PREVIOUS	January 2007 (15.10)	NEXT

Spikes Amid Flatness: a New Global
Topology Enabled by Components Standards

Under Influence of Components Standards Edge-Based Telecom Operation Spreads to Specialized Edge-Based Geographic Communities

November 30, 2006 Ewing, NJ --

Executive Summary

Introduction

The January issue begins with an essay summarizing five little known articles published by John Waclawsky in Business Communications review. Here John shows why the compnents standards process of the IETF, IEEE, WC3 has been able to create the Wealth enable by the Internet.

Components Versus Systems Standards

The introductory essay undertakes an exploration of the standards processes that serve either to enhance or degrade telecom’s economic value. It makes its case by summarizing and discussing five papers by John Waclawsky. The papers explain two different standards process that affect telecom: Systems standards processes and components standards. Systems standards are left over from the large-scale factories and investments of the industrial age that was required to make investments scale to meet the substantial costs of deployment of cumbersome large scale systems.

However the emergence of the integrated circuit helped to lower costs and enable the development of an architecture made up of loosely couple components made feasible a technology development process that is not bound to complex systems based telecommunications or the devices that enable these telecom systems. I refer to the components based standards processes that began in the 1980s with the growth of the early internet and as it also enabled cheap consumer electronics and computer based systems is now beginning to drive most economic development around the globe.

Indeed the only excuse for using non components standards is the protection of a market monopoly that has been rendered obsolete by better and more cost effective technology. Standards approaches may be likened to a plumbing system in which the valves can be arranged in such a way as to offer opportunities for new wealth creation or manipulated in a very different way to keep alive archaic and inefficient ways of doing old processes and make innovation too costly to survive.

All manner of open architecture, from PCs to the Internet, rest on the concept of components that can been refined and enhanced, as components, and therefore independently from the ways in which they can also be cobbled together to perform useful tasks. When a standard like Ethernet is not shackled to a complex system the way SONET is, we can see what a difference it makes. SONET was developed within the complex bell systems standards process. Having to interoperate in a tightly coupled system of extremely complex Bell network systems software, it came on to the market with a normal speed of well under a gigabit. SONET systems were very complex and had to undergo rigorous testing at Telcordia. When they shipped they carried multimillion dollar price tags.

Meanwhile, Ethernet as a component standard has had a thousand fold increase in performance over the past 15 years. The motivation was to deliver more benefit for the customer and make it faster and cheaper – a task at which the IEEE and a group of hardware manufacturers have been hugely successful. Vastly more cost effective components based gigabit and ten gigabit Ethernet dominates in IP network backbones.

This pattern has been repeated across the entire IT spectrum. Linked to open source software, it forms the edge-based network where anyone can plug in and innovate with a loosely coupled component that poses no threat of crashing the entire complex system. The disparity in wealth created by these different approaches is nicely documented in Figures three and four. Waclawsky’s detailed description of the contrasting processes and their results gives enormous insight into why the system standards based telcos simply cannot compete with the components based Internet leadership.

Fort Wayne Shows the Importance of Human Engineering Issues in the
Context of Municipal Fiber Builds

The Mayor, Graham Richard, looks at the city as a business chartered to give services to its citizens. It is expected to render these services in return for tax dollars with maximum cost effectiveness. The mayor has applied Six Sigma training, popularized as a way to make business processes more efficient, to city government. Strongly aware of the role of communications technology in determining a region’s global economic competitiveness, he is to the extent possible making broadband a part of Fort Wayne’s basic infrastructure. Finally he has hired an “I-teams coordinator” and charged her to look at city services from top to bottom and report on ways in which broadband can be applied to make them more cost effective and deliver more value to the citizens of the city. In this respect Fort Wayne has been more creative than any other city of which I am aware, and has a long list of innovative programs and approaches that other cities would do well to consider.

Quality of Service Difficult to Define and Even More
Difficult to Achieve

We had very vigorous discussion over QoS which turns out to be often hard to define let alone hard to agree on what uses are appropriate when.

Bill St Arnaud gives a very good definition: QoS is fundamentally a tool for the network operator to use in order to deliver explicit and/or implicit service guarantees with respect to different types of traffic (and customers) in the advent of congestion.

To some extent the QoS universe seems to divide along the lines of those who are telco oriented, selling bandwidth with usage caps, measuring and metering service and with limited available bandwidth, trying to give priority to voice or video links. Those selling scarcity are more likely to be wrapped up with QoS issues than are those who concentrate on raw gigabit or above kinds of Ethernet. But as with every attempt to look for rules, there are always exceptions. While over provisioning on a small network might look like the best solution, on a larger more complex network the amount of over provisioning that needs to be done to guarantee results would be extraordinarily expensive.

Those against are generally involved with data running pure IP networks and not phone companies. They generally do not bill for any kind of measured usage. On the symposium list John Waclawsky is the most prominent of the voices against. John takes a position that bandwidth can and should always be increased when link usage becomes crowded. John’s philosophy is to use priority queing only when increasing bandwidth on a rapidly growing link is not possible.

