Implications of Internet Governance
On the Competitive Internet Landscape
Keynote Remarks at
Re-Engineering the Internet
Conference
___________________________________
"The impact of modernity was dissolving for so complex, indeed so subtle a structure. For how could organic institutions be rationalized, when relationships were so intricate that the very attempt to define them would merely serve to accentuate differences? ... It is in the nature of successful policies that posterity forgets how easily things could have been otherwise."
-- Dr. Henry A. Kissenger, A World Restored (1964),
referring to the Austrian Empire, and the Treaty of Vienna.
* * *
"The nineteenth century invented the locomotive, and Hegal was convinced he had grasped the very spirit of universal history. But Flaubert discovered stupidity. I daresay that is the greatest discovery of a century so proud of its scientific thought."
-- Milan Kundera, "Jerusalem
Address" (1985), in The Art of the Novel.
__________________
I'd like to
begin by noting an important mistake in the conference agenda. I'm billed as an
"expert opinion" which I'm certain has been the catalyst to wry
smiles if not outright peels of laughter. As many of you here today know --
since a large portion of you appear as sources for my articles in CommunicationsWeek International -- I'm beholden to you, as my muse.
Shy an "expert," I'd rather call myself an "adept parrot,"
helpfully chirping back that which I hear from the real experts: you here
today.
Even if
I've never frantically called you on deadline, the fact you've taken the
initiative to attend this conference is a sign that you take these issues
seriously. And as such, you in private industry, government and representative
associations, are all experts.
But it is
not completely ironic that I should address you this morning. Although you are
experts -- you're also partisans. You have a stake in the outcome. As a journalist, I'm supposed to be
independent. (As an Internet user, clearly I'm not.)
How
Internet governance issues will impact the private sector is a vast topic. To
make my twenty minutes meaningful, I plan to divide the subject into three
parts. The first, is to identify areas where the Internet industry has a stake
in the outcome of the current debate -- and where it has a role to play. I'll
explain the nature of the debate, and also explain what I mean by the
"Internet industry." But I believe that these points can be better
made by others here today, so I'll only offer some simple thoughts, as an
outsider.
Secondly, I
wish to offer a brief analysis of the previous year's Internet governance
debacle. As a journalist, I strive for objectivity in my reporting. I am
therefore particularly grateful for this opportunity to temporarily cast aside
those sanitized robes, and instead critically examine what I believe will
become the legacies of the maligned process.
Finally, I
plan to take the conference agenda in hand, and tell you some dirt about some
of the speakers here today. I want to let you know what to listen carefully
for, which they will NOT say today. And I'll tell you why they won't say it.
Perhaps this tenacity --if not rudeness -- will compel one or two of them to
drop a word on the matters I raise.
In
considering these three areas together, I pray that I do not end up falling on
my sword. If I offend any one here, please understand that my only interest is
to communicate something truthful and meaningful, which in other fora such as impersonal
newspaper print, is occasionally appreciated. Ultimately, these are fugitive
opinions, recklessly delivered, and surely soon forgotten.
I.
Internet Self-Governance and the
Private Sector
Internet
governance is a new term for many. For some it is a frightful one. Others would
say it is impossible. A conference in 1996 at Harvard University was called
"Coordinating the Internet" -- which pre-supposed that the Internet
is something that requires order brought to it from outside of it. They didn't call
it "Internet Coordination," for example, that avoids this presumption
in a linguistic sense.
Whether we
use the term coordination, management or governance -- what is it we're talking
about? In short, the Internet is a distributed system, but at certain levels
requires central authority or it doesn't work universally.
Three areas
stand out:
* The
domain name system data;
* Internet
Protocol address number allocations;
* Special
engineering parameters.
The Net
functions more smoothly -- there's less a chance that something can go wrong;
more predictability -- when these aspects are managed, or coordinated. There
may be adjustments to this -- Is there a fourth role? Can one of these three
functions be left to complete distributed management? These are relevant
questions, but better considered at another time, by someone more competent.
The problem
today, of course, is that the Internet institutions that oversee these
functions need to adapt, or be created.
* They are
still aligned with the academic community that developed the Net, but which has
now been overtaken in importance by the private sector, of commercial backbone
provision.
* The Net
was developed in the US, while today the Internet is international.
* The
institutions that "govern" it evolved organically, and have a very
tenuous foundation in law -- and that doesn't scale very well.
* The
original designers comprised a small, diverse, yet generally cohesive
community, but today that community is enormous -- indeed, it no longer is
considered to be the Net's architects, but its users (and in some cases their
government representatives!).
