A Practical Navigator for the Internet Economy

The Metamorphosis of Level 3

How Level 3 Went from an Inter City Long Haul Fiber Network to a Metro Based End-to-End Alternative toVerizon and ATT for Enterprises

The Internet as an Ecology of Connectivity and special issue on Level 3. How to purchase this issue. $350 or $1400 group.

June 18, 2007 Ewing, NJ -- The August issue contains an introductory essay, my two part analysis of Level 3, an interview with Robert Feiuerstein, and a shortened version of Level 3's March 14 2007 Analyst Conference.

Executive Summary

Internet as Ecosystem pp 1- 4

In a discussion of Eric Raymond’s the Cathedral and the Bazaar, the Internet is seen to be neither cathedral nor bazaar but rather an ecosystem.

1. Who was really in control of Linux? Those who did all of the assembling - i.e. the "top" in Carr's view - or those, who at their leisure - on their own time and terms - perfected a particular piece of code? Who was in service of whom? It was not one-way.


2. This leads to the observation that the Internet is far more like nature - it is an ecosystem. Meta - parts and micro subparts develop on their own time to suit their own sets of needs. Unlike "natural" ecosystems, however, the Internet compresses time and space; in so doing it also increases the number of participants. This enhances its multidimensional complexity.

But to observe that the Internet is complex actually misses the point: it is highly adaptable & highly flexible: it enables. The output is incredibly rich and complex, but the process is simple ­ tools, connection, collaboration, intelligence, information ­ all are pushed to the edge. And the edge gets to make its own decisions about where and how to make connections across all available dimensions.

Metamorphosis of Level 3 -pp. 5-7

Level 3 has managed to change from an intercity carrier vulnerable to Qwest attacks into an end-to -end IP carrier connecting upwards of 7000 enterprise buildings in 120 metropolitan areas. This accomplishment should mean that the ILECs, unable to kill Level 3, will have to co-exist with it. in the future.

Level 3 has transformed itself from an intercity fiber carrier into an end-to-end solution for customers in about 7,00 on net buildings in 120 metro-areas in the USA. In 2004 Qwest began to test its Level 3 interconnection agreement in front of the PUC in its 14 state area. Three years later Qwest still attacks Level 3 and has had some success in getting state PUCs to require Level 3 to pay increasing fees to Qwest.

Level 3 is truly end-to-end for nearly 7,000 buildings that are attached to its network. At its May 2007 stockholders meeting it announced plans to add 700 to 1000 buildings per year ­ saying that 100,000 buildings were within 500 feet of its fiber optic cable paths. As long as Level 3 completes a successful integration of its acquisitions, and this integration is almost finished, what Level 3 has done after Qwest started to try to bleed it to death, looks to be a successful and historic accomplishment.

The COOK Report has long advocated the creation of an open access fiber network at community and city levels. What Level 3 has done is to create a National end­to-end IP based fiber network that as Bob Feuerstein points out in the interview that follows is open to a wide range of businesses and applications.

Robert Feuerstein Interview - pp. 8-19

The interview updates a 2005 talk entitled “A Commercial View of Optical Networking 2005-2010. Robert explains many often misunderstood issues about the use and capability of the technology. For example he touches on issues such as SONET versus mesh protection schemes that can affect network build and reliability issues if the are not wisely dealt with.

He observes the following: “The emphasis within Level 3 these days is on selling lit services and providing excellent services to our customers. Now back in 1998 -2000 time frame we did sell lot of dark fiber. Between 2002 and 2004 we got back almost all of that dark fiber. People found out that running a big dark fiber network on their own is a difficult, complicated, expensive task.”

He gives a detailed over view of Internet2’s net Infinera based network and a very useful overview of how Infinera’s ability to do OEO regereration in expensively enables Level 3 to avoid ultra-long haul problems an have more interconnect points.

Summing up strategy: COOK Report - If you are going to be vertically integrated, you have to do so at some level of scale to make it work. The metro assets are the foundation for the whole strategy because you can then do everything end to end with your own customers and without the interference of another carrier, which is the whole point of having done an IP network to begin with. Right? Feuerstein: Yes. You got it absolutely Right.

Level 3 March 14 Analyst Meeting - pp. 20-47

Kevin O’Hara: we believe that the appropriate question as we go forward isn’t what’s happening to price or demand but what’s happening to the relationship between price and demand. We say that because from the beginning of Level 3 we’ve under the hypothesis that the communications industry is moving towards a technology business model and away from the traditional utility business.