One finds that the introduction of QoS tools increases complexity in network construction and operation. Something that is already very complex. By increasing complexity QoS will also increase operational expense. The anti QoS folk maintain that the QoS tools that vendors sell are so complex as to require a PhD to operate them.

John adds a compelling point stating that, if networks are to be run reliably, they need redundant links so that if one or more links go down traffic will always be able to find alternative routes. However to for traffic to be able to fall over to an alternative route the link better not be run at more than 50% capacity. Protecting against multiple rout failures means that that links should be used even more sparingly. If links are kept lightly loaded, then they should not need extra QoS protection. John also is adamant that return on investment with QoS is negative. In other words QoS costs more to implement than can be saved by its use.

For Frank Coluccio the matter was not so much black versus white.

Frank makes a number of equally compelling points. He focuses on capacity planning and utilization issues in mission critical financial networks. Management in these networks must focus on handling growth. To do so it must do very frequent traffic measurement and it can run into situations where its ability to go out and buy more bandwidth can be restricted by management’s desires to put off network expenditures.

Editor: QoS is a tool that a service provider, depending on who that provider is, may or may not use to its strategic advantage. In an enterprise John would say that increasing bandwidth is almost always the first middle and end line of defense. Frank would suggest that even in an enterprise depending on managements objectives application of some QoS techniques would normally allow for a more balanced approach.

So You Want to Build you Own Fiber Network

In response to a request for information from a list member Frank Coluccio explains the different between conduit and ducts, choice of cables types of strands, installation of fiber by a facilities based provider or by ones own contractor, approximate cost per mile roughly a million dollars in a densely traveled corridor of northern NJ. Frank cites other source for buried conduit in rural, semi urban and urban areas

Susan Estrada introduces the concept of the middle mile bottleneck where Humboldt County in northern California has only one set of fiber coming into the county. “The middle mile project is to help reduce local 'first mile' costs because right now there is only one set of fibers that come into the area owned by AT&T on a relatively unstable path (mudslides, earthquakes, rock slides and tsunamis are all real threats.) As captives of this capacity, they are paying what they call "minibar rates" for services that other communities pay a heck of a lot less for.”

The county is looking at the costs of developing a competing pathway. Tom Hertz from Fiber Utilities gives figures for aerial fiber that range from 2 to 4 thousand dollars a mile in the Chicago area to $250 to 750 a mile in rural Colorado. New builds (construction only): $5/foot - aerial, $10/foot - buried, double those for metro/campus environments.

Bill St. Arnaud cautions readers not to forget support structure costs. He says: often the biggest cost of deploying fiber has nothing to do with the fiber installation itself - but support structure restoration. This can create the biggest variance in costs. Electrical utilities love to get their poles replaced by fiber builders - using the excuse that the weight of the fiber will over stress the pole, which 99% of the time is horse manure.

Frankston’s Connectivity and Gilder’s Fibersphere

Looking at California’s plans for public private infrastructure Bob Frankston comments that: there seems to be a disconnect in a state that lets everyone use billion dollar highways without paying any direct fee yet attempts to minimize the value of a few million dollars of connectivity by charging for each trip. Frank Collucio adds that the need to bring adequate public oversight to such huge task mirrored in many levels of complexity is especially daunting.

Frank then goes on to comment on Bob’s ideas about “connectivity’ as being similar to Gilder’s 1992 description of a fibersphere. George explained that every end user would possess a strand of fiber that was run to a central connection point where they were all fused into a central, optical core.

For short distances, passing optical signals in this manner through the fibersphere works out very nicely. For longer distances, however, they don’t work out so well, until at one point they don’t work at all. Were it not for the physics governing power dissipation through media, the fibersphere would have been just the thing you’ve been looking for.

Addressing could be handled by parties agreeing to connect at a certain wavelength or color. Scalability however is a problem. Optical spread spectrum or optical CDMA was developed in 2001 by a company call Codestream, but with the carrier business model dependent on selling scarcity not much progress was made.

Bob responds that he emphasizes connectivity as a question of reframing the issue- speaking of it in non telco terms. “The emphasis on a connectivity infrastructure from the edge: · It's a fresh start and thus we have an opportunity to define it · It emphasizes the connectivity not the particular means · It emphasizes the commons rather than the slivers. I want to connect to you not necessarily a web site. It is infrastructure and we share that. It emphasizes the pervasive aspect. It's not something that one has to subscribe to and we don't worry about up take since it is used by the community itself not just a few viewers. It is from the edge - we run our own wires and connect to our friends and so on. It's not a service, it's opportunity.”

Futility of the FCC Definition of Broadband

The premise for this dsicussion, as stated by Andrew Odlyzko, is - Governments should build long-haul and metro fiber and conduits, and rent them out to companies, which could experiment with various methods of solving the first mile problem of connecting homes to that fiber.