* Finally,
these institutions are still funded by the US government -- an anachronism from
when data networking projects were civil and military projects.
That's the
problem. The system has to change. And because there is a governance role, it
becomes a question of power.
The
Internet industry has a vital stake that the system works. The industry, as Jon
Postel, the head of the Internet Assigned Numbers Authority (IANA), notes in the most recent issue of CWI,
is no longer just the ISPs and router vendors -- it's the users. Users like Shell Oil or the government of France.
It is the banks like Barclays. And as such, since our salaries somehow worm
their way from payroll to our accounts over networks; and we're going to want
to purchase something from a Web site with those euros, or simply check our
balance while ballooning in Bogota; it's us too. Individuals have a stake in
the Net's reliable functioning.
If that's
the problem, and that's who it effects, then in what ways will the outcome of
the governance model impact users, and where can users play a role?
Let's look
at the three functions that require coordination separately, and note where
each bisects the interests of business. The most obvious is domain names, and
unsurprisingly, it is also where the most controversy has been.
Domain
Names:
Business
wants to protect their trademarks and brand identity. Most business people are
unaware of the issues and its seriousness. But those that are aware of it, for
example the director of electronic commerce at the US Council for International
Business, are livid at the idea of many generic TLDs constantly being created.
Every new name is a threat to the existing identity in other gTLDs. I'm
suspecting that some large companies with "international brands" or
public perception will seek moratoria on their names for any new gTLDs that are
created that can lead to confusion in the market place.
How this
will work internationally will be tough. Expect more dogfights and regional
rivalries. Even expect a contingent to seek unicode gTLDs - so users of a
Kanji-based browser can type in characters for top level domains. This would
fracture the Net, for as Christian Huitema, a French engineer at Bellcore in
the United States, notes, the DNS is based on the lowest common denominator:
the 7-bit ASCII character set. Without it, users in Texas might not be able to
connect to sites in Tangiers without special software that turns their
Latin-alphabet keyboard into Arabic.
No policy
will resolve the trademark and brand issues of TLDs - that, I'm afraid, will be
a concern that will exist forever. The best deal business can cut is to protect
truly international names by means of a mechanism to prevent their registration
before any new TLD databases go live. For what I'll call "rightful
duplication" -- Cisco Systems, Cisco Iron Mining-founded 1888, and Cisco
Medical Labs in Bombay, India (fictitious examples) -- perhaps a
".who" is in order. It would be a database-based TLD administered by
linking up all other TLD databases (including country code TLDs), that would
serve to list all shared second level names for any given entry. When the
registration is made or renewed, users can provide a 50-word description of the
site. In this way, cisco.com is a technology company and cisco.med is a health
clinic -- and hopefully less confusion will result, good routing, and better
health in Bombay.
The
Domain Name System's Root, the Data:
So who
manages the data? That's a techie question, and although I have some thoughts
here -- for example, the system needs to become formalized, with more structure
while maintaining the distributed model -- these issues can be better dealt
with in other presentations.
Yet as the
"DNS Blackout" showed last summer, when Network Solutions Inc. that
manages the data, sent corrupted files to the root servers, its stability is of
extraordinary importance.
Users
should ensure that their direct ISPs have failure recovery plans in place and
that backup files of DNS records -- at a minimum all the sites on the single
operator's network -- are kept in case of problems similar to those of last
July.
IP
Number Allocation and Oversight:
This, being
of much more importance than the names, has obviously been largely ignored.
Large corporates and network providers want easy access to IP number space so
they can have the utmost flexibility to modify and increase the size of their
networks. But the numbers are finite -- and will still be, even with IPv6.
Additionally,
the number allocation has to be well-managed. Don Mitchell of the U.S. National
Science Foundation once described to me what happened many years ago when a
number was duplicated. What engineers today call "router flap" -- two
networks advertising the same route -- can become bigger trouble if it's not
caught in time. Imagine a stream of packets originating from Europe are told by
one router that the destination is in Asia, only to then be told later in the
path the destination is in North America! Mitchell called it a meltdown -- the
routers get overloaded and crash. Keep in mind that the Net is great at routing
around single points of failure, so as one path gets hyper-congested, routers
resend the packets by other routes! The problem spreads!
The point
here is that there are two aspects to the IP number allocation process, and
they must be handled with care. And industry can easily play a role. Join the
European registry RIPE, or if you need address allocations, RIPE NCC (and
they're having their trimestrial meeting later this week in Amsterdam, in case
I've so frightened you that you want to join today!). There, you can play a
part in the rule-making process regarding the requirements and fees to receive
allocations, as well as ensure that the registry's number coordination function
is sturdy, and in cooperation with other registries.
