Very thought-provoking, and accurate, picture of the current state of the Internet information state. See also a Reporters sans Fronti?res article on the devisive Internet control situation in (for instance) Nepal: http://www.rsf.org/article.php3?id_article=14461

- ferg

Rebecca, this is an outstanding summary!

Balkanization is one of the justifications used by those that want to introduce split roots (among financial and other interests).

Part of the CN2 root that also happened was the introduction to that root, som alternative namespace of some new .IDN TLDs, making it possible to register domains ending with an idn tld.  (ie. <domain>.中国 [actually <domain>.xn—fiqs8s])

We’ll see a variety of additional justifications and motives for root division, and we’ve seen many over the past 10 years. 

is one of something that countries are looking at in an effort to re-introduce sovereign concepts to a borderless Internet.

China’s advantage in controlling its internet users is not v6 or anything else like that. It is that most chinese quite naturally gravitate only to chinese sites, hosted in china. What is physically located in china is quite easy for them to control.

The NGN is a fairly serious effort, and in cooperation with Japan and Korea - which too have deployed most of this in production, so that they are way ahead of most of the world in metro internet, broadband penetration and other things like ipv6.  “Standard” connectivity is grossly inefficient and slow, based on decades old technology, is what some NGN advocates say.

Things like proprietory wifi authentication protocols and a few other things might be viewed by some as part of the usual game of “one-up-manship” that goes on in the wgig and other international forums .. the argument in favor of such a statement goes a familar way - that the chinse want a countrywide intranet, but for that they need the protocols and other framework components to be deployed and in wide use.

It must be noted however that there are several other “new internet developments” that come out of China, ostensibly endorsed by the chinese government, that are widely derided as stupid and short sighted. Like the “IPv9 proposal” that was floated recently. 

As James Seng points out in an old circleid post - http://www.circleid.com/article/646_0_1_0_C/ - IPv9 was actually a much more familiar phenomenon - a tech company that came out with a highly stupid idea and launched a PR blitz to market it.

Given that the chinese government has a stake in lots of private or semi private industries, and given that business relationships there are fueled by “guanxi” - the right connections - what was actually overhyped PR promoting a dud product suddenly got elevated, in several people’s minds, to the status of official policy.

So there are several questions that can be asked here.  What is behind all these things?

Industry with a new marketable technology and enough political clout to get the government to endorse it?

The chinese government preferring an internet that is “different”, for reasons ranging from all-chinese URLs to greater political control?

Part of a coordinated PR strategy to make China appear to take the lead in Internet efforts, so that when they make proposals in international forums, they can also cite their highly advanced technical competence to operate what they propose?

I rather suspect that I’m one of the six blind man looking at different parts of an elephant.  So are these things I pointed out part of the same beast, or are they actually a rubber hose, a tree trunk, a wall etc, that I’m trying to construct a non-existent elephant out of?

Why shouldn’t people be able to create their own internet namescapes?

What ever happened to the end-to-end principle?  Why should users (and their proxies, ISPs) be refused the right to chose how they will use the internet.

If a community of internet users decides to establish their own name space on the internet then who are we to deny them that right?

If they are willing to live with the inconvenience of having their own name space then who are we to demand that they cease and conform to the naming dictates of outsiders?

If they don’t feel the need to communicate with us then why should we force it upon them?  To do so sounds like one of those arguments from the “direct marketing” folks about how their right to send spam or junk faxes trumps our right to control our own lives.

The argument you make could be adopted without change by telephone companies to demand that voice over IP be denied because existing POTS phone users can’t call every SIP phone.

Breaks with the past and community established boundaries are the natural results of innovation and human choice.

The argument that there must and shall be but one catholic internet DNS root is an argument that denies both innovation and human choice.

> The argument that there must and shall be
> but one catholic internet DNS root is an
> argument that denies both innovation and
> human choice.

Here’s what I wrote in Joi’s blog in response to just this often repeated, and just plain wrong argument.

Well .. it depends. That old Al Gore cliche about the information superhighway does have its points ..

You can exercise your right as a free and independent citizen of cyberspace (cue John Perry Barlow’s declaration of independence of cyberspace) .. and take a convenient short cut across the superhighway .. ending up as roadkill. Or you could say it is your way rather than the superhighway, strike out on your own and suddenly get this unpleasant feeling that you are all on your own, somewhere in the middle of a featureless indefinite plane. The key word is “critical mass” - which none of these proposals have got.

