Viewing the Internet as an Economic System

With the Approaching End of Routable IPv4 a New Analysis Richly informs Policy Makers of Consequences of Inaction

Executive Summary

An Analysis of IPv4 as a Medium of

September 30, Ewing, NJ -- The Internet will run out of routable IPv4 blocks within about two years.  A year ago  John Curran described in detail the need for a concerted migration effort lest large ISPs have to refuse to accept new customers or engage in policies that  would lead to expansion of the core backbone and fragmentation of the net.  John described  a situation of great complexity without agreement on the nature of the problem.  A year later not much has changed.  There is more discussion, a tiny increase in v6 while RIRs are acknowledging the inevitable and proposing policies that run the gamut from auctions of numbers to declaring missions complete and disbanding once blocks are  exhausted.

Tom Vest brings a very important analysis to these ongoing events that says that the crisis may be understood by acknowledging that the problem may be understood in terms of IPv4 being a medium of exchange with the Regional Internet registries playing the roles of central banks.  He shows why one set of policies leading to routing table bloat will bring inflation to the system while at the same time  a shortage of addresses will bring stagnation. While migration to v6 is the only long term  solution, Tom shows why the registries are a necessary mechanism for their member ISPs to use in managing the internet economy.

He says: “there is a finite quantity of actors and activities that may be supported by a finite MoE pool, and the system is likely to become increasingly difficult to sustain as you approach that limit. Also, unless you plan to simply stop growing your economy once you hit that upper limit, you need to think about what comes afterward – i.e., how to integrate a supplementary Medium of Exchange, or transition to an entirely new, less constrained one.”

Understanding that an isomorphism or one to one correspondence between the RISs and the worlds monetary economy exists allows the RIR’s to show their members how many many decades of economic policy making can inform the choices of those left to manage the current dilemma.  In Tom’s words every MoE system is an artifact of human design, and apparently every one is vulnerable to distortions and inefficiencies like inflation and deflation. Moreover, given its importance as the unifying feature of the overall economy, control of MoE creates conflicts of interest among competing participants within the system.

History reveals that every significant and enduring MoE system has been associated with the emergence of some kind of coordinating mechanism or maintainer, which functions as a check against inflationary and deflationary episodes and helps to eliminate or at least minimize the conflicts of interests that arise as a result of the centrality of the MoE mechanism itself. In conventional economic circumstances, where the competing participants are banks and similar financial institutions, two kinds of coordinating mechanisms have existed at different times and places: sovereign-backed central banks, and less frequently, privately-organized “banker’s clubs.”

In the Internet context, where the competing participants are routing service providers  -- ISPs -- this central mechanism started off as a basic ledger function administered by a highly respected, neutral individual, and ultimately evolved into what is most frequently called the “RIR system” – which possesses some elements in common with both central banks and “banker’s clubs.”   See his most important basic policy memo “Migration or Stagflation” at
While much of the discussion has been with Internet engineers, from discussing his research with him, it is apparent that efforts have to be aimed most directly at the higher level strategists who have the responsibility for maintaining this stuff as an economy something that the engineers do not have.  Acting on their own  the engineers won't do this. But without this kind of tutorial, the higher level strategists will never be aware that they should even think about it.

Open Access Fiber Definition? p. 26

We talk about open access fiber networks all the time but New Zealand’s Keith Davidson posed an especially astute way of  asking if anyone can point me to a working definition of an open access network, to which a country / economy, or a significant portion of a market has agreed to the definition. Other key elements:

- can a working definition be developed as public policy, or does it require legislative / regulatory intervention?

essential that the definition burrows down a level, defining open access at Layer 0, then at Layer 1, then Layer 2.

useful if anyone has ever developed a matrix or scorecard that could be used to measure the closeness, or distance from "open access" for various providers and their offered services.   It is a working definition at the network layers levels, from a competitive providers perspective that is required...

Coluccio: "it is a working definition at the network layers levels, from a competitive providers perspective that is required..."

At various times over the past thirty years, Open Networking has been expressed variously as comparably efficient interconnection in some position papers, or equal access in others, and other means of defining openness have been used.