Network
Parameters:
This is an
engineering issue. If you're a high-tech firm, either you or companies you work
with participate in the IETF as engineers. The mailing list is open and the
meetings are three times a year. Your "entree" to the process is via
the engineers who participate as engineers rather than individual companies,
yet of course they have the interest of the Internet community -- and the
bottom line of the companies that pay their salaries -- at heart. For the large
cooperate users and residential users, your input on these matters (if you
understand them!) is via these tech companies from which you buy products.
Beyond
that, network parameters don't concern you, and the best thing you can do to
assure the stability and longevity of the Internet is to leave the matter alone
and let the engineers who presumably know what they're talking about do their
work.
The Internet
and the Government:
I'm told
that the U.S. government intends to make policy recommendations that will give
users a larger role to play -- perhaps even establish a formal role for them.
That they deserve a voice, I don't think anyone disagrees with. So the real
question is what sort of role or voice will they have.
The
Internet has long suffered the gilded talk of being a distributed system that
puts power at the edges and can't be controlled. And the Internet community has
been held up at times as some magical democratic organization that empowers
everyone; and merits not might rules the day. I'm not so naive as to sell that
flavor of nostalgia. Part of it is true, but there's another part that makes a
mockery of those perspectives.
Indeed, my
own analysis brings me to conclude that the Internet has long been ruled by
benevolent dictators. And that the hour is at hand when the larger raft of
people want a say in the way it all works. Their values may be crass
commercialism rather than elegant engineering. But that may simply be the
reality of representative democracy.
Either way,
it is to this upheaval in the existing order of Internet institutions and
self-governance that I want to now turn. That this conference is held in Europe
is symbolic, since as Europeans, we've been down this road before, of fracture
and rebirth. We built our nations upon ideals. And we had the floor ripped from
under us as well. May those lessons not betray us today.
II.
Brigadoon, Athens, Gettysburg,
Paris:
Myth, Democracy, Civil War and
Revolution
"The
IAHC is a farce, the committee should have stuck with the spirit of the Postel
draft, which we all did have input into....
At your
name service, Eugene Kashpureff, ALTERNIC.NET"
As far as a
cyber-shot heard 'round the world goes, or gaining the world upon giving up
chains, it ain't historic. But it does belong to history.
It was
written by the so-called "DNS Terrorist" -- exactly one year ago
today, the 26th of January 1997, on the IAHC mailing list.
What have
we learned, one year later? And what are the legacies that we inherit, after
such divisive, painful and unnecessary conflicts; the IAHC process being
noteworthy for the reaction it provoked rather than what it accomplished.
I confess a
certain fatalism. I admire Arnold Toynbee, the Oxford historian who never tired
of pointing out that civilizations fall, and that to save the tumbling edifices
before they crumble, society must respond to challenges -- if it indeed is to
be saved at all....
I believe
that 1997 has been the year of such a historic challenge, for that ephemeral
concept called "Internet self-governance" -- when the institutions
that organically emerged to coordinate the Net were forced to molt, shed a
skin, to face new circumstances. They were unable to do it on their own --
perhaps the Net community leaders lacked legitimacy, perhaps the Net just
lacked people with sound political sense....
Today we
are on the cusp of a new era for the Internet, and we await the U.S.
government's plan to manage a stable transition for these institutions. We do
not know the outcome yet of the challenge we faced, but we know we are in the
endgame.
I suspect
we'll fail.
I see
Internet self-governance in a historical sense. The early days of the Net was a
splinter of time when dreamers applied engineering to enhance communications.
As such, they organized themselves on selfless, rational principles. Many, like
Dr. Postel, were volunteers in this community. They had the best interest of
the medium at heart.
Such self-governance
might be like Brigadoon, the mythic Scottish village. It only appears once
every thousand years. Like democratic Athens compared to Washington DC, the
first incarnation was the finest. The second time is farce.
I see the
past year as the Internet's first civil war. Guess what: Gettysburg. Both sides
lost.
The
Internet needed to rely on meatspace -- the offline world of politicians
practiced in the art of process and procedure to instill authority, legitimacy,
to the Internet's institutions. That is the mark of a failure. It might be
salvaged somewhat -- the European Commission and the White House want the Net
to run itself -- but the initial damage has been done. Public policy has
plunged its flagpole deep into the soul of cyberspace. It will never leave.
Indeed, the
civil war's factions are still upon the ramparts -- this will never go away
either. But that makes the Internet, and Internet governance, resemble
everything else: Politics have political parties. Nations go to war. What we're
seeing is the Internet beginning to look like everything we've ever seen
before. What's new with the Net: Nothing! It's finally becoming banal -- now we
know the halcyon days are past.