Now, we have everybody from the strongly independent alternate root people who resist what they think is too much IANA/ICANN/root server operators control, to some dictatorships who love the idea of a separate internet for their very own that they can control as they please ..

and various other people who want a different internet for reasons from “domain names that are all in our local language” to “I hate voip, its killing my telco revenues and this will help me kill voip in my country” ..all of them making common cause, and not very much sense, at the wgig for entirely different aims.

My feelings about this mess are much better put in a sermon by a 16th century clergyman who went all the way from rake to devout priest in the course of his life .. Meditation XVII by John Donne.

[...]

No man is an island, entire of itself; every man is a piece of the continent, a part of the main. If a clod be washed away by the sea, Europe is the less, as well as if promontory were, as well as if a manor of thy friend’s or of thine own were. Any man’s death diminishes me, because I am involved in mankind; and therefore never send to know for whom the bell tolls; it tolls for thee

The full text of this sermon is at http://isu.indstate.edu/ilnprof/ENG451/ISLAND/text.html - I just quoted what is probably its most famous part.

So yes, I do care if all these little pieces splinter off from the Internet .. I do remember a time when the internet was a disconnected set of islands, independent of the main. A time when people talked about leaf nodes and gateways, and had to make expensive international calls to connect elsewhere. So do we finally go back down that path again?

Response to Suresh Ramasubramanian:

You say that you “do care if all these little pieces splinter off from the Internet”.

That’s nice.

But if I want to be one of those splinters why should your choice prevail over my choice?

What you and the rest of the catholic-rooters are doing is to impose your choice on those who wish to differ.

To continue your metaphor - when the splinter gets big enough it’s hard to know which piece is the splinter and which is the board.  As was mentioned in the Circle-ID note by Naseem Javed - “English is the big mama of the business language on the global scene, but on the spoken side, Chinese is the big papa”.  So if we want to get into the debate over who gets to be the one name space that your so much desire than we had be better be willing to say that the current DNS name space may come up on the loosing side.

By-the-way, my argument is hardly “plain wrong” on either a technical or logical basis - My argument is that the internet is about choice today and tomorrow.  Your argument is that the internet users of today should not have the freedom of choice of those internet users of the past.  Your argument is that your choices should prevail.  My argument is that individual choice should prevail.

I am an American - my internet life is made all the easier if my language and my culture are imposed as a new form of techno-imperialism onto the 90% of the worlds people who don’t speak my language or live in a wealthy California seaside city.  So I have a natural sympathy for the internet world you are trying to create.  On the other hand, history teaches us that such efforts are ultimately futile and counterproductive.

So rather than trying to defend the undefendable we should accept the internet as a pluralistic system with multiple name spaces based on local, and even individual, choice.

Karl, my argument is that the internet is just too valuable to fragment.

What you propose will effectively reduce it to lots of split apart and separate systems.

It is not just english speakers who wont be able to access chinese sites under an alternate root, probably to download anime and canto-pop if nothing else.

Think of (say) korean and japanese people.

Then let them set up their own system. After all not too many koreans or japanese john q public internet users speak english or even learn it at school.

English has a rather interesting status in India, a country where all you need to go to encounter a completely different language, with its own script, pronunciation etc is to cross a single state border.  And two or three hours by air will bring you to states with languages that are completely different and alien, based off different root languages, completely unrelated script etc.  Without english as a “link language” we’d be stuck using sign language to communicate I guess.

As I said, you are free to set up an alternate root. Nothing stops you, except maybe a vague sense of regret if you have one about fragmenting what was pieced together from a whole lot of fragments (that would stop me, it doesnt appear to be a consideration for you, fine).

Go get some critical mass for it. And then find out ways in which your alternate root can inter operate with all the other alternate roots out there.  And find some way of resolving things like a .biz tld conflicting with an alternate root’s .biz tld.

Sooner or later you’ll find things tending right back to what you wanted to leave behind you in the first place.

Using a facile catchphrase like “techno imperialism” completely trivializes the issue, by the way, besides being flat out wrong. Comparing what is effectively a lingua franca to the “no taxation without representation” of the 1770s, or Gandhi et al versus the british in India in 1947, just doesnt make sense to me.

Let get real. There is the world digital ecosystem. It needs a network service. It used Tymnet I, Tymnet II, OSI, Internet. Question is “is Internet II its next bet or do we need something new?”.

To answer the question we know:

- there are four poles of authority (power and competence): technical, political, economical and societal. Different requirements.
- the network doctrine therefore needs to be centralised (nations, languages), decentralised (infrastructures, communities) and distributed (user-centric), what means a multitude of network partitions of different granularities.
- the average Wi-Fi user has the communication and processing power - and the nuisance, innovation and standardisation capacity - of a nation 10 or 20 years ago. Money will not come from their internet access, but from their intelligent servicing. So there is money for needed complexity.

