Published today, this is the introductory essay to:
IP Everywhere: Riding the Internet Tsunami
The State of the Internet 2000
At the beginning of the new millennium the Internet is globally
triumphant. Blindly embraced by politicians who do not understand
it, the Internet is touted as the engine that will power the economy
of the new century. The lover's embrace by the rich and powerful has
helped to make it a fad which is increasingly in control of the stock
market. Internet companies are too often valued not out of sound
understanding of their business models and cash flow dynamics but
because they are Internet companies. The Internet has emerged in the
1990s as perhaps the supreme "disruptive" technology in the sense
popularized by Clayton Christensen in his 1997 book The Innovator's
Dilemma.
Christensen found that "disruptive" technologies generally "under
perform established products in mainstream markets," however they are
generally cheaper and simpler to use than the sustaining technologies
that they ultimately dethrone. Examined from a slightly different
perspective, "disruptive technologies give customers more than they
currently need or are willing to pay for. [Source is TeleGeography
2000 . Review follows.] This is one reason why such technologies are
usually commercialized first by customers serving niche markets such
as data networks or CLECs."
Because "disruptive" technologies bring to market a such a radically
different value proposition than was here-to-fore available, those
practitioners of the older sustaining technologies find themselves
unable to quickly adjust their business models and infrastructure to
compete head-to-head with the new "disruptive" technologies. They
may also be constrained by the very service models which have made
them essential to their customers.
To name just two examples central to the Internet's triumph as a
disruptive network: in the last three years, we have gone from the
ancient sustaining wisdom that telecom networks must be intelligent
at the center, to the new paradigm demanded by the Internet--that the
intelligent devices at the edge of the network must find nothing in
the middle that gets in the way of communication between them. Also
we have traveled from the certainty of knowledge that publishing is
an activity whereby elite gate keepers package knowledge for the
masses to a world where every computer owner can become a publisher.
In this new "disruptive," Internet-driven world the old realities of
our once comfortable industrial world are turned sideways if not
inside out. The trail that has been blazed by the new media and
mirrored by the old media, is simply to hop on the bandwidth wagon,
and hype processes that are not well understood yet. Jeff Bezos, the
CEO of Amazon.com, a company that has yet to report a profit, becomes
Time magazine's "Man of the Year."
This report will give an in depth look at reasons for the success of
the disruptive Internet. We have technology analysts, e-commerce
analysts and legal experts bringing their respective expertise to
bear as they tell their separate stories of who is winning and who is
losing. What we lack is a look at the larger picture of how the
interaction of these areas will determine success and failure over
the longer term.
This report will also examine in depth certain critical areas of
technology change. As it does so, it will show how the efforts of
pre-Internet telecom entities to channel and shape the legal and
political control of this technology no longer fit the competitive
multiple protocol environment under which these entities must work.
The processes developed to service the pre-Internet, circuit
switched, telecommunications systems simply do not afford policy
makers any way to deal fairly and harmoniously with Internet
disruptions. There are no longer appropriate pigeonholes and
regulatory boxes that can sustain the old way of doing things.
Democracy and technology no longer mix. If technology is to be
consistent with public values, the public must understand the
technology and be able to move from that understanding to a grasp of
the stakes.
Understanding the stakes is difficult. Technology itself is
increasingly complex. Point A does not always lead to Point B, even
if Event C occurs. The impact of a new technology on is increasingly
difficult to judge. Consequently we are likely to find that many of
the assumptions on the basis of which policy has been developed over
decades are no longer valid. For example this report will show that
the buyer's market for bandwidth is filled the potential for
financial peril as well as profit. And it will show that the obscure
debates over end-to-end network transparency can exert transformative
influence over how the Internet is organized.
Ultimately, as the system grows sufficiently complex with fewer and
fewer people competent or even able to direct its course, some may
ask when does chaos set in and take control? Or when does Internet
coordination become merely a series of panic stricken interventions
where the "coordinators" lurch from one crises to the next while
trying to keep all the "vested" interests in their same positions of
dominance? Already technology that our political leaders do not
understand is driving the formation of public policy. Thus,
democratic values independent of the technology are no longer the
primary public policy goals. These are additional "disruptive"
effects of the triumphant Internet.