In civil
war, there comes a moment, after neighbor has turned upon neighbor, that they
stop. Someone picks up a deserted inner-tube, patches it, fills it with air.
Re-builds a house. Carves a violin. Walks upright after the hurt and
uncertainty of a crooked spine.
The
Internet has little patience for such reflections. Chris Ambler, who went so far as to sue Jon Postel and
others over his claim to own ".web" -- since he said he owned it
first -- now works for Microsoft Corp. Has the civil war ended? Who was even
there to have seen it? A small handful of tired policy makers, geeks and
reporters.
Nevertheless, we're living in times of a veritable "Constitutional Convention" -- the matters we will debate today are similar to the issues others groped to solve at other times. It is institution building, it is social organization. What we decide -- or gets decided for us -- will have lasting significance for users and for the technological underpinnings of the medium itself.
It's for
that reason my own newspaper has shed so much ink this past year, covering the
topic. And I believe these facets of the problem, though heavy on abstractions
and short of depth, are of profound importance to telcos, ISPs, vendors and
users.
As a
French-American, I share two heritages that help me understand some of the
intricacies of the challenges the Internet community faces, as a sort of
"perpetual revolution" plays itself out.
In the
American tradition, the "great experiment" was democracy, and soon
the young republic was faced with the issue of centralized versus decentralized
control. It wasn't called that specifically, but Hamilton's drive for a
central, federal, national bank, and Jefferson's plea that the states remain
more powerful and that the nation retain its agrarian roots, was essentially
one of power, authority and freedom. Everybody loved Jefferson. But Hamilton
won.
In Paris,
they didn't fare much better: The revolution that gave birth to the universal
rights of man was succeeded by The Terror.
My
continual references to the political -- the last networking layer -- may be
appropriate if I'm right that the key challenges are now more political than
technical. IP can't route around human nature. As Magaziner readies his report,
I sense that the Internet is set to move into a new phase -- the geopolitical.
Nations and regional blocks will eye up networking technologies, topologies,
and resources under the rubric of strategic interests. The warm water port will
be replaced by the root nameserver.
Already, as
a precursor to this, note that Australia's incumbent carrier Telstra has sued
the FCC over the accounting rate regime. Aussi Web surfers, they contend,
effectively subsidize U.S. users, since those down-under bear the full cost of
circuits across the Pacific. David Conrad of APNIC, the Asian IP number
registry, this month indicated he would resign, citing the political and
advocacy role the organization now is likely required to play.
Of course
meatspace for the moment matters more. A few days after President Clinton
boldly announced his administration's electronic commerce initiative last July,
both he and the policy's architect, Ira Magaziner, flew to Europe. Magaziner
went to an EC conference in Bonn to sell the plan to the first world,
ostensibly America's allies. Clinton was in Lisbon. The word
"Internet" never fell from his lips -- he came to enlarge NATO.
In the
conference documentation I provided an article from WorldLink magazine this
autumn, where I call Magaziner "the best hope in the Internet community
that the Internet can be free of off-line bureaucracy." The magazine is
published by the World Economic Forum -- the people who host the yearly Davos
conference, which starts at the end of this week, and which I assume Magaziner
is flying off to. I guess he'll talk in generalities, spoon-feed the issue to
the business leaders and politicians there. He knows any Internet transition
policy he cooks up needs buy-in from the rest of the world.
It was
inevitable that the Internet would become enmeshed in politics. It was always
there, since the government, the U.S., gave rise to it. But the transition has
so far been botched. Some would say the transition plan of Internet
institutions hasn't yet begun. I take the view it started a long time ago.
Nobody noticed. We mostly debated distractions.
Look at the
Internet governance participants: They're the walking-wounded. The factions
still exist. As do the tender untruths that seem so innocuous because the lie
is in what is left out rather than what is spoken.
But before
us today is the establishment. Let's not forget that: You are all the winners,
in a way. You have influence, are a part of the "The Process." You're
quoted as experts in the press. The prediction I'll make is that your comfort
isn't going to last long. I see plenty of more Amblers and Kashpureffs on the
horizon.
III.
Listen Closely to What They Don't
Say!
Vaccination
theory says that a small dose of disease allows a body to build up its natural
defenses. And if I may play the role of a virus -- which is very easy for a
member of the fourth estate -- I now wish to tell you about what you WON'T hear
the other speakers say.... Ultimately, this may serve as a polite preperation
for the firestorms to come, in other venues.
(Delivered
extemporaneously)
# # #