[discuss] The decentralization of IP addresses

[willi uebelherr]

Dear friends,

this is the contribution from Jefsey on the Governance list to this 
discussion.

many greetings, willi
Coro, Venezuela


At 18:59 29/11/2015, Jean-Christophe Nothias wrote:
>Third, challenging the Internet architecture seems to be a red line,
>something that no multistakeholder/status quo champion could ever
>discuss, debate, think of. They should think twice. And not because
>of the ITU,  but

>1. because of the US obstructive stance,
>2. and because technology calls for innovation and disruption

JC,

Let me be clear about this in order to not create unnecessary confusion 
or dispute. The red line is about the Internet medium layer architecture 
vs its Catenet basis. This is not a question of technical dogma but 
rather of technical focus, options, experience, capacity and 
societal/political stability.

A. I will explain why it is a UNIX/NETIX perspectives opposition
B. I will shortly explain the root of the confusion
C. I will explain the current open trend
D. I will eventually consider Willi's position


A. the UNIX/NETIX opposition

The internet (cf. IEN 48 
<https://www.rfc-editor.org/ien/ien48.txt>https://www.rfc-editor.org/ien/ien48.txt) 
has architected the Internet project as the ARPANET catenet, along the 
Louis Pouzin terminology understood by ARPA as "roughly [meaning] "the 
collection of packet networks which are connected together."" Vint 
explains that it is not enough for a practical implementation and sets 
the objectives and constraints of an ARPANET catenet internetting. He 
then documents his own objectives. There are two targets and one key 
contribution.


1. Vint Cerf's first objective (specific to the internet use of the 
catenet):

to permit the internal technology of a [TCP/IP] data network to be 
optimized for local operation and to be readily interconnected into an 
organized catenet. This means that everyone must use an inter-network 
optimized technology. And the IP local addressing scheme must extend to 
the global network. This differs from the two other parallel contributions:

1.1. by the ITU: to build a catenet through an external technology 
(X.75) optimized to support local technology interconnections, with a 
local technology (X.25) optimized to use that international technology 
and using its global purpose X.121 addressing scheme.

1.2. by Tymnet (which was the only internationally used technology from 
1977 to 1987) of which architecture used a meta-technology and 
addressing scheme approach to interconnect every local and international 
protocol and addressing scheme, and eventually services (my responsibility).


2. Vint Cerf's fundamental contribution

This is Vint's main contribution because it is universal. He states: 
"The term "local" is used in a loose sense, here, since it means 
"peculiar to the particular network" rather than "a network of limited 
geographic extent." A satellite-based network such as the ARPA packet 
satellite network, therefore, has "local" characteristics (e.g. 
broadcast operation) even though it spans many thousands of square miles 
geographically speaking." This, together with Louis Pouzin's catenetting 
actually defined glocality as a local virtual network global reach. I.e. 
what I call a VGN (virtual glocal network). This is something difficult 
for IETF people to consider because they are only referred to twice in 
RFCs as being outside of the "end to end".

2.1. In RFC 1958 (internet architecture): "The network's job is to 
transmit datagrams as efficiently and flexibly as possible. Everything 
else should be done at the fringes."

2.2. In RFC 5895 (mapping characters in IDNA2008): "It should be noted 
that this document does not specify the behavior of a protocol that 
appears "on the wire". It describes an operation that is to be applied 
to user input in order to prepare that user input for use in an "on the 
network" protocol. As unusual as this may be for a document concerning 
Internet protocols, it is necessary to describe this operation for 
implementors ... This because local typing, etc. is to be supported by 
local subsidiarity in order "to reduce the surprise for users and is 
likely to be slightly (or sometimes radically) different depending on 
the locale of the user".

This means that mapping/unmapping is to happen outside "of the wire" 
(end to end), at the fringe. IDNA2008 may call for fringe to fringe 
operations. Those are "OSI presentation layer six" operations. However, 
there is no "presentation layer six" in the internet layer stake. Hence, 
the possibility of presentation layer six based "network application 
services". I called them "Extended Services" in 1984, when I created the 
Tymnet/Extended Services department. They came above the TCP like 
internet "value added" services, above the IP like catenet "basic 
services". Problem: one of the T/ES services was to transparently map 17 
million IP addresses (RFC 923) to X.121 addresses in order to deploy its 
global applications as network open services rather than edge 
proprietary businesses.