The basic precept in most of these schemes, however, has been usually to allow the attachment by competitors onto incumbent providers' backbones or subscriber loop networks (and vice versa), no matter whether they were collocated physically in an incumbent provider's offices or they met each other at some other mutually agreed-upon meet-me point, as would be the case in a carrier hotel, or through some other 'virtual' method supported by fiber extension.

Goldstein: Regulate too much in the name of "open" or "neutral" networks and you break the market. There is literally no way to define (draw a clear static boundary) around "abuse"; it is a dynamic function. Publish a strict Rule of "abuse" and abusers will find a way to violate the spirit of the law within the letter of the law.

Role of Video, p. 32

Budde: Fortunately we do see that telcos can change.  BT in the UK and KPN in the Netherlands are the leading lights. There are also some glimpses of similar changes in for example France, Sweden and New Zealand. For totally different socio-cultural reasons and often based on totally different models developments in Japan, Korea, China and Taiwan are also far more advanced than those in the USA, Australia and some of the backwater regions of Europe. This indicates that the cat can be skinned in different ways - there is not one size fir all models.

But in all those markets we do see that wholesale is valued and once this gets further developed, we will then see content and service providers starting to use video as input and they will add value to it. In healthcare and education this will see money being spent by governments as well as by private health organisations (perhaps as much as 25% of all future broadband revenues). Utilities will pay for smart grid applications, media for entertainment and so on.

So if by "video" we "really" mean all rich multimedia files, I agree. But if we only mean machine-made series of pictures displayed one after the other to create the illusion of motion, I think we are missing much of real value, and of real quantity needing to be transferred.

Odlyzko: I fully agree, and such big files definitely have value, experiments show that while video telephony is not all that attractive, voice plus a good whiteboarding facility matters a lot to people.

Comcast Caps-
Roberts Comments  p. 38

On August 29th Comcast Corp, the largest U.S. cable operator, said on Thursday it will cap customers' Internet usage starting October 1, in a bid to ensure the best service for the vast majority

Goldstein: When bandwidth hit the several-megabit range, sharing movies became practical.

When bandwidth hits the 100 Mbps range, sharing HD movies will become practical. If the price is zero, people might download movies just to taste them, or collect them.

Yet there is no such thing as a free lunch still applies; "practical" to a user who has an "all you can eat" plan is not the same as "practical" to the provider responsible for providing that user with backbone connectivity.

Kenjiro Cho: My view on the Japanese situation is: - an average Japanese residential user consumes traffic by orders of magnitude more than an average American (simply because capacity is available). - 10% of heavy hitters consumes 80% of traffic, currently driven by p2p. but this is the nature of Internet traffic.   Even if we eliminated p2p, 10% of heavy hitters would still consume 90% of traffic for videos and other rich media content. - a bandwidth cap helps as a short-term workaround.

Now South Korea (from somewhat fuzzy estimates) and Hong Kong (from solid government statistics) do have much higher levels of traffic, about 6x the US level. But Japan, with its big pipes, does not. And that is a puzzle!

In general, I hear all these stories about TCP-P2P traffic filling all the pipes and requiring major changes in traffic management policies, and shake my head. There are simply very few cases of traffic growing explosively. As Kenjiro writes, in Japan traffic growth is below the rate of technology improvement. Now this may be because of those traffic caps, but if so, then perhaps the industry should rethink its policies, and start thinking of stimulating traffic growth!

P.S. Those 900 GB caps for NTT, or 250 GB for Comcast, are far above what the typical user is likely to reach. But they may have a big effect in stifling traffic. As I have documented in various papers going back a decade, and as Kenjiro mentions happens in Japan right now, even small usage pricing or other limits (especially if, as in the Comcast case, there is no simple answer.)

Growth of Data traffic, p. 44

Rood: Rollie,  I think you mix up two things in your response, and that the issue on (global) traffic statistics is more subtle.

If bandwidth is cheap and you throw it at a problem to compensate for electricity and location costs, or if. IT-staff is expensive so you outsource to an ASP or grid etc. the longer run effect would be that traffic would go larger distances and enter more backbones (or on the backbone companies would start to sell more wavelengths and dark fiber strands for private corporate networks from business facilities to the IT-facilities etc.)