The reality is that more often than not government that wants its own infrastructure is forced both by ideology and lack of cash to declare itself in favor of a public private sponsorship in return for which all to often the incumbent controls the result.

Unbundling of physical networks is felt to be the wave of the future but no one can explain how we will get from here to “there”. For the most part the political will is simply not there.

What Is Keeping Pots Alive?

Farooq Butt hypothesizes that POTS is part of the social fabric and it's very hard to pull the plug and drop out totally. He speculates that the apparent staying power of POTs is due to ita rock solid reliability, the comfort of safety, a local stake in the network, utter simplicity of usage and billing, strong privacy, assured ability to connect and the warmth of known brands?

Waclawsky: I think POTS is all about OLD money restrictive voice stuff, not data. The Internet is all about data and NEW money. Yes, they are two different networks - different worlds and different requirements. But the Telco's have been believing their exaggerated self-importance with these similar arguments and they have missed all the wealth generated by the Internet.

Complexity isn't a network problem. It is a device selection problem for the end-user (we can make simple devices to make phone calls, can't we?). Our children don't see POTS through rose colored glasses.

Does everyone on this list honestly believe the 100+ year old Telco's have had the ultimate answer for location, privacy, pervasiveness,....etc, and simplicity via a framework anchored by a rotary phone?

Editor: The answer seems to be complacency and inertia on the part of many of the users and the device makers who in turn are captive to their carrier sales channels.

For the rest of the issue you will have to subscribe.

Contents

Spikes Amid Flatness: a New Global Topology Enabled by Components Standards

Under Influence of Components Standards Edge-Based Telecom Operation Spreads to Specialized Edge-Based Geographic Communities

An Exploration of the Standards Processes that Enhance
Telecom’s Economic Value p. 2
DARPA and Then the NSFnet Pry Open the Closed Telco Universe p. 3

Understanding the Creation of Wealth in Information Technology

Components Standards Versus Systems Standards p. 5

Market Commoditization: Enemy or Friend? p. 8
A Fundamental Question p. 10
What the Systems Standards Groups Have Done p. 11
IMS - a Further Effort to Maintain the Closed System p. 11
Systems Standards Goal is to Control Customers Once Again p. 12
P2P as a New Refuge for Innovation p. 13

Fort Wayne Applies Broadband to Deliver Better and More Cost Effective City Services

Mayor Graham Richard Demonstrates that Understanding Urban Constituent, Competitive and Operational Problems Comes Before the Application of Technology p.16

Symposium Discussion October 17 - November 17, 2006

Quality of Service Difficult to Define and Even More Difficult to Achieve p.30

So You Want to Build you Own Fiber Network?

Frankston’s Connectivity & Gilder’s Fibersphere p. 49

Futility of the FCC Definition of Broadband p. 54

Sidebar: Do Studies of Economic Impact of Broadband Networks
post installation Exist? p. 63

Executive Summary p. 67

Symposium & Interview Contributors to this Issue

Affiliation given for purposes of identification - views expressed are those of the contributors alone

Kevin Barron, Network Manager, Kavli Institute for Theoretical Physics, U. Cal Santa Barbara
Mike Bookey author America at the Internet Crossroads, Principal Pachen Light Consulting
Farooq Butt, Chief Strategist, Motorola’s Mobile Devices Group
Frank Coluccio , President DTI Consulting Inc., New York City
Robert Cromwell Senior Assistant City Attorney in Seattle specializing in telecom regulation
Sean Donelan Security Analyst Cisco
Susan Estrada, President and Founder of Aldea Communications and of First Mile.US
James Enck, Eurotelco Blog and telecom Analyst Daiwa Securities, London
Bob Frankston developed Visicalc and Lotus and later home networking at Microsoft
Fred Goldstein Principal of Ionary Consulting, author of The Great Telecom Meltdown
Tom Hertz CTO Fiber Utilities
Olivier Jerphagnon, Strategist Calient Networks
Bruce Kushnick Founder Teletruth and author 200 Billion Broadband Scandal
Kevin Marks Technorati Engineer, Coauthor Micro Formats, Quick Time Developer
Chris Marsden, European Telecoms Analyst, Rand London, UK
Sascha Meinrath Executive Director CUWIN and community wireless expert
Andrew Odlyzko Director Digital Technology Center, University of Minnesota
Doc Searls Editor Linux Journal and Doc’s IT Garage blog, ClueTrain Co-author
Bill St Arnaud Director Ca*Net4, Canada’s high speed research network
Graham Richard, Mayor, Fort Wayne, Indiana
Chris Savage Attorney CRB, Washington, DC
Dirk van der Woude Civil Servant Amsterdam and fiber expert
John Waclawsky Chief Software Architect, Motorola
Steve Wolff, created NSFnet Backbone – now R&E advisor with Cisco