If we compare with the preceding (both successful) partitioning experiences [Tymnet and OSI]. We had 50 to 100 nations and tested a few languages. We now have all the nations, 20.000 languages, millions of communities and billions of direct or indirect users. ICANN described in ICP-3 that the Internet always developed by users’ experimentation. So, there should not be a problem, should we not be delayed by the integrism of the Internet Church of the Legacy Days (we are probably the only ones to have answered the ICP-3 call so far).

Millions of “standardizers” can create, promote and progressively filter applications. They can also organise a network architecture of usage. But if they are to share the current “mono” by default internet architecture parameters (one DNS, one numbering space, one user class, one language, one IETF, one ICANN, one dominant Government, one Governance, etc. etc. and worst of all “one single language tagging” commercial based system as discussed at IETF), they must engage into competition and into business/political/vital protection from the others (cf. the long expected US Statement of Principle). Partition is here. And right now it is balkanisation: multilingualism, numbering, bandwidth, culture, etc.

Now, let consider there is no more scarcity because the architecture parameters are tuned to “multiple”. Instead of being subject to balkanisation we can benefit from compartmentalisation. Like in a ship’s hull. It increases stability, security, risk and pollution containment, surety, etc. ... and _unity_. The Internet was called “the [physical] network of the networks”, the Internet is now to become “the networks [of usage] within the network”, with every user being a member (or a user) of as many externets as he wants.

Robert Tr?hin and Joe Rinde did it in 1977 (and OSI had it partly built-in) with Tymnet. They conceived the root name principle (creating a super TLD concept - later used in RFC 920): in a network a root name is the name of an external network (externet), of its class of users, of its group of resources, of its Registry, of its Gateway, of its Reference Centre (like the IANA), of its Governance. An externet can be an extranet or an external network look-alike within the main network (this is the rule in a global network). Domains are externets sharing the same class. ICP-3 documents the use of multiple classes and calls for their (over delayed) experimentation (we propose ccTLDs and everyone to carry it - the UNITRY project).

This means that China, soccer’s fans, cities, etc. can develop their _same_ “own Internet” either as a local balkanisation, or as externets co-operating to the solidity and to the unity of the global unique and pervasive Internet continuity. Obviously there is work, thinking, experimentation, testing, etc. But more than anything else there is a need for us all to convince Legacy Worshipers this is no schism but the future of their own maturing baby :-). They’re other steps ahead: vernacular, multimedia, multimode, and smart. This is - in part - already prepared by IAB/IETF through OPES analysis - networked smart virtual boxes.

Balkanisation is probably a necessary phase: it will convince them to study and implement the solution other technologies already tested as appropriate to the digital ecosystem. We could have saved a lot of time, money, people’s pain and efforts: the DNS Integrism is probably the worst enemy of a stable, sure, secure, politically free and innovative DNS. But this is also the real world.

Second response to Suresh Ramasubramanian:

You say that “the internet is just too valuable to fragment.”

That begs at least two questions:

The first is what is the internet?  I always define it as the open system that transports IP packets from one IP address to another.  In that definition DNS a layered-on application.  Your definition of the internet seems to be at a different level. You seem to hold that a single namespace in DNS is the defining characteristic of the internet.  I’d posit that the ability to communicate is a more important, and definitely more fundamental, defining attribute of the internet than a single unflawed DNS name space.

The second question is this: (let’s assume your definition that internet==one DNS namespace) Why is the internet “too valuable to fragment”?

That question requires value judgement or two.  The position you take seems to necessarily require that your values supersede those of other people.  One can take that position.  However one ought not to be surprised that other people don’t agree.

I tend to feel that the more important element is not that there be one holy DNS phone book but rather that any two people who chose to communicate via the net have the ability to move their packets to and from one another.  I’ll leave the question of how they rendezvous up to them.

The position that the catholic-rooters take is disturbingly similar to that of an inquisition: that any deviation from “the” one true DNS is a heresy to be stamped out at all costs.

The keeper of this catholic root zone - NTIA and its stepchild, ICANN - have adopted a highly paternalistic, almost missionary, attitude that is so starched and Procustean and so tilted in favor of a few industrial interests (e.g. intellectual property protection even beyond what is provided by, and often contrary to national laws) that it is no wonder that there are many who wish to go elsewhere.