This extended addressing service not only allowed global competition 
with the US, but it also put the Internet in bad shape because the 
presentation layer six concerns security, languages, and intelligent 
exchanges (through formats). The issue was architectonical: was EDP to 
be computer (US UNIX industry) or network (Tymnet/PTT NETIX) centered?

In addition, it was a national security issue: UNIX internet systems had 
no protection against non-US protected accesses established through 
T/ES. It was also a strategic political/industrial issue within the 
deregulation context where AT&T was dismantled, killing its own X.25 
technology development. Tymnet then started proposing X.75/TCP/IP/Tymnet 
technology to the seven FCC regulated "Baby Bells" (Regional Bell 
Operating Companies) the same as they had already leased and operated 
the five FCC regulated IRCs (international records carriers) and all the 
other foreign Operators and PTTs (except BT) throughout the world.

As a result, the Tymnet 100% parent company was purchased by McDonnell 
Douglas, the military-industrial leader of the time. They closed my T/ES 
by mid-1986 and sent their own people to the first IETF meeting. The 
digisphere had to be NSA-compatible.


3. Vint Cerf's second objective

Vint Cerf's second motivation was "to allow new networking technology to 
be introduced into the existing catenet while remaining functionally 
compatible with existing systems. This allows for the phased 
introduction of new, and obsolescence of old, networks without requiring 
a global simultaneous change."

In blunt and clear words, it was to seamlessly expand new TCP/IP 
features in order to compete and overcome the Tymnet and ITU's X.75/25 
capabilities.

In other words, to do worse than me! They fired me and froze Vint's 
ambitions: this was the "statUS-quo" strategy.

However, Vint Cerf persisted, created ISOC, chaired ICANN, and joined 
Google. His TCP/IP technology was more adequate to handle open source 
than Tymnet and more versatile that the ITU approach. It called for 25 
years but he eventually reached the verges of his second objective, 
powerfully threatening the status quo. While the WCIT was to show that 
the various foreign NSAs objected to the US NSA's global surveillance.

As a result, the State Department supported the ISOC/GSN cooperation, 
the OpenStand statement, the minority vote in Dubai, the Snowdenia, the 
NTIA statement, the Lynn St Amour/Don Tapscott report, etc. that 
eventually led to the ICANN reshuffling, and to Jari Arko's 
<https://www.ietf.org/blog/2015/01/taking-a-step-towards-iana-transition/>2015/01/08 
blog post stating :

"Our work is not yet complete. There are a number of steps still in 
front of us. They include the following:
     * Both the numbers and names communities need to complete their 
proposals. We at the IETF will continue to engage with them with their 
work, just as they assisted us with ours.
     * Later, the IANA Transition Coordination Group 
(<http://ianacg.org>ICG) will assemble a complete proposal and gather 
community feedback on the result. When ready, they will submit the final 
proposal to the NTIA.
     * The NTIA must then consider and approve the proposal.
     * Finally, it must be implemented. "
An IETF/WG on the IANA Transition had supported this IETF allegiance to 
the NTIA, hence its decision to become the technical body of the 
"GAFAMUSCC" RFC 6852 "global community" embracing "a modern paradigm for 
standards where the economics of global markets, fueled by technological 
advancements, drive global deployment of standards regardless of their 
formal status. In this paradigm standards support interoperability, 
foster global competition, are developed through an open participatory 
process, and are voluntarily adopted globally. These voluntary standards 
serve as building blocks for products and services targeted at meeting 
the needs of the market and consumer, thereby driving innovation. 
Innovation in turn contributes to the creation of new markets and the 
growth and expansion of existing markets."

I appealed this with the IESG and IAB in order to make sure that this 
was the true consensual decision of the IETF. 
<http://iuwg.net/index.php/History#Evolution>http://iuwg.net/index.php/History#Evolution. 
With the consequence documented there: the decision to start an XLIBRE 
(<http://xlibre.net/>http://xlibre.net) RFC 6852 global community for 
those wishing to research and test aside from the US/Google technically 
correct use of the world digital ecosystem. Thinking of themselves as 
their own VGN masters or Intelligent, Internet Users (IUsers), rather 
than as ICANN (IN) DNS Class consumers.


