[governance] What is a: necessary/efficient/ultimate Number

[Karl Auerbach]

yehudakatz at mailinator.com wrote:

> The question is:
> With encompassing the whole realm of iGovernance, How many People (Seats) are
> needed to run the Internet?

Running "the net" requires several tasks at several levels (global, 
regional, local, personal.)

The most important task is that of maintaining IP level connectivity. 
Although that is largely nearly automatic, it does require the 24x7x365 
cooperation between many (hundreds) of autonomous routing authorities 
(often called ISPs).  That system is largely driven by enlightened self 
interest.  However, as I mentioned in a previous post, there is an 
potential future need for there to be external mechanisms to provide for 
end users, or their agents, to obtain end-to-end assurances across a 
sequence of providers.

DNS is operated by several different groups.

At the top by perhaps one quarter of a full time person who must daily 
prepare and inspect a root zone file.  That file has, on average, 
perhaps one change per day, and it is quite small (less than 20K bytes 
compressed) and easily disseminated.  The harder part is validating it, 
which is easily done with digital checksums (digests).

At the root layer there are 13 distinct entities, some of only a few 
people, who operate the 13 root servers.  Many of those servers are, 
thanks to the innovation of those operators and not to ICANN, are now 
replicated around the world using a technique known as "anycast".

These root server operators are entirely independent of ICANN.  Those 
operators are free to sell their positions, to date mine the query 
stream, to give preferential (or impaired service), or even to cease 
operation altogether.  These people do a very good job, but their only 
obligation is self imposed.  ICANN has promised to oversee these folks 
and establish enforceable service level obligations, but has failed to 
do so, thus leaving network users in the lurch.

There are many other jobs needed to run the net on a day-to-day basis, 
but those are typically localized, such as maintaining the contents of 
lower-layer DNS zone files, configuring and monitoring the zillions of 
local naming and security/access control systems, etc.   The life of a 
network sysadmin is never quiet.

The vast bulk of the things talked about in the context of internet 
governance are things that have absolutely nothing to do with the 
day-to-day or even the year-to-year operation of the internet.

For example, the entire mess that ICANN has created about top level 
domains is a construct that is entirely political and protective of 
certain status-quo interests (such as trademarks.)

Imagine, for example, a supermarket.  It has many products on its 
shelves.  Imagine another supermarket; it too has many products on its 
shelves.  For the most part the products are from the same vendors. 
There are a few boutique products that are sold only in one store or the 
other.

DNS is like that.  We go to a root, just as we go to a supermarket, and 
select the top level TLDs that we want - .com, .net, .org, etc...

If a new vendor comes along and has a new product, that new vendor has 
to convince each supermarket manager to carry that new brand.  Whether 
that newcomer obtains a place on the shelves of most supermarkets 
depends on both the efforts of the vendor and the acceptance by the 
public.  It's called "building brand recognition".

If there is a vendor of a brand, let's call it "Coca Cola", has built 
its brand and some newcomer comes along and tries to label its products 
as "Coca Cola" there are mechanism in the existing legal system to deal 
with the issue.  The supermarkets need not be involved and there is no 
ned of some global authority over supermarkets to supplant the workings 
of the trademark enforcement system.

The same thing can occur in DNS.  However, it is ICANN's position that 
there must be exactly one supermarket.

Were we to recognize the fallacy of that dogma we would see that nearly 
every aspect of internet governance done under the ICANN banner is work 
that is not needed, is excessive regulation, anti-competitive, and 
duplicative of the existing legal system.

You mentioned ENUM - apart from the fact that a lot of VOIP telephone 
numbers contain no numbers at all but, rather look like email addresses 
- ENUM does not seem to be taking off.

Most of the effort to administer ENUM will be a front-end effort to 
establish the delegations to the countries.  After that point it will 
require minimal maintenance on a daily basis and the workload will move 
to the national and local provider levels.  And that will require a 
number of people with a lot of computer assistance.  It will also 
require a huge troubleshooting staff: regular expressions are hard even 
for experienced people; they can become a nightmare and a source of 
many, many operational problems.

There are other matters - IP address allocation requires a few smart 
people.  The RIRs have a good handle on this.  Right now ICANN's 
(actually IANA's) IP address policy can be described as "when a RIR 
asks, ICANN/IANA grants".  That layer can be eliminated and left to the 
good efforts of the RIRs working together (which they do rather well 
already.)

IP protocol parameters, another IANA job, is largely simply assigning 
the next number in sequence.  There are some more intricate allocations 
that do require some thought, but the IETF has created a system of 
experts to help out with that.  All in all we can figure that protocol 
number assignments can be done by a handful of people.  And it is a job 
without political overtones.

The job of recognizing ccTLDs is something that is akin to recognition 
of what and who is a sovereign.  That's something that is highly 
political.  ICANN has done a bang-up job on that - recognizing, for 
example, a purported ccTLD operator on the basis of a handwritten 
document on an otherwise blank sheet of paper, and bearing an 
unauthenticated signature of an unknown person.

As for non-ccTLDs: That can be readily mechanized by creating an 
inexpensive and very lightweight system of applications that requires 
the applicant to promise to adhere to widely accepted and used written 
internet technical standards.  If they so promise, their application is 
accepted and they enter either an auction or lottery to compete with 
other applications for some number of TLD slots that are made available 
on a periodic basis (I've suggested 10,000 a year, or one million over a 
century, but even if it were 1/200th of that, or about one per week it 
would be a vast improvement over the current stasis.)

I hope that this answers some of your questions.

Certainly what we have seen to date is massively bloated over what is 
needed.  That is the result of the capture of the existing bodies of 
internet governance and their transformation from internet governance 
into bodies that protect incumbent industrial interests.

		--karl--
____________________________________________________________
You received this message as a subscriber on the list:
     governance at lists.cpsr.org
To be removed from the list, send any message to:
     governance-unsubscribe at lists.cpsr.org

For all list information and functions, see:
     http://lists.cpsr.org/lists/info/governance