I use religious terminology in my commentaries on these matters because it is my sense that the arguments raised to support the status-quo, NTIA/ICANN root zone and legacy root servers are of a techno-religious nature.  Clearly they are deeply held, but I perceive them to have little, if any, grounding on demonstrable technical requirements and instead to be value judgements about what the proponents would like the world to look like.  And I tend to see a lot of that world view having a lot of the same utopian unreality that I saw on the streets of Haight-Ashbury during the late 1960’s.

No, Karl .. you use religious terminology because after you left ICANN with extremely bad vibes about what ICANN was doing, what you do is to religiously criticize anything that can be associated with an ICANN activity. I’d almost make jokes about religious wars except that I’ve never really agreed with the rationale behind the medieval crusades, or jehads through the ages (both of which do mean religious war)

Thinking of the internet as merely a layer 1 to 8 set of pipes that facilitate end to end connectivity completely misses the part about layer 9 - the people using it.  It is not just a road system.  So as I said, find critical mass for it. Use it if you like. But sooner or later most such efforts die out entirely because of very few people using it, or several sets of people all building their own roads into the middle of nowhere.

Come to think of it, you can have a road in the middle of nowhere, leading from nowhere in particular to nowhere else in particular. But, you’ll have to find people and cars to fill that road.

Third response to Suresh Ramasubramanian:

You make an assertion that I left ICANN.  That is untrue.  ICANN, contrary to the promises made at its formation, eliminated all publicly elected members to the board of directors.  Had ICANN not eliminated public elections and public directors I have no doubt that I would have been elected for another term.  So it is much more accurate to say that ICANN left the public and its representatives.

But apart from that, you still have not once answered my initial question - Why do you demand that we violate the end-to-end principle and require people to use the NTIA naming system rather than allowing them to use a naming system of their choice?

You have danced up one aisle and down the next avoiding answering that question; so rather than attacking me, please tell us why your name space is of such importance that it is must be imposed on every single person on this planet whether they want it or not.

The internet grew from the rejection of the old dogmas of the telcos.  The opinions you espouse resemble those dogmas in the insistance that users are unable to make choices, that the “correct” use of technology must be defined by engineers, that other choices are to be presumed ill and dangerous.  The atmosphere surrounding the internet is starting to resemble the self-protective cloud created by AT&T and its captive agency, the FCC, during the early 1950’s as epitomized by the Hush-A-Phone
decision.

Neither you nor I are in a position to impose our choices - the community of internet users will chose.  And that is exactly the way it ought to be.

No demands there, Karl. Just a plain statement of opinion.

Start as many as you like. Just see what sort of critical mass they get

Then, once they achieve critical mass, see how soon they start tending towards merging again to get back to a point where they split off in the first place.

This is not an issue of the end to end principle at all. It is an issue of consistency and stability of the domain name space.

At least having a common set of roots means that the end to end principle is maintained by having anybody go from one point to another without modifying his resolv.conf, installing a plugin for his IE or any of the other things that will be required for your pet theory to succeed.

Now for what is again my opinion, and what you will probably consider a grautitous insult, though it isnt intended to be one. 

Rephrase my “you left ICANN” to “when you and ICANN parted ways” if that suits you better. 

Please take your “I hate icann” and “the current operators of the internet are a control-mad cabal” blinkers off for a few moments and go back, read just what I wrote, I’d thank you.

I’m getting rather tired of repeating myself.

Dear Suresh,

Just because you type English, it does not mean everyone has to. The Chinese don’t need to. Similarly for the DNS roots. If you don’t read Chinese, why do you care where their webs are?

Regards,
Colin
P.S.  Japanese *can* read Chinese, the character set is mostly common, but the words are sometimes quaint.

Colin - thanks

Sure they dont need to. But sometimes they’d like to, just like lots of non chinese speakers want to access everything from a recipe for peking duck to cantopop songs.

> The Chinese don’t need to.

Some of them would like to.  And those who do shouldnt be prevented.

> Similarly for the DNS roots. If you don’t
> read Chinese, why do you care where their
> webs are?

Because I work for an ISP that’s based out of hong kong, that has customers that are other ISPs, portals and free webmail providers, around the world, and because I have users around the world?

Sure I do care about that - what my users want is what I want too.  And if I want a great recipe and a detailed history of peking duck, I should be able to google up something like this without having to worry about whether I’ve sacrificed the right number of goats at midnight, or whether my ISP has set the appropriate root servers for china in their resolvers, or whether I have a keywords plugin for that installed

> P.S. Japanese *can* read Chinese, the
> character set is mostly common, but the
> words are sometimes quaint.

Hai, Colin san. Wakarimasen to the extent that I know to feel my way around japan or china, and the waiters and cabbies there cringe at the way I mangle their language whenever I happen to be there.

