A Practical Navigator for the Internet Economy

The Paradox Of Commoditization

Trying To Save What Is Inevitably Lost, We Lose What Could Otherwise Be Gained

The monopoly control of customers by Legacy networks is destroying the economic benefits that could be obtained from the on going pervasive and inexorable commoditization of telecom and information technology. We face a paradox. While we have eyes, we cannot see.

We act as though we could wish away what is happening to new products and prices. But the fact is that the on-going commoditization of technology cannot be undone. Products will continue to get better but they will also continue to fall in price. In the face of these dynamics jobs will melt away. The only growth in the industry will be come from a variety of education, customer support, strategic evaluation and consulting positions. The only additional growth can come from use of the technologies in an open architecture that preserves the freedom to innovate.

 

If we adopt the mind set that commoditized telecom and IT is basic infrastructure, we can struggle to keep the infrastructure open. In doing so we shall also keep open a seedbed of new economic development and new creativity and technology. In addition to a foundation for new business, this course will also support opportunity for further growth in customer support and education. But, if, as seems currently more likely, we follow the course of permitting the legacy industry a closed monopoly in order to save itself, we prolong the current agony and forego what economic development and growth could flourish in an open environment.

The Commoditization of Everything

We are witnessing the commoditization of the entire industry. It is not just telecom. It is telecom and all of information technologies. Both industries are finally maturing across the board. While new products are appearing, they cost less and do more. They bring a different kind of economic value. We are no longer likely to see the creation of any new industry giant like a Cisco or Microsoft. Open in architecture and cheap to produce the new products are staggering the industry precisely because they are one or more orders of magnitude less expensive than the closed and proprietary systems they replace.

While most new products are still designed in North America, Japan, or Europe, the majority of their components are made and assembled in Asia. There RAM is a commodity endlessly replicated in multi billion dollar "fabs." Around the 'corner' in Taiwan and other areas, commodity disk drives are mass-produced. A terabyte in the pocket is not far off. Commodity open source Linux and open source web services stand ready to do battle against Microsoft's closed systems.

When new software is needed, it may be designed in North America or Europe. But the code is written in Bangalore, or Moscow, or Shanghai. Hua wei is sued by Cisco for doing what is in effect a commodity knock off. Back 'home" a handful of folk do the integration, first of the software, and then of the firmware and prototype hardware. They ship the result back to Bangkok or Kuala Lumpur for replication and assembly. Container ships bring the boxes back to ports like Yokahama, Newark, or Antwerp for sale on the shelves of Best Buy and Comp USA and other warehouse retailers. Prices are driven inexorably downward.

This new, cheap and powerful hardware and software is being installed in networks with fiber cores terminating in monopoly controlled copper local loops. If the network is the Stupid network, in other words if it is an open access end-to-end Internet, boxes running commodity Ethernet can switch and route packetized information as appropriate. With the addition of a VoIP gateway card, the same devices can achieve vast cost savings by turning voice telephony into an application that rides alongside other digitized and commoditized applications such as email, web services, and television.

But most networks are not open Internets. They are copper based networks with last miles subject to a monopoly controlled, content centric approach. They are bastions of legacy technology using an infrastructure far more complex and 10 to 100 times more expensive than that of the Stupid network. The legacy telco network is one where the monopoly must cut its own throat to try to compete with open architecture Internet upstarts that would take away its more profitable business customers. In other words, while the prices it receives for its services plummet, it must write down the value of its plant and equipment to a level where, given its smaller income, it cannot maintain its current cost and employment structure and sustain the ability to pay its debt.

When voice no longer rides on the TDM transport that was especially designed to carry it and is just a packet-encapsulated application on an IP network, the new central office is no longer a building housing five million dollars worth of equipment. It fits on a desktop using SIP, SIP proxy servers, and ENUM databases. It costs well under five thousand dollars and delivers an entire range of services not possible to derive from now obsolete TDM hardware costing a thousand times more.

Commoditization Makes More Job Losses Inevitable

If telecom in the United States has lost 500,000 jobs in the past three years, with the inevitable demise of the LECs, it will lose another 500,000. Attrition in computer hardware and software should cause 500,000 more positions to evaporate. What is left will be administrative, financial product planning and marketing. Unfortunately, in this brave new world the marketers will be figuring out how sell $100 products in Best Buys rather than $100,000 systems to enterprises.