What I have observed, when analysing FCC international traffic statistics from the 1995 - 2004 period is a different phenomenon.

In 1995 voice still dominated international capacity and the US traffic had it's strongest traffic relations with neighbouring countries (Canada, Mexico), a group of usual suspects in Europe (UK, France, Germany, Switzerland and the Netherlands), Japan in Asia and some immigrant countries (India, Phillipines scored very high too.  In 2004 Internet traffic had just surpassed voice traffic and international bandwidth for leased lines also had grown far beyond circuits for voice.

The picture for the US international circuit status from the FCC at that time was strikingly different. The largest capacity relations were with the UK, the Netherlands, Canada, Germany, Japan,
Sweden, but countries like India, despite all the hoopla about outsourcing did not advance that much in their ranking.

Predatory Society  p. 47

COOK Report:  Jamie Galbraith, son of John Kenneth has taken Veblen’s Liesure Class - added on his father ‘s New Industrial State and come out with a book that reads like and economics version of the Shock Doctrine. Unlike Naiomi Klien, Jamie Galbraith has some darn good solutions.

I see his world view as compatible with Carlotta Perez's whether she agrees of course is a whole other matter,

I blogged about it here

and here this entry contains the final pages of chapter 10.

In the last entry he makes the valid point that business that wants to go anywhere innovative in a technology sense cannot under a predator state and that progressive business has strong reason to oppose same.

I found Galbraith’s Chapter 10 - the rise of the predator state so compelling - it makes sense in my mind of what George Bush has done to us in the last 8 years - that I have used MacSpeach Dictate to do a virtual transcription of about 7500 words - about 90 percent of the chapter.  People do need to buy the entire book however.

Gigabit per Second  FTTH in Amsterdam  p. 52

Wagter: We (GNA, BBned, InterNLnet (ISP)) have just concluded a 3-day test of a 1 Gbps FttH connection for consumers over our network. The price levels of equipment are dropping to a level where 1 Gbps seems to become commercially viable, and the consumer equipment for fast and practical home networks (including set top boxes for full HD) is appearing on the market. So it seemed a good idea to do some testing to prepare ourselves.

Included in the test is something we expect to appear as a business model, more or less derived from Korea (web disks) and the fact that users on FttH networks in the Netherlands increasingly take a paid (20 euro/month or more) subscription to news servers. Apparently people are willing to pay for access to content, this indicates that a new business model is possible.


Viewing the Internet as an Economic System

With the Approaching End of Routable IPv4
a New Analysis Richly informs Policy
Makers of Consequences of Inaction

Part One – Background                                                  p. 1
The RIRs Formulate Transition Plans                             p. 2
A New Way of Thinking                                                 p. 3
IPv4 as the Internet's "Gold Standard"                           p. 4
What is the Goal of the Analysis?                                  p. 5

Part Two - IPv4 as a Medium of Exchange                     p. 6

Part Three - The “Systemic” Charts Explained              p. 10
System Population Utilization                                        p. 12
General System System-level Utility and Carrying Capacity    p. 14
Similar Dynamics of Both Systems Compared                   p. 16

Part Four - Migration or Stagflation? Policy Choices
and Challenges in the Coming IP Addressing Transition     p. 18

Why Policy Administrative Roles of RIRs Are
Similar to Functions Performed by Central Banks                p. 20

Conclusion                                                                          p. 23

Symposium Discussion August 17 - September 17

On Defining an Open Access Fiber
Network                                                                             p. 26
Net Neutrality Delusions                                                     p. 29

The Role of Video in Open Access
Networks                                                                            p. 32

Comcast Declares Bandwidth Cap - Larry
Roberts Comments from Japan                                           p.38

Internet Traffic is Growing Fast
but Capacity is Keeping Pace                                              p. 44

Commentary Predatory Society:
James K. Galbaith in a Perezian Context                            p.47

Test of Gigabit per Second FTTH in
Amsterdam                                                                         p. 52