Annex 3. A practical innovation and complex deployments strategy

From IUCG - Internet Users Contributing Group

The current work of WG-IDNABIS leads one to think that IDNA might exemplify a practical IETF innovation strategy towards a stable and faster deployment in spite of the complexity of the Internet system.

.

Contents 1 Introduction 2 Rough model introduction 3 The model gap 3.1 The three ways to fill the gap 3.2 The IETF options 3.2.1 There is plenty of architectural room for innovation 3.2.2 There is not so much room 3.2.3 There is insufficient room 4 Taking advantage from the difficulty 4.1 Indirect enforcement 4.2 How to obtain it 5 The example of the WG-IDNABIS case 5.1 Users' ML-DNS concept contribution 5.2 The practical risks 5.3 The so-called impossible multilingualization of the Internet 5.4 Other areas for a similar strategy 6 Security Considerations 7 IANA considerations

Introduction

The Extended Relational Space Model (XRSM) that the IUCG (Internet Users Contributing Group) is working on, has a distributed, systemic, and user centric network architecture approach that could extend the RFC 3935 core values. A first draft has been submitted on the precautionary principle. and the general introduction work has been engaged, which should help better understand the positions that will be used in its framework, and the resulting characteristics of its "Internet PLUS" model.

However, this work requires time and reflection. The current WG-IDNABIS debates show that some earlier considerations could be of use in order to explore how the IDNA situation might lead to, and exemplify, a practical strategy to obtain a stable, innovative, and faster deployment of the IETF propositions in spite of the Internet system complexity.

Rough model introduction

Among the other fundamentals, the XRSModel considers three main strata in e-semiotic (usage of signs to uter semantic meaning) communications:

• physical, plug to plug interconnectivity - the telecom strata.
• logical, end to end interoperability - the datacom strata.
• nootical, mind to mind interintelligibility - the metacom strata.

The "decentralized, end to end, dumb network, protocol on the wire" Internet makes up most but not all the logical strata. The DNS includes some semantic aspects that belong to the nootical strata. Jon Postel identified it when he forbade namedroppers from discussing and confusing semantic and protocol issues (at that time they discussed the English meaning of the TLDs tags).

The model gap

Along this model, there is a gap to fill between the top of the Internet and the bottom of the Intersem (Multilingual and Semantic Internet) metacom layers. This gap minimally concerns:

• the missing Internet session and presentation layers.
• extended services that the OSI model did not consider and that were included in the first architecture of the international network.
• new requirements that we (will) learn from the Intersem analysis and testing.

The three ways to fill the gap

There are at least three ways to fill that technology gap.

• The usual way of the IETF, which is in charge of the entire existing billion systems Internet legacy. It is detailed below.
• the various proprietary ways, integrated services providers or alliances (to differentiate from basic access service providers), will naturally push forwards through business coopetition.
• the way lead users call "Internet PLUS", standing for "parallel layered usage systemic". It consists, for each layer, in proposing those users with one or several alternative modules that extend the IETF architecture parts the user needs. This layer (alternative) extended building block approach will probably be more agile because it is not bound to constraints that the user may not be interested in, such as backward compatibility or interoperability that he never needs. Internet PLUS building blocks will , otherwise, have to be completely interoperable with the IETF architecture and solutions. Yet, they will add extra features, the total pile of which may result in the missing services.

The IETF options

The IETF architecture is usually confronted with one of the following three cases:

There is plenty of architectural room for innovation

This can result in a network technology enhancement if the three (IETF, ISP's, and User's) approaches described above can agree upon a solution.

There is not so much room

Therefore, a new deployment is usually constrained on a short-sighted basis (cf. RFC 3968 rationale by IAB): large commercial ISPs are then the probable way to finance and deploy. This will be in the hope that the situation will unlock further on, one way or another. This may result in long term detrimental impacts and major network unbalance, or even balkanization. The users, being less represented, cannot oppose this easily. This is even more so the case if the propositions are confuse and call for long debates. When propositions are confuse, and may bias the network context, the debate may become detrimentally controlled by some commercial interests.

There is insufficient room

There is insufficient room within the IETF end to end decentralized network territory. Therefore, the IETF must go beyond the network border, in turn trespassing and venturing in the user application area, where it has no enforcement capacity and can only engage the credibility of its image, while its documents may be disrespected by some.

