[discuss] Real world Impact of multiple roots

Jefsey jefsey at jefsey.com
Mon Feb 3 14:42:19 UTC 2014


Ouahou!!! Michel,

Would this /1net mailing list become interesting???
Comments are interspread.

At 13:17 03/02/2014, Michel Gauthier wrote:
>We are in total agreement. When I use the word "intelligent ways" it 
>is the IUse (intelligent Use which refers to technical use 
>[hardware/software] as opposed to usage that refers to community brainware).
>>The words are a technical statement: if the root splits, users
>>might be directed to a different web site than the one originally
>>intended to have a given URL.
>Absolutely.
>>There is no judgment in that text,
>>nothing about "freedom" or seeing "what one does not want to see."
>>Those are your words, not an implication of the IAB's words.
>These are my words and RFC 6852 words. The RFC 6852 modern paradigm 
>implies that the running code (documented by IAB) has met a living 
>mode success (resulting in global communities and a huge bounty). My 
>evaluation, or historic judgment, is that this was accepted by usage 
>for due to the capacity of flexibility in its use.
>>FYI, the IAB's words originated as a formal comment to ICANN
>>sent in September 1999 (under my name, as IAB Chair at the time):
>>https://mailarchive.ietf.org/arch/msg/ietf-announce-old/ut9aeZoSbY4_GxMdmhyEDBX_Dvg%3D/

Well we can thank Brian!!!
And Mike Roberts for the question and ICP-3.

I immediately pasted at the IUCG at IETF to be used in the debate on 
digital security and awareness 
http://iucg.org/wiki/Main_Page#Digital_Security_and_Awareness.
http://iucg.org/wiki/19990928_-_IAB_Technical_Comment_on_the_Unique_DNS_Root

>Deep thanks for this input. Now, we know where the problem comes 
>from. I suppose that it will considered on the IUCG site?
>As an analyst, however, I see where is the point of controversy. You 
>start in assuming:
>1. The Internet, to remain a global network, technically requires 
>the existence of a globally unique public name space.

Correct. EVERY communication requires that. This why we have natural 
languages. crown-root is the same as telling the grassroots: "quote 
your languages in ICANN English". The problem we often face is that 
we are confusing the information theory and the communication theory 
we miss. http://en.wikipedia.org/wiki/Communication_theory

They quote Shannon who creates the problem when he said ""The 
fundamental problem of communication is that of reproducing at one 
point either exactly or approximately a message selected at another 
point.". This is not the problem of communication but of end to end 
transmission. In writing this he limited communication to passive 
content and forgot fringe to fringe active content and overall 
intelligent content.

The internet is a transmission system between peers. In the initial 
phase these peers were networks. In the second phase they are 
relational spaces organized with different technologies.

>2. The DNS name space is a hierarchical name space derived from a 
>single, globally unique root.
>3. This is inherent in the design of the DNS system.
>Point 3 looks related to 1 and 2, but is only related to point 2, 
>because there is no logical relation between 1 and 2.
>- as a Anglophone you have a crown-down hierarchical pragmaticaly 
>local vision you translate into the "unique root" concept,
>- as Francophone I have an equalitarian heterarchical metavision: I 
>can only read point 1 through a "unique grassroot" concept.
>Would this memo of yours have been published as an ISO draft, i.e. 
>bilingually, it could not say this. This would have saved us many 
>hours of misunderstanding: the controversy would have clarified by 
>then as it has now. Because the DNS works, until now. This now is 
>when ONS, e-sovereignty issues, Snowden, etc. issues are raised to 
>be discussed in Sao Paulo.

I think this is something that every Francophone participating into 
the ISO standardization process can confirm and we saw applied when 
ICANN attempted to "internationalize ISO 3166", i.e. globalize ICANN 
through standards, as it tries now to do it through an international 
meeting. (I think this is an episode they don't want me to harp on 
:-)). What is fun is that it resulted from a BoD decision in ... Rio.

>This being said: point 1 and point 2 are perfectly true. They are 
>orthogonal. This means that point 3 is not a deduction but an 
>implicit induction. The root inherence is not in the name space but in the DNS.
>
>In any case your quote and your passed text is IMHO the most 
>important contribution to the Sao Paulo preparation. It justifies 
>the ICANN's role IRT. the IN-root, and the complementarity in the grassroots.

Well, this is, IMHO, as you partly mention it at five strata:

- interconnect (basic services)
- internet (value added services)
- intertech (extended services)
- interuse (applications as a protocol)
- intersem (semiotic facilitation to intercomprehension).

Each entity has its subsidiary role and responsibility on a necessary 
grassroots basis, because this is an open network not army.

>> > Please do your home work on this issue before showing yourself entirely
>> > out of scope; One thing you can do is to read the ICANN authoritative
>> > position on the matter.
>> > http://www.icann.org/en/about/unique-authoritative-root.
>>The IAB was glad to see that at that time, since it entirely agreed
>>with the IAB position.
>Absolutely: "As described in a recent proposal within the IETF, this 
>"class" facility allows an alternate DNS namespace to be operated 
>from different root servers in a manner that does not interfere with 
>the stable operation of the existing authoritative root-server 
>system." Moreover that this recent proposal was introduced by your 
>successor as the IAB chair.
>ICANN there clarifies the point of controversy. In your position and 
>RFC you presuppose that the DNS system is the only possible system 
>to navigate the global name space and that it has only one authoritative root.
>ICANN maintains that the DNS covers the entire name space, but adds 
>that it also permits multiple non interfering visions of this name 
>space (classes). This confirms that it adheres to the grassroots 
>concept and that it avoids the interferences with its own root.
>My part is not technical, only logical. It is not up to me to say 
>that the DNS as documented by the IETF can do it, but what ICANN 
>says is that its current policy enforcement is based on the belief 
>that it can do it.

Frankly no one knows if the DNS can sustain the entire name space. I 
believe yes because ITF people are good at spotting end to end flaws. 
The only difficulty is the ambiguity of the classlessness that has to 
be reconsidered in a fringe to fringe context, i.e. outside of the 
internet: at the intertech level.

To know it there is only one solution: to test it. I carried for 
nearly two years a community test ("dot-root") as per ICP-3, after 
the IETF people explained me that their job was not (as ICANN had 
suggested it) to carry tests. This has not disrupted the internet, 
and has shown that the "crown-root" could be concerted and that the 
crown-space could be democratically organized by its stakeholders at 
the root administrator level. Now the "HomeRoot" testing is to be 
carried to check at grassroots level if the DNS stayed consistent and 
what are the real advantages in terms of speed, security and control 
and disadvantage in terms of governance.

>> > If you want to
>> > know more about the internet as it is and disucss it less as it is not;
>> > http://caida.org.
>>Which particular aspect of CAIDA's view of the Internet do you mean?
>>There's a vast amount of information there and at other measurement
>>sites.
>
>It is just a link among others where to learn some political 
>humility when facing the technical reality as other CS sites remind 
>us the same political humility in front of the societal reality, 
>or  business sites in front of the economic reality.

May be could you look at https://www.dns-oarc.net/ ... NB.: as always 
the money barrier interferes with the technical aspects. But if you 
have an US grant ...

jfc






More information about the discuss mailing list