Later: The cost to move a DVDs worth of information over an IP network is about an order of magnitude or less than the physical alternative of shipping it through the mail or going down to the neighborhood store and renting one.

. . . The metro company strategy has been a little bit different. We believe as one measures these things the gross margin is not an accident, it’s not an independent variable. Gross margin is a direct result of what you sell, whom you sell it to and where you sell it particularly if you can sell it on net. We made the metropolitan acquisitions because we believe it gives us a significant strategic advantage.

. . . . Jack Waters: When we speak of the Level 3 inter city network, we really mean the network that interconnects our metro networks. This network is comprised of the original Level 3 network plus Wiltell, Broadwing, the Telco inner city assets and the Progress Telecom inner city assets. The network has grown to 47,000 route miles which is a proforma number after the integrations. Today the number is actually larger and is really over 70,000 route miles. Based on our European expansion we serve 16 countries. We continue to believe that this network is the most scalable in the industry with the breadth and depth that is unmatched.

When many people think of Level 3 they think of the inter city network and, while the inner city network is incredibly important to us, it’s our metro facilities that are really one of the greatest sources of our competitive advantage. These networks allow us to have such superior growth margins because they let us get closer to our customers without using a third party.

We serve a total of 125 metro markets with 25,000 fiber route miles as the largest competitive footprint in the US. Today connects to 6,500 aggregation points, cable head in, wireless switching centers, internet traffic exchanges and enterprise buildings and these metro networks have extensive reach and are within 500 feet of well over 100,000 incremental enterprise buildings. Our goal is to add 750 to around 1,000 building in 2007 and we expect that sort of pace to accelerate for many years to come. . . . .

From 1998 to around 2004 we saw incremental, intra-platform efficiencies.

Then in 2005 a new technology became available. Photonic integrated circuits or optics on that chip which significantly reduced our cost to move information. Now this technology is very similar to the shift from discreet transistors to microprocessors. It is essentially taking a whole bunch of discreet optical components and placing them on a chip. By doing this, our cost for new optical gear is dropped by over 70%

. . . . we like Ethernet a whole lot and it really boils down to the economics it gives us. Ethernet is a pretty widely understood technical term but the reason why it’s so attractive to us is it really means a broad supply chain, it serves telecom carriers, data centers, storage companies, enterprises and even consumers. What that means is the demand for Ethernet technology is huge and so is the market. To us that means we’re one in 10,000 buyers of technology rather than one in a handful of buyers of telecom specific equipment. And the impact of the technology quite frankly is stunning because of the fact we’ve been able to take advantage of the Ethernet we’ve been able to drop our cost for core IP network technology by over 90 percent.

Sureel Choksi: We define the wholesale market broadly to include about 1,000 companies across a number of industries including cable, wireless, carriers, voice service providers, the federal government and number of large tier 1 systems integrated. We estimate that our customers spend approximately 75 billion dollars a year on communication services of which we target about a third or 25 billion from a service or geography perspective

. . . . . Jack mentioned that the Level 3 network is within 500 feet of 100,000 enterprise buildings that does not include about 20,000 wireless towers where the Level 3 network is in close proximity and as data applications over wireless networks continue to grow the resulting tower backhaul requirements grow as well. Given our deep metro facilities we think that represents a future opportunity for us as well.

Jim Crowe: However, it is the traditional Telcos though that really have the power if it [network neutrality] moves to the FCC. Who’ve got 100 years of experience of participating in a reasonable fraction of that period competing with the RBOCs and I can tell you they have plenty of lawyers, plenty of lobbyist and if you want to move to their turf, you want to give them an unfair advantage let the FCC start to arbitrate that neutrality. Far better to let a market sort through the issue, far better to let customer preference decide the issue. Today it’s becoming increasingly clear that it is socially unacceptable to differentiate between packets; let’s hope that remains the case. But if not, far better to have recourse to the anti-trust courts than to the FCC.

. .. . Sunit mentioned our customer concentration trends are improving. I thought I’d give you some specifics because this is a question again we get quite often that is are customer concentration issues going to be a problem for Level 3. Well including the SBC contract, that’s the contract that will go away with accordance with the figures that Sunit provided, including that contract we expect our customer concentration that is the top 10 customers to decline from 45% of total revenue, . . . - we would expect that to decline to 30% of total revenues of the fourth quarter of this year.