B. The root of the confusion

The root of the confusion is that, as I indicated it, the ARPA, Tymnet 
and PTT models covered both the lower and upper layers. Therefore, 
people used PSS (packet switch PTT services) and the Internet as a 
global digital solution without differentiating the layers.

This was increased by the habit to confuse the internet and the web.

Today, most of the World Digital Ecosystem Governance considerations are 
internet centric. With laws around the world not making a difference 
between "Internet", "the Internet", "the internet" and "internet" for 
what is actually the Catenet Model for the ARPANET-internetworking.

In addition, there is an addressing system confusion between centralized 
(Copernican, geocentric), decentralized (Newtonian, heliocentric) for 
what is distributed (Einsteinian, cosmological).


C. The current open trend

There certainly is a US effort to build on the 1986/2012 momentum to 
keep industrial, commercial, and political control of the WDE (world 
digital ecosystem). However, experience and technology show that this is 
a BUG. This bug is to want to "Be Unilaterally Global". This was a 1986 
misunderstanding due to the 1977 push to the international catenet given 
by the FCC (VAN license to Tymnet and Telenet and naming to Tymnet). 
However, this was only a US "go" that matched the European "OKs" 
simultaneously gathered by Robert Tréhin (head of the Tymnet European 
Operations, TEO) in building the public catenet. In essence, an 
international network is multilateral. Communications' multilaterality 
is managed by the ITU.

When we interconnected the Internet catenet to the public global 
catenet, the US had to protect it from the non-UNIX systems. This led to 
the US strategy of replacing the ITU by the NTIA.

Firewalls have been deployed. IAB has eventually engaged in working on a 
secure protocol stack. It is time now for the BUG to be fixed.

Actually this is urgent. Before the IoT deploys significantly because, 
by nature, it has to be BUG proof. No one anywhere in the world wants 
their fridge to be under US NSA and Google surveillance.

However, the way it is made must not hurt the network development and 
stability. So Vint Cerf's second objective can only be deployed by 
subsidiarity. To develop and deploy additional compatible services that 
will eventually be able to replace the existing architecture.

The XLIBRE trend seems to be to:

1. capitalize on IP for the catenet for the basic services.
2. consider alternatives to the internet "TCP" added value. This has 
started with XMPP, named content networking, SDN, etc.
3. develop and deploy "intersem" extended services experiments for a 
multi-vendor LIBRE (LIBRE even of the Libre) smart interoperability.

The interest of this is that it respects the experience acquired since 
the late 1960s. This is what I call "reconsiderative" innovation, which 
is neither "incrementative" nor disruptive. If I would start it all 
again, how would I do it, now that I can use all that I have learned and 
others have developed since then?


D. Willi's position

More often than notWilli's position does not make IETF technical sense.

There are two ways to react:

- In bashing Willi for his lack of technical understanding.
- In protesting against the IETF for not publishing standards that:
--- Either permit developers to meet Willi's technical needs
--- Or documenting their RFCs within a graded framework the first layer 
of which is understandable by Internet Users (IUsers) without needing to 
be smart Intelligent Users (IUsers).

You will note that my language uses the same term ("IUser") in both 
cases, and in many other cases such as is the case for 
(http://xlibre.net/index.php/IUse) individual, informed, independent, 
innovative, industrious, inventive, insatiable, imaginative, impartial, 
impecunious, inevitable, inflexible, insisting, insupportable, 
inexhaustible, ingenious, interactive, interdisciplinary, interested, 
interrogator, interventionist, irreducible, irritating, etc. lead user.

For a merchant, the customer is the problem, for a technician it is the 
user. And the customer and the user are king. I am not sure I understand 
what Willi may ask, but I am sure I am to consider it carefully because 
IETF and technically correct people are sustainable and 
"incrementative", Willi asks us to be disruptive and what I call 
reconsiderative. By the way, this is what the IAB reconsideration of the 
protocol stack is exactly about.

The real need is for all of the Willis of the world, which Willi 
represents, feel that their dataspheres are secure, at ease with their 
networking experience, and the master of their own glocal digitality.

jfc