Speaking from Odessa

Odessa sits on the Ukrainian edge of the Black Sea. It is a town that, like much of the former Soviet Union, last saw great prosperity before World War 1. The twentieth century was a catastrophe in these parts,

communism being no friend to economic development, despite its claims, and war of the kind fought  between Nazis and Communists devastating to all life and proper values. Odessa’s downtown consists of elegant three and four storey buildings with Imperial facades. The downtown streets are cobblestone. The sidewalks likewise are in stone and brick, with this qualification, that the idea of a smooth continuous surface along the length of a street seems not to be the practice. Look where you are going is the guiding principle. Unguarded pits are there for the unwary.

My hotel is the “Bristol”, an elegant rococo pile whose refinishing kept all the late nineteenth century décor and room arrangements and backfilled them with modern conveniences: bathrooms, TV, Wi-fi. Outside the plush environs of the hotel, society is rapidly transforming from poverty to modern scramble: cellphones, electronically verified transactions, food stores with liquor and fresh vegetables, and the flourishing of a service culture where service is an industry and not a feudal condition.

The occasion is an Internet conference sponsored by RIPE, the European assigner of IP addresses, which sponsors Internet education in its regions. I am here to give a speech on Internet exchange points, to which subject I will get in a moment. The charming aspect is that the conference (ENOG 10) is the product of cooperation between Russians and Ukrainians, whose countries are supposedly at low-level war. In fact, I noted from newspaper reports that as soon as Tsar Vladimir Putin saw that the Ukrainian parliament passed laws allowing linguistic rights to Russian speakers in the eastern (Russian-speaking) parts of Ukraine, hostilities settled down. This was before his Syrian intervention.

On the topic of Internet traffic exchanges, there is an excellent paper by Dennis Weller and Bill Woodcock for the OECD, Internet Traffic Exchange: Market Developments and Policy Challenges.

Two vital and large facts emerged from their study. The first is that

The performance of the Internet market model contrasts sharply with that of traditional regulated forms of voice traffic exchange. If the price of Internet transit were stated in the form of an equivalent voice minute rate, it would be about USD 0.0000008 per minute—five orders of magnitude lower than typical voice rates. This is a remarkable and under-recognized endorsement of the multi-stakeholder, market-driven nature of the Internet.

The second major finding was that almost all of the Internet’s exchange points work on what is a handshake deal.

A survey of 142,000 peering agreements conducted for this report shows that the terms and conditions of the Internet interconnection model are so generally agreed upon that 99.5% of interconnection agreements are concluded without a written contract. That these “rules of the game” are so ubiquitous and serviceable indicates a degree of public unanimity that an external regulator would be hard-pressed to create. The parties to these agreements include not only Internet backbone, access, and content distribution networks, but also universities, NGOs, branches of government, individuals, businesses and enterprises of all sorts—a universality of the constituents of the Internet that extends far beyond the reach of any regulatory body’s influence.

In Bill Woodcock’s mind, the value in the Internet is created at the Internet exchange point (IXP). Imagine an old movie where ten telephones cover the desk of the business executive. The comedy of those old routines lay in the businessman not knowing which telephone was ringing or what telephone carried which of the three concurrent conversations he was engaged in. Now suppose a telephone exchange unites the ten thousand people connected to one telephone with the ten thousand people connected to the second telephone. The number of telephones on the desk is reduced by one, and the number of telephone directories needed is likewise reduced by one. The utility of the network to each subscriber has climbed. And where does this increase of nearly metaphysical value occur? At the exchange point. The more participants in the exchange point, the fewer the telephones on the desk (to mix the metaphor).

The job of the Internet exchange point is exactly similar. The cost and benefit of the IXP is measured in more pragmatic terms by reductions of the costs paid for transit (carriage across a network) versus the costs of establishing the exchange point. If the former exceed the latter, then an exchange point should be established. There are some reasonably precise calculations that can be done to show when traffic can be more cheaply exchanged than transited.

As Bill Woodcock explains:

Netflow provides a count of the volume of traffic headed to and from each other network on the Internet.

If you combine Netflow data with routing table data (or even just ASNs [autonomous system numbers]) from an IXP, you can tell how much traffic you should be able to deliver (or receive) at that IXP.