Those who dare to assume the role of enterprise strategist in the
midst of these changes must understand that their successful
stewardship will depend on their grasp not only of what Gilder calls
the ascendant technologies. They must also develop at least two
other skill sets. One is an understanding that, as Lessig has shown
so well, the success and visibility of the Internet has brought to it
a level of attention where the legislators and the regulators will
impact it whether the rest of us like it or not. [For Lessig's Code:
and Other Laws of Cyberspace see
http://cookreport.com/lessigbook.shtml ]
The other is a skill set that will help strategists to evaluate the
stakes behind the questions of what will become the engineering
agenda for the Internet. Compromises have been made in network
management that can spell trouble further down the road. To scale
network growth and cope with available IPv4 numbers, design decisions
compromising end-to-end connectivity were made in the early 1990s. As
the long IETF discussion published in the February 2000 COOK Report
[and reprinted pp. 259-272 below] shows, there is a fallout from
these decisions. The fallout is an incipient battle over how some of
the intelligent edge devices of the network will communicate through
the center to their intelligent counterparts at the opposite edges.
The approaches taken now as to early implementation versus delay and
revision of IPv6 will impact the diversity of the infrastructure of
the future Internet. Arcane but very important debates like these
will determine whether issues of control over end-to-end uniformity
and network transparency take priority over everything else. Will
time spent toward these ends mean that other more critical
engineering concerns are ignored? A feared consequence is that the
Internet's ability to continue to scale may be endangered.
Technology Changes
The year just ended saw a continuation of the bandwidth revolution
begun by the real take off of WDM technologies in 1997. New
developments in DWDM (Dense Wave Division Multiplexing) and optical
switching have further multiplied the bandwidth available from a
global binge of fiber deployment. A revolutionary drop in the cost of
data storage combined with increased network speeds is making it
possible to deliver data across a wide area network more quickly and
more cost effectively than across the bus on the motherboard of a
single computer. These changes made the dream of the network becoming
the computer - a dream that was first articulated earlier in the
decade -seem likely to become true. They also made possible the rise
of a new application service provider industry.
The year also saw the full flowering of new global telecom providers
built on the inexpensive infrastructure made possible by the new
technology. Qwest, Level 3, Metromedia, Williams, Enron, Global
Telesystems, Global Crossing, Next Link, Teleglobe, and Above Net to
varying degrees are all examples of these new telecom giants. While
Teleglobe has been around for a number of years in a rather different
form, and Williams is a reincarnation of an earlier venture, all of
these new players with the exception of Metromedia have been profiled
within the pages of the COOK Report. Questions of interconnection
and peering are still critical. Equinix's neutral Internet Business
Exchanges to be built globally under contract with Bechtel will be
the most high profile model for fitting the new backbone players
together and enabling cost effective interconnection.
As bandwidth consumption and fiber deployment soar the purchase of
bandwidth has become especially tricky. Prices of bandwidth bought
in bulk over long periods of time are plummeting and options of
purchase are increasing. (See Figure I Fiber Optic Cost Trends and
Figure X Bandwidth Market Matrix from TeleGeography 2000 on pages 2.
and 161.)
It is a buyer's market but a dangerous one as AboveNet found out when
14 months ago it paid over 8 million dollars for a 25 year IRU on a
trans Atlantic STM-1 only to see the price of such an IRU plummet to
less than 3 million a year later. Options are available for those
who make this kind of mistake. MetroMedia Fiber Network (MMFN)
bought AboveNet last year assuring it of a direct and affordable
supply of bandwidth. As David Isenberg pointed out in late December,
according to MMFN's president, "a DS3 from the telco costs $3,000 a
month, and a comparable MMFN fiber costs about $5,000. But the fiber
can be lit at OC-12 - [a speed that is]14 times faster than DS3 - for
$500 more per month (assuming 10-year depreciation). This works out
to about $400 per DS3 per month." The lesson is that as new
companies are built on shifting assets and technology, agile fast
moving players will find many options. But, as the market
solidifies, such options won't last indefinitely.
An important new business involving the sale of surplus bandwidth at
exchanges is beginning to emerge. We learned about RateXchange in a
phone conversation with Ross Mayfield its CEO on December 30. In New
York and Los Angeles Mayfield is running what he calls The Real-Time
Bandwidth eXchange. He says that RTBX is a switch-based exchange
that facilitates the entire transaction of bandwidth between
interconnected carriers. Benefits include significantly lower
transaction costs, anonymous trade, immediate delivery, and guarantee
of payment. He seeks to create a win-win situation where an ISP with
a temporary bandwidth surplus can sell bandwidth to an ISP with a
shortage. His ultimate goal is to commoditize trading in bandwidth
by means of the sale of contracts for future delivery.