As a result of this upheaval, customers will be left even more on their own. They will need help to figure out how to put the products together and assess what combination of products most effectively meets their needs. Until everything is truly plug-and-play and automatically-configured when attached to the network, education of the customer on product capability and system integration is the remaining critical area of competition. It is also the only bright spot for future industry employment.

Commoditization dictates competition. But now that the companies are in trouble, competition on the part of the legacy, monopoly-owned, circuit-switched side of the telecom business is being allowed to disappear. Until the legacy companies go bankrupt and swap out their obsolete equipment, there can be no benefit to anyone from commoditization.

The innovation and cost performance benefits of commoditization are all on the side of the open access, end-to-end, packet-based inter-networks. So far such networks in competition with each other for market share cannot make a profit.

The mind-set of the political and regulatory system cannot comprehend the resulting paradox where the most productive and advanced networks cannot make money because founded on commodity technology, they can be cheaply cloned with cookie cutter reliability. Staring into the headlights of the onrushing train wreck, it is blinded by fear of the destruction of shareholder equity and putting people out of work. It is seduced by the complaints of the incumbents who are selling the false premise: "Give us monopoly and we will have incentive to build." Determined to protect legacy interests, it consequently tilts the playing field in the US against the commodity players and against innovation.

The truth is that, even with monopoly, will they not build. They will instead die, unless somehow, they managed to get use of the packet-switched, commodity-based technology successfully outlawed.

To our great misfortune we do not yet understand that Commoditization has turned telecom and information technology into a basic enabling infrastructure like the electric, the water, the sewer and the highway grids. This new commoditized, powerful, and cheap technology can be used to deliver value and new jobs through preserving for everyone the ability to tinker and to innovate. Commoditization has removed the incentive for value creation from the monopoly network and left it with only the incentive to squeeze every penny of return for as long as it can prevent encroachment.

The Legacy networks can and do use the new IP, commodity technology. But they are prevented by their debt obligations from being able to acquire enough of it. Furthermore, even if they could implement it extensively, their business model assumes a monopoly over access to transport. Because they must do it all, they are paradoxically denied its fullest advantage. The find themselves with no choice but to use packet switching and VoIP in an effort to sustain their traditional way of doing things. Protecting what they have, they lose what gains the new inexpensive equipment could offer.

For example they would seek to offer international VoIP over their own dedicated network while a new competitor can dispense with the sunk cost of maintaining a physical network by simply renting access to an Internet that others maintain. Since the competition has only to rent access to transport and run voice as an application on that transport, it can offer service that is unencumbered by legacy costs. The business model of control of applications and customers prevents productive investment.

The paradox of commoditization leaves us with our uninhibited creativity as our only new source of economic development and growth. If all we do is drive prices down while maintaining the old structures, all we do is drive more people out of work. Failure to understand this leaves the legacy networks in power and able to preserve their obsolete assets by killing creativity, innovation, and experimentation.

The backyard tinkerer has for the past century been the principal source of wealth in the United States. In the inexorable transition to a new commodity-based world, given current policy, we are exporting the freedom to be a tinkerer to Canada, Sweden, Japan and Korea. Having to compete in a global economy, we are sacrificing the viability of our resulting economic infrastructure in a foolish attempt to shore up legacy networks that can no longer serve as adequate means of competition.

Lessons carried home from Spring 2003 Voice on the Network where most were upbeat and pitching in to spread the new application

April 7, 2003

Shortened Executive Summary

Find out how to order single copy ($400) or group license ($800) for just the April - June 2003 issue.

Choose Both the Technology and its Organizational Structure, pp. 1-4

Ascertaining how to build a new and sustainable telecommunications infrastructure is not just a matter of choosing a cost effective sustainable technology. Every bit as important is choosing and implementing the business model and organizational structure in which a technology will be operated. Voice over IP is, as this issue of the COOK Report demonstrates in great detail, a technology that is being implemented throughout the telecommunications world in very different ways. VoIP can be done in such a way with H.323 and Megaco as to attempt to preserve the vertically integrated control of the carriers. it can also be done with SIP in such a way as to move unprecedented control over their communications into the hands of users at the edge.