Taking advantage from the difficulty

This last situation may turn out to be very interesting if it is adequately and carefully handled. The reason for this is that in such a case, the IETF usage application layer recommendations become legitimately enforceable and, taken as such, as the default specifications, the Internet PLUS applied answer will respect and impose via its usage. The enforcement capacity of the Internet PLUS is that by nature it remains within, but it also enlarges the network ends.

Indirect enforcement

As a result, every user who chooses to adopt an Internet PLUS approach de facto enforces a mandatory respect of the IETF positions to those wanting to interoperate with his machines.

This does not prevent special usages in special usage cases by informed users (by essence, Internet PLUS users should be more informed or be specifically assisted users).

This additionally assists a progressive adhesion, transition, and technology enhancement to the point where ISPs will probably find their best interest in joining the bandwagon, in turn adding credibility to the whole proposition.

If the IETF proposition is inadequate, the Internet PLUS users will provide an earlier warning and probably more documented information.

How to obtain it

The only constraint for the IETF is to make sure that it keeps a specification balance. This means that it should not

• make its specifications overly specialized in favor of any class of ISP interests, Internet PLUS developers, or users. Therefore, the specifications do not become overly restrictive for the other classes.
• over delay them, precluding Internet PLUS developers to develop their extensions as interoperable ones.

In the same way, IETF documents should not be overly general or become confuse. This would lead one of the categories to take the leadership in imposing de facto its own solution, all to the disadvantage of the others. This phenomenon is common and is known as "internationalization" vs. "multilateralisation" (multilingualization in the language area), as the unilateralization of one single standard instead of a multilateral normalization. This may help in the short term and survive the medium term but is likely to lead to earlier obsolescence and costly redesigns.

Actually, IETF WGs should consider publishing an explanatory rationale. This rationale could include advice for application and Internet PLUS module developers as well as for a best practice for users. This would prevent architectural rigidity, standardization tensions and governance conflicts.

The example of the WG-IDNABIS case

Coming from the existing IDNA RFCs, the WG-IDNABIS discusses IDNA updates "for the Internet to work better", whatever the Internet governance may decide.

It is in this way that it trespasses the network ends and encapsulates the Internet into a sort of architectural "advised practice layer" that Zone managers and people can disregard, in turn risking the trust that users have in the IETF.

Users' ML-DNS concept contribution

This is why the various ML-DNS projects may turn out to be the best way to implement and enforce IDNA and progressively turn the Internet into a truly Multilingual network. However, an ML-DNS architecture should be considered as a part of an Internet PLUS framework in order to be of full and quick use.

In this WG, the group of Internet Lead Users initially asked the Chair, on behalf of many others users from various language oriented mailing lists, about IDNA. They asked whether he intended to make it the "ML-DNS" (whatever it may be and that is to be discovered) that the WSIS expects, or to have a progressive strategy towards it.

The practical risks

IDNA exemplifies the various practical difficulties that an IETF decided upon innovation may meet.

• a charter not taking into account ML-DNSs (whatever it may be that the users want) as an opportunity to adequately use it.
• delay in producing a stable document (a few months to possible several years of documentation instability)
• a set of complex documents.

The so-called impossible multilingualization of the Internet

This was because one cannot bundle the Legacy Internet with multilingualization: Jon Postel understood it, RFC 3536 does not even want to use the word, and Unicode is not ready for it.

This is why a user-oriented IDNA uncoupled yet interoperable multilingual ML-DNS (whatever the users want it to be) is of interest. In using it, the Internet user may immediately, progressively, optionally, support the Intersem. In turn, the Intersem will probably be an IDNA killing client.

Other areas for a similar strategy

The same approach could most probably be used for v6 and DNS security.

Security Considerations

The Internet PLUS is not an alternative technology. It is not even a technology evolution. It is a concept whereby when a functional extension is possible and further on needed in the pile, it can be implemented upon the users' decision, even if it locally reduces some global constraints, as long as its real usage (may be not its full capabilities) remains fully interoperable.

This concept may obviously raise security concerns and lead to a discussion about new security related solutions. However, such concerns and extensions should be discussed for each of them, one by one, prior to implementing each of them. Furthermore, its investigation includes the documentation of an "Interest" charter for the secure and responsible use of the Internet as its own test-bed.

IANA considerations

The concept of Internet PLUS is based upon a distributed vision of the Internet systemic diversity. This also concerns the IANA that is to be used as part of a Multilingual Distributed Referential System. However, Internet/Internet PLUS interoperability calls for a strict respecting of the IANA tables and parameters.