Home / Blogs

The ITU and IPv6 Transition: Controversy at the IGF

McTim

At today's "Managing Critical Internet Resources” session of the Internet Governance Forum 2009, the ITU agenda on taking a role in IPv6 distribution once again reared its ugly head.

In a heated exchange, Professor Dr Sureswaran Ramadass, the Director of Nav6 an ITU consultant/fanboy squared off with the new ICANN CEO about competition in IPv6 address distribution. It seems clear that the ITU hasn't given up their hope of becoming a player in distributing IPv6 addresses to their members.

His question, setting off the contretemps, was "Since there are so many IPv6 addresses, why can't we have additional organisations giving it out?" (start at about the 53rd minute of this video)

The answer, of course, is that we CAN have additional organisations, but we SHOULD NOT. There is a policy in place for developing emerging RIRs, the ITU however just doesn't meet the criteria.

These criteria were conveniently summarized for the ICANN Stockholm meeting

I doubt that the ITU could get any objective observer to think that they could meet these criteria, unless they radically changed their way of working of course, which isn't likely.

The reasons we SHOULD NOT have multiple RIRs per continent are various. The first is that RFC2050 (BCP 12) calls for a relatively small number of RIRs, and that they be of continental dimension. Another reason is that it would promote de-aggregation leading to greater bloat of routing tables. In addition, having the ITU set up a registry (or registries) to compete with the existing RIRs would lead to what is known as "RIR shopping". In other words, if a provider didn't like something about the policies set by an RIR community, they could go to an ITU IPv6 registry. This is exactly what Ramadass seems to want. This is generally seen as a very bad idea in the global IP addressing community, as RIR shopping is seen as a way to circumvent the Bottom Up, Transparent, Open, Consensus driven (my acronym for this is BUTOC) way that RIR communities make IP address policies.

The biggest applause of the whole exchange was saved for Dr. Nii Quaynor, the Dean of the African networking community, when he pointed out that he was very happy with the way the RIR system works, as it allows Africans to set their own policies in an open and transparent way, and this multistakeholder self determination is a positive step in development, and any change in this way of collective working is not in the interest of Africa, developing countries, or in the interest of a single Internet.

By McTim, Internet policy and governance consultant
Follow CircleID on
SHARE THIS POST

If you are pressed for time ...

... this is for you. More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Vinton Cerf, Co-designer of the TCP/IP Protocols & the Architecture of the Internet

Share your comments

Problem with link? Michael D. Palage  –  Nov 17, 2009 10:52 AM PDT

I have been trying to access that plenary session from the UN website for the past couple of hours with no luck, however, every other plenary session appears to be available. Anyone else having trouble with that link?

Geographic RIRs are a political sop Karl Auerbach  –  Nov 17, 2009 11:34 AM PDT

In order to meet the claimed goal of improving IP address and routing information aggregation IP address registries should be created in conformity with the overall clusters of internet connectivity.  In other words RIR boundaries should roughly match areas of rich connectivity with boundaries drawn between those areas.  That is not necessarily the same as geographic regions.

Geographic regions as a basis for IP address registries is merely a political sop to gain regional political support.

My last conversation with Jon Postel was on exactly this point - that IP address registries should conform to IP address aggregation and not political correctness and that such registries should come into existence or be merged as the lumps of connectivity of the net evolve.  We were in general agreement.

The ITU has a good record of handling telephony identifiers; there is no reason to believe that institutionally they would be any better or any worse than any other RIR except for their penchant to recognize countries and large, paying, corporations as having a voice in ITU decisions while leaving mere biological people and small companies on the outside.  The RIRs, particularly ARIN, have set a good standard of openness to which any new RIR ought to aspire.

There isn't any particular need for a new RIR, whether that be the ITU or not.  In fact there is some merit to the idea that we have too many RIRs already.

But if there were a need for a new RIR, the ITU, but only if it opens it doors wider and can avoid the kind of brutal country-going-rogue that we saw this week at the IGF meeting when presented with a negative message on a sign about a book, would not necessarily be objectionable.

It may be somewhat illuminating to remember that the IPv6 address space began life, back in the early 1990's, with one half of the entire space, a carved out to Netware/IPX networks.

If there was once a reason to give half of the IPv6 space to essentially one company then some descendent of that idea, even if withered and faint, might remain viable to justify some IPv6 space to the ITU to use as its sandbox for allocations, particularly if we recognize that gifts of RIR-ness to gain political support has been around for a decade.

Additional registries better serve economic and language needs, not routing aggregation. John Curran  –  Nov 17, 2009 7:00 PM PDT

Karl -

Like Jon, I agreed that the regional registry boundaries would ideally match regions across which aggregation could be used to the benefit of the routing table.  The challenge which such an approach is that individual ISP's wish to control their network routing across their backbones, and do not want to have a single aggregate which represents multiple network in a single region/country.

The reason we have regionalization in the registry system is to provide for more relevant policy making and support, as different regions of the globe have different circumstances in terms of language and economic conditions.  This approached has worked very well, with the RIR's working together to help with LACNIC and AfriNic formation when they were appropriate.  I'd recommend having discussions with folks in those regions about whether they feel better served as a result.

It is true that each additional Internet Registry causes multinational service providers to have to deal with an additional administrative overhead, and means that have to manage additional IP blocks and the resulting routing thereof.  In fact, an additional registry actually results in more routes over the same period for the same amount of customers, as large providers have to drawn their space in smaller blocks across more registries.  For this reason, the addition of any Internet registries needs to be very carefully considered, and presently these tradeoffs are reflected in the ICANN/NRO ICP-2 global policy on recognition of regional Internet registries.

/John

To post comments, please login or create an account.

Related

Topics

New TLDs

Sponsored byAfilias

IP Addressing

Sponsored byAvenue4 LLC

Domain Names

Sponsored byVerisign

Cybersecurity

Sponsored byVerisign

DNS Security

Sponsored byAfilias