According to Googin's Law the network may be either a valuable monopoly or a valueless commodity. In VoIP the argument rests in part on whether of not VoIP via SIP is run over an application neutral transport layer. This is the most cost effective approach, but it is also on that will deprive companies of the ability to control their customers.

A regulated circuit switched telephony environment with little technology innovation meant that tariffs (prices) could be set with little regard for actual cost of operation. In this safe world the Internet becomes a Trojan Horse. It permits communication that is based on a totally different technology foundation. The combination of the availability of the Internet cloud and continued technology innovation means that the old world of different tariffs for voice and data is not sustainable. The cost of transmitting data is getting ever cheaper. The new seismic shock is now that voice and data are becoming from an economic point of view, the same. The impact on telco revenues will be major.

Analysts also had better begin to look at the impact of commoditization. If voice is just an application like email or web services, then an innovative company like an ITXC can compete against the incumbents by doing a VoIP overlay of the existing Internet rather than building an entire network from scratch. We received today (April 7, 2003) from Tsering Gayltsen Sherpa the first VoIP phone call from base camp at Mt. Everest. The future belongs to the users and not to the huge telco and cable cos. The only question is how long the transition to the user-provisioned future will take.

Technical Protocol Issues pp. 5 - 17

Schulzrinne: The important distinction is whether a network is application-neutral or not. SIP designers have taken the position that if you want QoS, security and the like - great, but these are not specific to voice services, multimedia services or even SIP-initiated services in general. To the extent that NGN represents a vertically integrated, application-specific network, it has nothing to do with protocols, but rather with a fundamental notion of what the Internet (or packet communications in general) should look like.

As a side note, I find the smart/dumb distinction less than illuminating, except that it captures the notion of an application-neutral transport mechanism. Even in an open network, providers can offer lots of 'smart' network services - after all, a full-fledged SIP proxy running scripts, Java servlets and CPL probably does more per-call processing than your average 4ESS or 5ESS switch. [Editor's Note: The Stupid Network has nothing against smart services as long as they are located in edge based devices under the control of end users.] NGN is not the next-generation Network, it is the Nothing-Goes-Network (except the set of applications blessed by a carrier).

Shockey: Thank you very much Henning that phrase completely encapsulates what I've been trying to describe to Gordon and many others as the carrier-centric vs application-centric framework.

I think we are getting to the crux of the matter here. The Internet was designed as a application neutral transport layer. Delivering real time voice is only an application. SIP was designed as a application. H.323 was designed as a service.

The stunning economic efficiencies of the application neutral transport (aka Internet) continue to push into the economic center of carriers business models. What concerns me is the anarchy this is clearly unleashing. Carriers' revenue models are totally dependent on the classic voice service. If that starts to collapse, as I believe we are now seeing, the first instances of the "Perfect Storm" of the last two years of the telecom industry become "Apocalypse Now".

Sip Characteristics, pp. 18 -22

You can decide to move that hosting service wherever you please because you have control of your DNS records. The same thing is true for VoIP. As long as you have control of your DNS records, you can point them to your SIP services. Move the services and changes the pointer. As you choose where you get web services you can also use your DNS to choose where and how you get your SIP services.

SIP a Key Component, p. 22

Shockey and Rosenberg: "SIP was designed with certain assumptions in mind. First was scalability: Since users could reside anywhere on the Internet, the protocol needed to work wide-area from day one. Users could be invited to lots of sessions, so the protocol needed to scale in both directions.

Carrier Walled Garden SIP, p. 25

Editor's Note: The carriers are well cognizant of SIP and are planning their own implementations. At the end of February it became clear that at least one of the ILECs - Verizon - was flirting with a walled garden use of SIP - sliced and diced in keeping with the company's fiduciary responsibilities to maintain as long as possible the ILEC's control over its customers.