But they’d still need a gateway of some sort if there was a different set of roots for chinese domains.  And they might then have to add a different gateway for any domains that happen to be hosted just across the south china sea in Taiwan.  Add another gateway for korean sites (maybe a keywords plugin for your browser as well).

Interesting. I think your friend Tim betrays a lack of technical knowledge by pointing to IPv6 as a problem. Many Asian ISPs already are IPv6, most networks are dual-stack, and ICANN announced last year that all root servers can accommodate IPv6.

No doubt many of you know the Chinese market better than I do. But I think it’s multi-lingual that may create the separate Chinese Internet the government wants, all while positioning it in a way to stroke nationalistic fervor. A quote from the Director of CCNIC:

The Director General of CNNIC, Mao Wei, said, “We have already established a framework of registrars nationwide in China, and we and our registrars have been very active in assisting corporations in China, in particular the small and medium enterprises, to understand the need to claim their domain name resources that are completely in Chinese characters. Within China, the recent launch of Chinese Domain Names has attracted many corporations in registering their own Chinese Domain Names. To handle the growing demand and interest, 15 states, including Beijing, Shanghai etc. have been designated as regional centers in the effort to continue educating the public.” He added “Now that China has become one of the world’s most dynamic economies, we believe that foreign multinational companies can expedite their marketing goals in China by registering these Chinese-character domain names since the emerging Chinese Internet consumer is predominantly a non-English speaker and strategically best approached via the Chinese language”.

It is therefore natural for us to ensure that foreign entities who wish to protect their Chinese language domain names in the .公司 and .网络 extensions are able to participate early on in the process. Therefore, we are very pleased to partner with i-DNS.net to bring this early opportunity to people outside of China now.”

“It is heartening to see that 20 years after the Internet Domain name system in English was invented, and 7 years after we at NUS invented and pioneered the concept, and 6 years after we at NUS conducted an Asia-Pacific deployment testbed with a dozen nations and languages, and after 5 years of lukewarm commercial deployments in various parts of the world, with the majority being a misguided attempt at trying to get multilingual peoples to accept two-language hybrid domain names (e.g. the Chinese-English ‘multilingual.com’ names from Verisign), finally we have broken the political logjam and witness a major community ? the Chinese ? turning the promise into daily Internet reality” said Prof. Tan Tin Wee of NUS, who pioneered this technology at NUS in 1997 and 1998. Dr. Tan is widely acknowledged as the father of the modern multilingual domain name movement and was also the former chairman of the Asia-Pacific Networking Group, APNG, which under his chairmanship ran the Asia-Pacific IDN testbed in 1998/9.

I can only regret that all these good intents are just underlining the need to both partition the Internet to address the demands of nations (like China, and Chinese names are not only registered in China, since foreign entreprises wanting to operate in China, find with them a way to get accessed sites dedicated to the Chinese market; but most of all USA who just took-over the root, as Jon Postel did once), and the necessity to keep a single, stable, united internet.

But again and again, united, stable and single does not mean uniform!! To the countrary! Please Suresh and Karl, stop disputing this non-sense Nessie story. Alt-roots are all over the place and they are declared legitimate by ICANN. Please read ICP-3 and let proceed. Anyway, alt-roots the way they are currently understood are NOT the solution. We face today a forest and not a root. The problem is to keep it organised and concerted.

Karl, frankly, the Internet Church of the Legacy Days is not that much in ICANN but in the IETF. Suresh, that Church includes the worst fundamentalists I know. And I weight my words considering their worldwide impact on economies, cultures and most probably life of people. And then? Everyone who keeps disputing rather than proposing a true solution (actually probably a set of solutions), delays it and its huge benefits too. In a way heis worse than them, because at least they - sometimes - believe they are protecting the world, at least their world. I indicated my solution and I work at it. Either we work together on it, or you propose a better one and I work with you.

Christopher, we certainly have identified the need for multilingual names. Now we need scalable solutions. The IDNA is certainly not a serious solution, even if it may help many to understand what we need in ... not delivering it. Co-root are not a solution either if they are not correlated together - and this will not be done by ITU, UN or anyone else. This will be done by ICANN or by the people, and their Governements, as a grassroots process - the USG being the first one (this is already done). Now, we have to deploy true and good solutions. IMHO this is too late to avoid a certain balkanisation and mess. But this may help in shocking IETF geeks, and making them move.

For example, nothing will prevent individual roots or “alias spaces”. But we need to understand how to correlate them, to support them, to serve them, so they are a stabilisation of the network rather than the end of the DNS.

Please let stop all these arguments, and let work all together to the developmennt of the MGN (Multilingual Global NGN).