do 
have better resource sharing capabilities locally though, and I want to 
enable them.


Bob


>- Sally
>http://www.icir.org/floyd/

____________________________________________________________________________
Notice: This contribution is the personal view of the author and does not 
necessarily reflect the technical nor commercial direction of BT plc.
____________________________________________________________________________
Bob Briscoe,                           Networks Research Centre, BT Research
B54/77 Adastral Park,Martlesham Heath,Ipswich,IP5 3RE,UK.    [phone]