QoS Not needed for VoIP, p. 29

On February 10 on NANOG, Woodcock: QoS is completely unnecessary for VoIP. Doesn't appear to make a bit of difference. Any relationship between the two is just FUD from people who've never used VoIP. Namely the folks who don't want end-users doing their own VoIP because it means lost revenue on circuit-switched networks.

Evslin & ITXC, p. 33

Evslin: With the old N squared minus one topology, demands of connecting every switch to every other switch your costs rise dramatically as the network increased in size. On the other hand, if I connect my switches to the Internet, I have only to connect them to a cloud rather than directly to every other switch on the face of the planet. I have an "n" cost instead of an "n squared" cost and I also have a simpler network. What many carriers also don't understand is that if two switches are connected to each other, the ports on the switch are dedicated to that connection. If it is nighttime in New York and London and no one is talking through those switches, those ports are wasted.

Alaskan VoIP, p. 40

Westlund: That is correct. Amstrong had a great vision. He is gone but we are in a position now where we can begin to implement what he failed at. That is to integrate cable properties and be able to use cable telephony to cover the last mile. We have been looking at multiple options including a PCS wireless license that give us a way to cover the last mile when a cataclysm hits.

We do voice over IP as part of the transport protocol within cable telephony.

Black Market VoIP, p. 44

A bright guy (multiple University degrees) with a large family discovered that he can buy some cell phone SIMS, some GSM radio gear, some VoIP gateways and some satellite gear and insert all the "local" calls he wants into the cell phone system of the his home country. Except that the calls originate from half way around the world. With a friend in New York and in London, they approach some carriers and offer to terminate 4 T1s of calling capacity to their city for $0.175 per minute. The carrier says, you pay for equipment and we will send you the minutes. They agree and the adventure begins. [snip]

At 20,000 minutes per day on each of 4 T1s, they see $10,000 per day gross margin. They immediately set aside $4,000 that will cover sales commissions and other costs in New York. But even $6,000 per day is a respectable figure by New York or by London standards.

Shortly the local crime-family discovers the illegal profits and they reach an agreement to pay some of their new-found wealth to protect themselves from being turned in to the local authorities

Legal By Pass in Europe, p. 46

In Europe it is perfectly legal for a company to go out and buy equipment that provides access to the cell phone system in order to provide by pass of some of the very high tariffs that apply when you call from wireline to cell phone and from one cell phone company to another. Even the carriers that are used by the cell phone companies are promoting the sale of these machines that provide this special access. [Snip]

Japan & Broadband (1) p. 50

It is indeed a victory of the unbundling policy of MPHPT [Ministry of Public Management, Home Affairs, Posts and Telecommunications]. The FCC's new "deregulation" policy is doomed because the ILECs will not deploy DSL in a way such that it cannibalizes their telephone business. We should ask why Japan's policy worked while the American 1996 Act resulted in the massacre of CLECs? [snip]

Japan Broadband (2), p. 56

COOK Report: Was Softbank the only player that really knew how to take advantage of the bargain price for copper and fiber in NTT's network?

Aizu: Softbank, or its subsidiary, Yahoo Broadband (BB) is by no means the only game in town. The ADSL game is quite intertwined with the game between the Ministry and NTT. This is still part of the long lasting war over the breakup of NTT. [snip]

NTT Fiber to the Home. p. 60

I live in the suburb of Yokohama, 30 miles south of Tokyo. My house was more than 3.2km (2 miles) away from the local exchange and the ADSL line( 8Mbps Service) connected only at 600kbps(down). So, it was a relatively easy decision my switching to FTTH although there was a 30,000 yen(U.S.$250) installation fee and my monthly fee would increase by 3000 yen (U.S.$25).

Broadband in Korea, p. 62

Aizu writes: "When it comes to broadband deployment, it is clear that East Asian countries, not northern European countries, are leading the world. Among them, Korea is by far the most advanced. As of April 2002, Korean broadband subscribers reached close to 8 million or 16.7% of the population and almost half of its households. Considering that Korean household Internet penetration was less than 5%, with 731,000 subscribers in 1996, the rapid growth of the Internet in general and of broadband in particular in Korea is quite significant." [Editor's Note: On February 23 Kilnam Chon added "As of December 2002, Korean broadband subscribers numbered more than 10 million or 22% of the population and around 75% of its households.]

