On Sep 20, 2005, at 5:43 AM, Michael.Dillon@btradianz.com wrote:

...

I'm kind of surprised that I hadn't seem mention of it here before now, but Om Malik points out in his blog that Google is reviewing bids for it's natioal DWDM network:

http://gigaom.com/2005/09/19/google-asks-for-googlenet-bids/

There seems to be a trend whereby anyone who can aggregate sufficient traffic to warrant their own IP network is doing so and offloading the so-called public Internet. In the case of Google it is reminiscent of the way the television networks aggregated broadcast content way back in the 60's.

Ten years ago, the idea that there could be a public Internet which anyone could use for any purpose was rather new. Is this concept now on the decline?

--Michael Dillon

The chief tradeoff here is not "public" vs. "private", but rather returns to layer 1/2/3 transport vs. returns to content and end-user services. Barring regulatory impediments, every content provider that reaches a certain size ultimately concludes that outsourcing transport is suboptimal -- either too expensive, too variable, or too restrictive. The "certain size" that they have to reach is largely determined by the ratio of transport outsourcing costs (opex) to infrastructure ownership costs (capex), which vary differently with scale (security concerns, service type/vulnerability, and tax laws also affect the tipping point). Once ISPs cross that threshold they build, and once they build they start thinking about all of the other content and services that they might deliver with this new capability. AOL, MSN, Yahoo, and now Google are just a few obvious examples. It works the other way also. Sometimes changes in technology, telecom law, or tax policy cause the service/infrastructure ratio to flip, and you get infrastructure divestitures. The 1996 Telecom Act opened up the wholesale access segment to many more/new players, and dramatically reduced the costs and difficulty of reaching dial customers. That prompted many operators to enter the transport business, but others to divest. Arguably the whole Internet thing itself was a product of a similar flip in 1984, when the US "value- added services" sector was created by regulatory fiat and set apart from the regulated basic access telecom segment. Huge improvements in the cost and availability of transport over the past five years have resulted in a massive expansion of the infrastructure insourcing group (visible by the accelerating demand for ASNs). I think of this as a kind of democratization of control over transport. Did this trend undermine the "public Internet"? I don't really see how. No doubt the massive shift in US telecom law currently underway will also have a major effect on these patterns. However it won't be some idealized "public Internet" that will shrink as a result, but rather the segment of the Internet that is not directly controlled by incumbent, facilities-based access network owners. Whether or not this latest shift represent a good thing or a bad thing remains to be seen -- however the examples of other countries that are dominated by facilities-based telco-ISPs certainly doesn't not inspire much optimism. Tom