When Web based interfaces for creating in almost-real-time
single-wavelength circuits between two points become common,
Mayfield, and by then many other players, intend to have created the
infrastructure to make this possible. Given that the Internet for
the past few years has been in a blind race to make enough bandwidth
available to meet huge demands, this technology will be likely to
give a welcome rationality to this marketplace. For the first time
it should offer a means of efficiently and quickly bringing available
bandwidth into balance with demand, and thus allow venture
entrepreneurs to build new infrastructure where it will be used
profitably
As the February COOK Report's lead article on the state of wireless
shows, we have many new developments in wireless reaching their
maturity and pulling the whole structure towards more ubiquitous,
always-on connectivity and applications. Wireless, becoming digital
in 1999, became also broadband. Almost all the kinds, number and
varieties of activities conducted via the wired Internet are now
beginning to be carried out on the Internet in wireless form. This
vast growth of wireless infrastructure has fed on itself to lead to a
continued explosion of net use and net traffic. Large growth in cable
and DSL connectivity in 1999 also contributed to the increase in
traffic and in the number of always-on connections.
In 1999 the Internet reached a critical mass where every player in
commerce and business had to have a presence. It became unthinkable
not to have an Internet strategy. The growth in Internet
participation fed on itself and made possible business models which
came to subsume every part of human economic activity. These changes
mean that we are looking at a telecom world in which the old
standards of evaluation fail us.
Technology Changes Divide Management into Irreconcilable Camps
The revolution has given us two opposed ways of thinking about and
acting to achieve network organization. One may be described as the
Bellhead intelligent network and the other as the Nethead stupid net
work. [See http://www.tmdenton.com/netheads3.htm and p.206 below.] In
the first intelligence resides in expensive complex smart switches in
the center of the telephone network linking stupid edge device
telephones. In the second fat bit pipes linked by idiot savant
routers connect intelligent computers on end user desks. The
intelligent "Bellhead" network is generally organized to be run from
the top down by an extremely rigid hierarchy. The business model is
a monopolist and feudal one with the lord of the manor ensconced in
his hilltop castle.
The "Nethead" architecture is organized through the creation of
structure by cooperating autonomous groups. In addition to being
important mental constructs, both architectures are now grounded in
billions of dollars of competing and not generally compatible
infrastructures. These competing infrastructures are the foundations
on which telecommunications, electronic commerce and individual
access to vast knowledge will be determined in the new century. The
older infrastructure (the Bellhead) is inherently more subject to
vertical integration and organization. In the Bellhead environment,
vertical, centralized, and rigidly-controlled hierarchies are used
to enable a single centrally determined and centrally priced palette
of communication services, while the newer Nethead companies more
often tend to be organized horizontally as companies whose market
strength is in the delivery of one or more layers of the protocol
stack.
This report proposes the outlines of a new business model structure
for identifying and understanding Internet players who will prosper
and ride the disruptive wave successfully. Growth areas in Nethead
oriented companies are for specialists in bandwidth, interconnection,
content hosting, application outsourcing, email. A player here is
either an infrastructure-providing specialist, or one who coordinates
and manages technology specialties horizontally to provide services
to end users. At the transport, infrastructure-providing, level you
have the continued growth of new telecom greenfield players such as
Level 3, Williams and GTS to name but three. These companies are
taking advantage of new technology and new market conditions to grow
from nothing into billion dollar plus companies in only two or three
years although the long term judgment of the market is not yet in as
to their appropriate valuation.
Opposed to this new disruptive Internet business model is the old
solid and stable vertically integrated telecom model followed in the
United States by the ILECs. ATT, with its attempt to buy control of
the cable industry to have a foundation from which to deliver all
telecom services, fits into the same category. MCI WorldCom with its
attempt to become one of the biggest world wide integrated phone
companies by acquiring Sprint also exhibits the telephone headed
business strategy of a gigantic vertical integrator of dissimilar
technologies and services. By the same token the AOL Time Warner
merger represents the old industrial age vertically integrated model
of the marriage of content and a delivery mechanism for the content.