By comparing the current cost of sending and receiving that traffic by the present methods (some or all of which may be over expensive transit) to the projected cost of delivering the traffic at that IXP, you can estimate the effect of participating at that IXP on your overall Average Per-Bit Delivery Cost.

So, if you’re a well-instrumented and self-aware network operator, this is how you make your decision to build out to a new IXP.”

This brings me to a subject way beyond the reach of my abilities, the value of growth in networks. There is a dispute among the mathematicians (or maybe it is a dispute among the investors) about the value to be attributed to increases in the size of networks. The first ‘law” is attributed to Bob Metcalfe, creator of Ethernet. He is said to have remarked that the value of a network increases roughly with the square of the number of its participants (Metcalfe’s law). As huge amounts of investment money have turned on such proclamations, it is worth examining the question. How much value is added by the addition of members to a network?

Several mathematicians took issue with this. Bob Briscoe, Andrew Odlyzko and Benjamin Tilly wrote an article in the IEEE Spectrum in 2006, called “Metcalfe’s Law is Wrong.”

Like Moore’s Law, which states that the number of transistors on a chip will double every 18 to 20 months, Metcalfe’s Law is a rough empirical description, not an immutable physical law….

The foundation of his eponymous law is the observation that in a communications network with n members, each can make ( n –1) connections with other participants. If all those connections are equally valuable–and this is the big “if” as far as we are concerned–the total value of the network is proportional to n ( n –1), that is, roughly, n2. So if, for example, a network has 10 members, there are 90 different possible connections that one member can make to another. If the network doubles in size, to 20, the number of connections doesn’t merely double, to 180, it grows to 380–it roughly quadruples, in other words.

We propose, instead, that the value of a network of size n grows in proportion to n log(n). Note that these laws are growth laws, which means they cannot predict the value of a network from its size alone. But if we already know its valuation at one particular size, we can estimate its value at any future size, all other factors being equal.

Imagine a network of 100 000 members that we know brings in $1 million. We have to know this starting point in advance–none of the laws can help here, as they tell us only about growth. So if the network doubles its membership to 200 000, Metcalfe’s Law says its value grows by (200 0002/100 0002) times, quadrupling to $4 million, whereas the n log( n ) law says its value grows by 200 000 log(200 000)/100 000 log(100 000) times to only $2.1 million. In both cases, the network’s growth in value more than doubles, still outpacing the growth in members, but the one is a much more modest growth than the other. In our view, much of the difference between the artificial values of the dot-com era and the genuine value created by the Internet can be explained by the difference between the Metcalfe-fueled optimism of n2 and the more sober reality of n log( n ).

The authors of the article – which I recommend as inoculation against faddish investment fevers – also warn that:

We admit that our n log( n ) valuation of a communications network oversimplifies the complicated question of what creates value in a network; in particular, it doesn’t quantify the factors that subtract from the value of a growing network, such as an increase in spam e-mail. Our valuation cannot be proved, in the sense of a deductive argument from first principles. But if we search for a cogent description of a network’s value, then n log( n ) appears to be the best choice. Not only is it supported by several quantitative arguments, but it fits in with observed developments in the economy. The n log( n ) valuation for a network provides a rough-and-ready description of the dynamics that led to the disappointingly slow growth in the value of dot com companies. On the other hand, because this growth is faster than the linear growth of Sarnoff’s Law, it helps explain the occasional dot-com successes we have seen.


Back to the issue of exchange points: the great value of Weller and Woodcock’s research paper, to my mind, is the emphasis they place on the fact that the Internet is a product of commercial, and not just technical, innovation.

The evidence, they say, points to the ability of people in all economies to self-organize Internet exchange points, the effect of which is to decrease dramatically the costs of communication. The moral of the story is that we ought not to emulate the highly regulated, inefficient and politicized markets that have characterized treaty-based voice telephone exchange, as overseen by the ITU.

Five orders of magnitude (between 100,000 and 1,000,000 times) cheaper is the kind of number that regulators and politicians ought to understand. Do they? For this is the great success of the Internet, which many countries do not appreciate.


To top