FCC Feb 20 Decisions, p. 67

We are telling anyone who depends on the PSTN in this country to do his or her business that he or she is stuck with what the telco gives. Powell has given us privately owned toll roads instead of an open infrastructure of interstate high ways.

Googin's Law and the FCC Decision

Scott Berry (a telecom consultant based in Darien, CT): Let me take it one step further, and clarify my earlier comment: The only way a commercial enterprise will invest is if there is a monopoly. Without the monopoly, their investment becomes worthless (a commodity). There are only two ways to play the game *at low risk*--ensure a monopoly beforehand, or plan on a commodity. The former is the ILEC plan, and the latter is the Municipal Fiber model. [Snip.]

Computer II and III, p. 73

On February 25, 2003 Francois Menard wrote to Cybertelecom mail list: My understanding of what the FCC has released is that all that has been said to date means little yet to the imposition of computer I/II/II rules for access to the underlying infrastructure on a state-by-state level at a tariffed rate. I am being told that "just and reasonable rates" are however out of the window. Because of FCC preemption, ILECs would only have to file rates when they are non-dominant carriers (not subject to just and reasonable conditions).

Spectrum, p. 78

Editor's Introduction: Although thinking about spectrum is growing much more sophisticated than was the case in the old property rights regime of the eighties and early 90s, keeping it relevant to the performance of increasingly sophisticated technology is very difficult. One of the key fora for this task in the Open Spectrum mailing list founded last year by Larry Lessig, Yochai Benkler, David Reed and Dewayne Hendicks.

Edge Based v6, p. 84

Editor's Note: Our publication of an extensive discussion on the problems and prospects of IPv6 on February 3 kicked up a fuss. As a result of some 'letters to the Editor, we sent the 17 pages to Jim Bound who is Chair IPv6 Forum Technical Directorate and Chair North American IPv6 Task Force. We asked Jim to review and comment on it. On reading the material Jim found that by and large he agreed with what we had published. An unexpected dividend however was further discussion between Francois Menard and especially between Jim Bound and Bob Frankston. It became clear that Jim was not aware of Bob Frankston and David Reed's work in pushing the concept of Edge Based v6. (Maybe we should publish this under a Department of "Pot Stirring and Weaving?") It was quite fascinating to facilitate the further discussion below and observe the results. We had speculated in our last issue as to how the ideas of Edge v6 could be moved forward. The discussion below, that the participants have graciously consented to our publishing, shows some possible paths. [snip]

Evaluation of the Future, p. 87

With Moore's law and a host of associated corollaries preceding apace, the very investment requirements for telecom are being reshaped. WorldCom's physical switching infrastructure, acquired for 45 billion dollars five years ago, has just been revalued post bankruptcy at 10 billion. Investments of the size undertaken by WorldCom in a single huge company won't happen in the future. No one can afford them. What is not understood is that no one will need to be able to afford them. What is also not well understood is that we don't know how to evaluate the economic impact of the decisions being made. This article will illuminate the economic regulatory complexities of an industry caught in chaotic transition.

Roadmap for Decision Making, p. 89

COOK Report: The problem we face is that we don't know how to value the changes taking place in telecommunications nor can we adequately value the role of telecom in our economy as a whole.

Not only is telecom a huge and vital industry but because it is now such a basic part of our economic infrastructure, getting the answers right about telecom means not only getting a sick industry straightened out but it means getting something that is the foundation for our entire economy fixed. The basic economy that can be built on an end-to-end IP network is going to differ in significant way from what can be built on the archaic circuit-switched network.

The problems before us show not just the importance of a single industry. Instead at a meta level they impact the entire economy. In this context, the tools that we have to make judgments about the economic impact of the choices before us are too keyed to the top down central command model of a traditional circuit switched smart network infrastructure. We are asking how to protect the business models or profits of the incumbents, and we are very likely asking the wrong questions.

Paradox of Commoditization, p. 102

The monopoly control of customers by Legacy networks is destroying the economic benefits that could be obtained from the on going pervasive and inexorable commoditization of telecom and information technology. We face a paradox. While we have eyes, we cannot see.