The Two Approaches to Slug it Out in 2000
We find that one battle of the year 2000 is to be between the
horizontal versus vertically integrated business models. The
horizontal model that is TCP/IP-based over gigabit or 10 gig Ethernet
over fiber - a stupid network with smart peripheries where new low
cost services can be cheaply and quickly inter connected like lego
blocks. Internet players tend not to be vertically integrated.
Instead they are bandwidth and transport players like Williams,
Enron, Global Telesystems or end user aggregations of services like
MindSpring, Earthlink, or AOL that run over someone else's
infrastructure.
Disruptive Internet companies are nimble and quick compared to the
older players like WorldCom that are still driven by empire building
in search of elusive economies of scale. In a vertically integrated
company, you may think you can cut costs by eliminating duplicative
services, but instead you have a management nightmare of complex
systems where glitches can mushroom into multiple week outages like
MCI's frame relay collapse during the summer of 99. Or ATT as it
attempts to build what it sees as the "network of the future" out of
CATV policy and infrastructure. These players tend to see the
Internet in terms of what has come before - in other words in terms
of what is already familiar to them. What has gone before is
regional feudal monopolies, based on exclusive land-based rights and
exclusive land-based responsibilities overseen by public regulators
enforcing a social benefit/social cost analysis. And it well may be
that this model may be with us longer than we would like. We should
not think it can be so easily replaced, or that such replacement
would have only good consequences.
Can the old vertically integrated intelligent network players
compete? Or should they have some kind of protected quid-pro-quo
status as the baseline common carrier? It will be much more difficult
when wireless means cannibalization of their local loop T-1 circuit
income rather than new income. Look for them to engage in regulatory
plays like ICANN.
The horizontally organized players can see the disruptions coming
much more clearly, push the development envelope better, and achieve
economies of scale by specializing in transport services, or storage
services or application outsourcing. The vertical integration of MFS
into the realm of its telephone company operators was what made the
Network Access Points (NAPs) fail with WorldCom in the 1996 -1997
period. (Some maintain that with a carrier running the NAP, it was
then put in the counter productive position of competing with its own
customers.)
In contrast horizontal aggregation of services will enable Equinix's
neutral exchange points. Vertical integration is needed to make all
the pieces of the intelligent bellhead network operate and to pay for
relatively few and very expensive centralized switches. To make
vertical integration work, one needs to control as much of the
telecom environment in which one operates as possible. Even so some
astute observers point out that, as the Internet industry continues
to grow, even the horizontally oriented Internet companies are likely
to make vertical market plays if they are sufficiently cash wealthy
and driven to expand market share. We must strive for educational
processes that will make clear to regulators and the public alike
what economic and policy consequences will flow from the two
different approaches to the market.
Ignored Perils of Scalability and Architecture
The plug and play nature of Internet interoperability means that the
over all network can be quite hierarchical, because there is room for
many companies of many different sizes. Consequently, the management
of each company can be quite flat and and be able to respond rapidly
to changing conditions. Internet interoperability encourages
specialization and rapid technology development because the
specialized pieces will fit together and work together. However the
Internet is now set to begin paying penalties for its success. The
triumph of IP in global telecom has made legacy companies believe it
is critical to their futures to dominate the implementation of IPv6
in order to preserve illusory and unattainable uniformity for their
goal of continuing in central control of network operations.
Let us summarize the argument we have just made. The Internet is
having a disruptive impact in three areas of global importance.
First, disruption in global commerce. As trade of goods and services
migrate to Internet locations, then individual economic power is
derived from one's position within and one's subtle understanding of
the new infrastructure.
Second, disruption in the restructuring of global telecommunication
services to accommodate Internet requirements and efficiencies. As we
have pointed out, the Internet is enabling new infrastructure that
can provide all manner of telecommunication services at a tiny
fraction of prior costs.
Third, disruption by the revolution of storage technology which
re-enforces and amplifies the plummeting costs of bandwidth
provision. As a result, some computers will expand into
configurations which look to them something like local intranets,
with extended storage lodged on the Internet rather than inside the
case, because in some situations, external Internet storage will be
cheaper and safer than local storage options. This trend toward
external Internet storage may shape the net into the global
repository for information and knowledge. While the Internet was
originally designed to be decentralized and not subject to single
points of failure or control, its current commercial success makes
today's Internet a target for desperate and daring stratagems to both
obtain ownership wealth and the privileges and prerogatives of
control, containment, restriction, and punishment.
