[ppml] RE: [arin-announce] NRO Response to ITU Comments on the Management of Internet Protocol (IP) Addresses

Gregory Massel gregm at datapro.co.za
Tue Nov 16 17:25:32 EST 2004

A number of African governments have the feeling that it is of strategic
importance for their countries to become involved in global technology
governance as a means of influencing policy such that it encourages
development of resources where historically these are limited and
discourages policy that facilitates advancement merely for existing
users/developers of those technologies.

The fundamental problem with the policy making structures of organisations
such as ARIN and the NRO, is that policy is effectively set by members, ie.
organisations or individuals who are making use of IP addresses right now.

To the best of my knowlege, there is little or no representation in ARIN's
or the NRO's policy making by individuals/organisations/governments/whatever
whose interest it is to consider those who do not currently use IP addresses
but who may begin to do so in the future.

The huge debate and lead time for the acceptance of ARIN policy 2002-3
clearly demonstrated this issue. The people who most stood to benefit from
it were NOT ARIN members. Yet the people who influenced the repeated
deferral of it and who finally voted to adopt it were mostly existing ARIN
members who were acting out of good faith rather then their own interests.

Even still, the ARIN policy making process is disproportionately represented
by Internet Service Providers who stood to gain significantly from a slight
barrier to IP address portability whereas if their customers were
represented, there would be a much greater push for a lesser barrier (and
possibly assignments even as small as /24).

The ISP approach to 2002-3 was typically along the lines of "what is the
effect on the routing table and can my current technology support it?" The
views of corporates and small businesses were not assessed, but my guess is
that they would probably have been along the lines of "Routing table?
Aggregation? Huh? Our ability to change ISPs effortlessly is paramount. The
technical experts can work out ways of improving router technology to cope
and ISPs can carry the cost as we're paying them for service."

Yes, it would have been possible for many non-ISP businesses to become ARIN
members and influence this policy by sheer numbers. But how many are even
aware of ARIN or know that they can become a member? Mostly they rely on
their ISP to handle their Internet-related needs. And so their interests
went largely unconsidered in the 2002-3 discussions.

What African governments are saying is that because many Africans are
technologically behind, it is important that they - as guardians of their
citizens (many of whom do not yet have the understanding or knowlege to
influence policy on technology) - to do it on their behalf and provide that

In my opinion, one of the biggest mistakes many Africans make with
technology (particuarly the Internet), is to confuse policy making with
policy implementation.

Looking at the ITU recommendation, one of the biggest faults I can find in
it is the notion that each African country needs to receive IPv6 address
space in order to set policy relating to IPv6. In fact, this would be a
costly and inefficient exercise and still fail to address how African
goverments can influence global IPv6 policies.

AfriNIC is going to face a massive challenge in the future. It has been
modelling itself on existing RIR's and in doing so has attracted the support
of the Internet community. It has also attracted the support of various
African governments (in particular .ZA which is a big sponsor). However, in
order to sustain itself in the future, it will need the continued support of
African governments and they will put pressure on it to set policy that
looks after the interests both of current IP address holders and persons who
may in the future require IP addresses (but who have historically not been
exposed to the Internet).

Anyone who doubts this need just look at the example of Namespace ZA, the
perfect case of an Internet community supported organisation that was forced
into becoming redundant through legislation because the ZA government
decided that it was too representative of the current Internet community and
not representative enough of the FUTURE Internet community (ie.the general

It is no co-incidence that the ZA government are such a big founding sponsor
of AfriNIC. I have no doubt that their enthusiasm for an African
organisation to set policy for the African region is because they feel it
will be more in touch with the needs of the technology have-nots and will
have a greater focus on education and outreach than an organisation like
ARIN whose membership is mostly a lot more technically knowlegeable and
therefore requires a greater focus on things like request turnaround time
than on education.

The key here is to persuade the ITU (and much of the rest of Africa), that
initiatives like AfriNIC are already in existance to address the sort of
concerns that they have and that they should use AfriNIC as a channel of
contributing to IPv6 policy rather than trying to create an alternative that
breaks a working global model.

I have no doubt that AfriNIC will be under significantly more pressure from
African governments than ARIN is from the US government or RIPE from the EU,
however, I do believe that it is up to the challenge and that familiarity
will assist it significantly.

I also believe that the AfriNIC policy formation model will need to change
significantly in the future if it is to address the concerns of
representivity that are much more difficult in Africa (because of the
digital chasm, never mind digital divide).

I urge anyone providing input to the ITU on their proposals to point out the
existance and relevence of AfriNIC and how an alternative country-specific
allocation approach will undermine AfriNIC, divide Africa, and in doing soo
weaken Africa's ability to stand united and in doing so have greater
influence on global IP addressing issues.

--Gregory Massel

More information about the ARIN-PPML mailing list