. . . Look at today’s communications market. We have portal companies, we have application companies, we have search companies, we have advertising oriented companies, we have software as a service, we have increasing numbers of social networking companies. All of those are the application communication companies of 2007.

Traditional communications companies are struggling to keep up with the pace of change of that development. No one thinks that traditional communication companies will dominate in those spaces. They won’t dominate of course in hosting or the network intelligence layers. No one believes that we’re going to see domination by the traditional companies. The real questions remain of course in access and backbone. Will access and backbone comply? And we’d simply say we’re already doing that. That is our strategic approach. We want to develop the first end to end company with extensive national metro facilities and a backbone network connecting. If we’re able to then all of those other organizations represented in the blue box there, the content providers, the cable providers, the wireless providers, government, enterprise those will all be markets for us and they’re far bigger than any reduction that we might see because a few large companies move some of their traffic to their own facilities.

Level 3 Concluding Analysis - 48

No longer is it predominantly an inter city fiber carrier inviting customers to come to peering points to hop on its backbone. For enterprises it is now a viable and desirable alternative to ATT, Qwest and Verizon. Closing in on 7500 “on net” buildings, it can give customers there cost effective end-to-end IP service. It has eliminated the last mile bottle neck by means of which all three incumbents would like to strangle it.

It has also organized into business divisions focusing on all the reasons that an enterprise, university, or government agency would like to buy broadband. It has found a business model that can stand up to the LEC incumbency and its business model and infrastructure is the best reason for hope that ATT will not be able to kill the internet in the US and stifle completely the development of the infrastructure we need so badly.

However control of what information flow I have observed from my interactions with the company over the past eighteen months is quite centralized. There is much focus on image and uniformity of message given to the public.

I wager that those thought leaders who understand the Internet would applaud what Level 3 has done and would help it get its story out to every wider circles, if only Level 3 would encourage its own people to have more contact with the outside world.

For the complete issue you must subscribe.


The Internet as an Ecology of Connectivity

Why Words often fail to Explain this Internet “Thing”

Understanding and Explaining What the Internet Enables Still a Difficult Task

The Internet is an Ecosystem p. 2

All Perception is Relative p. 3

The Internet is an Ecosystem of Connectivity p. 4

The Metamorphosis of Level 3

An Introduction to this Special Issue on Level 3 by Gordon Cook

Qwest versus Level 3 p. 5

End to End for the Enterprise Market p. 6

So What Does Another Analyst Say? p. 7

A Commercial View of Optical Networking 2005-2010

A 2007 Update from Robert Feuerstein, Level 3

Purchasing Drivers p. 8

Running Your Own Dark Fiber Net Not Easy p. 9

Internet2’s New Network p. 11

Protected Service ­ It’s Coming p. 11

Layer One Mesh: Approach with Care p. 13

Wavelengths on Demand p. 14

Pricing Trends p. 15

Ultra Long Haul Transmission p. 16

Structural Issues in the Physical Network p. 17

Vertical Integration as Winning Strategy p, 18

Level 3’s March 14, 2007 Analyst Meeting

An Impressive Display of Transition from an Intercity Provider to an End-to-End IP Network Solution

Market and Strategy: Kevin O’Hara, President and Chief Operating Officer p. 20

Disintermediation as a Source of Bandwidth Demand p. 22

Size of Network p, 25

We Are a Logical Consolidator p. 26

Integrating the Acquisitions p. 27

Network and Services Jack Waters, CTO p. 28

Infinera and the new Photonic Integrated Circuit Technology p. 30

Ethernet p. 30

Video Content Distribution p. 32

Services p. 33

Wholesale Markets Group, Sureel Choksi, President p. 34

Strategic Value of Building Location p. 35

Business Markets Group: Raouff Abdel, President p. 37

European Markets Group Brady Rafuse p. 38

Content Markets Group p. 40

Why Customers and Solutions Focus Is More Important than a Product Focus p. 41

Jim Crowe: Summation p. 42

Appendices p. 45

Level 3 – My Concluding Analysis

Level 3 Simply Doesn’t Get Its Story Out p. 48

Working In a Condition of Vulnerability in an Open World p. 49

Some Recommendations p. 50

Postscript p. 51

Executive Summary. p. 52

There is no symposium discussion and hence no contributors for this issue

For the complete issue you must subscribe.