Control Points: Ephemeral or Real?
One view of how services in the new telecom world will be delivered
is by an architecture of an idealized Internet where TCP /IP will
always be able to be sent from end-to-end. Because of a series of
compromises involving private address numbers, corporations now sit
behind firewalls that demand the imposition of many translation
devices. Thebellheads see these devices (NAT boxes) as obstacles to
the imposition of a centrally controlled communication path like
IPv6. The betheads don't see any of the bellhead's reasons for
concern. They are happy to use multiple protocols and bridges over
otherwise non interoperable network links. The nethead's outlook
focuses on extensibility while the bellheads, through trying to
influence the roll out of IPv6, focus on control.
ICANN which has garnered most of its support from bellheaded legacy
organizations, has waged a generally successful campaign to sell
itself to the press. The major press, having nothing other than
ICANN's sound bytes from which to claim a clue, generally has no idea
that anything is amiss. ICANN meanwhile is determined to error on the
side of control concluding that it has the ability to enforce a
uniform roll out for IPv6.
The two sides, one riding a disruptive technology and the other
trying to extend the life span of the old sustaining technologies,
sit in a hostile face off as they roll out competing business plans
for the 'converged' future of voice and data. But the Cook Report
argues that the netheads and the bellheads are fighting about
dominance and control issues, when there is emerging evidence that
they must, instead, focus on scaling and inter operability issues.
We are very sensitive to Sean Doran's concerns about scaling of
network architecture and routing which will enable the Internet to
cope with the coming onslaught of broadband soon to be generated by
the increasing numbers of intelligent devices in homes and small
businesses at the edge of the network. [See pages 150 - 152 below.]
But in the current competitive environment this is very little talked
about.
Therefore one must consider whether the Internet technical community
while doing battle over the presumed uniformity necessary for IPv6 is
actually depriving itself of the time and effort needed to face more
intractable problems of scaling, including protocol diversity. Up to
this point the Internet has scaled by a series of kludges like the
use of CIDR and RFC 1597 assignment of numbers for private intranets.
These are kludges that leave no one happy. While the Internet is not
likely to suddenly fall off a cliff, if these issues are not attended
to, performance is likely to get uglier and less reliable.
Thus we predict that what is otherwise an arcane dispute among
network engineers is likely to become a basis for an operational
strategy attuned to the structure, economics, and philosophy of the
vertically integrated telcos of the majority of the planet. These
conflicting issues meet head on in the distribution of IPv6 and over
ICANN which has placed itself in a position from which it could
attempt to coordinate the implementation in an effort to provide the
uniformity that the bellheads want and believe they need in order to
stabilize their world.
We find that a vertically integrated market strategy is usually
telco operated and at odds with the horizontally oriented and more
cooperative Internet market strategy. However in the current chaotic
period of expansion and emphasis on building market share, these
categories do not always hold. For the changes that we have
chronicled in the COOK Report for the past eight years are focused on
building market share as well as in the creation of new, successful
businesses based on openings provided by the "disruptive" technology.
When this happens, a business like 3Com may mature to a point where
the MBAs looking at the balance sheets and stock prices begin to
drive things more than the engineers.
As one astute observer confided to us: "vertical integration has
become the way that everyone is trying to rule the Internet economy,
and I have seen very specific presentations from 3 Com and others
saying explicitly that this is what they are hoping to do. It's the
way to maximize shareholder value, but at the expense of an Internet
that works well and that is user-friendly (by which I mean users that
can innovate, not just click mouses). Similarly, I believe that we're
past the nethead vs bellhead debate into something far scarier, which
is where netheads and bellheads both realize that they're going to
have to integrate their networks. Moreover they may be committing to
doing so even when they know that they don't have time to agree about
big chunks of the end result. In other words, while the PSTN may be
getting dumber, intelligence is being injected into the Internet in
some interesting and unexpected ways."
Under such conditions amazingly powerful technology will almost
certainly run into snaffus originating with kludges in protocol
design and IP numbering made in the first years of the take off of
the commercial Internet. When this happens more than web site
performance will suffer. Also the profits of those companies which
have marched forward fueled by nothing but blind faith in the
continuing scalability and robustness of the Internet will take
severe hits.
Alternatives?
So are there alternatives? Ed Gerck who has participated in some of
these debates has some wise suggestions in a long IETF discussion
reprinted in the February 2000 COOK Report. Gerck advocates that the
battle should not be thought of as either the uniform roll out of
IPv6 or the failure of IPv6 and thus being fated to live with a
network with some hosts blocked by firewalls or NAT boxes. There is
a third way marked by diverse protocols that can inter operate.
When we asked him for further comment he replied "I saw no
counter-argument raised to it - rather, it fell like an eye-opener to
the reality that Nature is based on diversity. Those that believe in
an uniform Internet as the only way to achieve end-to-end security
are actually still locked in the network paradigm of the 70's where
network administration dictated orders to the entire network, by
design. So, they will most naturally fight a multiple-protocol
Internet, because they cannot intellectually cope with it."
"However, in the Internet paradigm of the 90's, now truly as networks
of private networks in a progression over 30 years as catalogued by
Stef, [Einar Stefferud]we already have multiple protocols in
coexistence at various levels - what matters, so we have learned, is
not that there must be one unique protocol at every place and time,
but that different protocols, at different places, at different
times, and doing different things are indeed able to work together
when the objectives are the same. And this, quite revealingly for the
success of the Internet IMO, is how we humans prefer to work
together, how commerce thrives and how we can enrich each other's
living experience."
"So, communication protocols such as IPv6 must not be based on an
excluding model, where they must kill any other protocol, but on an
including model where interfaces are provided for backward and
forward compatibility. Otherwise, even if IPv6 would be adopted by
force on 90% of the Internet - this would still leave out 10% (which
is projected at 40 million end-users in 2000) and would immediately
raise questions about IPv7, IPv8 and so on. The only way to address
the backward- and forward-looking questions without requiring the
whole world to change standards at whatever cost at whatever time, is
to provide for inter operation."
"Thus, that is why I wrote that tools for IPv4/IPv6 inter operation
will be needed .... and valued, including NATs as a fundamental
building block - even though they began as a stumbling block in
another context," Gerck concluded.
Consequently, one must weigh whether the Internet technical community
may be actually depriving itself of the time and effort needed to
face more intractable problems of scaling, including protocol
diversity. Those who would pretend to do large scale strategic
planning in the midst of the Internet revolution had better
assimilate the lessons of Larry Lessig's Code and Other Laws of
Cyberspace. As the Internet moves into the legal and political
arena the question for analysts to ponder is not a naïve belief in
the unregulatability of the Internet. They need a much clearer
understanding that the operation and impact of the Internet can be
determined by the legal system, by the way in which its dominant
protocols are coded, by the architectural environment in which it
operates and by the local customs that determine what behavior is
socially and commercially ethical and acceptable.
Gerck's statement about the need for inter operability tools for IPv4
and IPv6 is a good summary of why Lessig's seemingly arcane issue of
network architecture and control is politically very important. For
in Lessig's language, it offers yet another example of how the
technical decisions that change the shape of Internet architecture
will indeed render it more subject to legal and regulatory control.
The "disruptive" Internet is in the midst of what looks to be a
general triumph. During our eight years of reporting on the growth
of the commercial Internet, the COOK Report has developed a unique
perspective that synthesizes technology issues with their operational
and engineering counterparts. We believe that synthesis of these
intricate processes as described in detail in IP Everywhere, falls
outside the scope of more specialized analysts. We also believe that
those who become familiar with these complex developments will be
able to formulate more successful business strategies. For the first
time the technology is likely to be as much affected by policy
decisions of network design and global regulation as by the older
unregulated Internet frontier mentality which drove the mad rush to
add users, scale the network, and increase market share. Winning
players must take into account the daily onslaught of new technology.
They must also make the right judgments about the governance and
standards wars, and the rapid shift in market economics. The winners
will be those who get the right answers to this complex mix of
questions.
Gordon Cook January 24, 2000
****************************************************************
The COOK Report on Internet Index to 8 years of the COOK Report
431 Greenway Ave, Ewing, NJ 08618 USA http://cookreport.com
(609) 882-2572 (phone & fax) IP Everywhere: Riding the
[EMAIL PROTECTED] Internet Tsunami - 392 pages
just published. See http://cookreport.com/ipeverywhere.shtml
****************************************************************
