Just a couple of questions regarding this, though I do support what I see as the thought behind the policy.

First of all, how do you define a non-profit entity/network. In my opinion, you cannot give space under this policy to a "community" network unless it is a legal entity. Now, in the case of things like the Wireless User Groups and other community networks, how many of these are actually legally defined entities. As has already been raised, even with these legal entities, how many of the countries have ways to define "non-profit" organizations? If we get this definition right and define the criteria for qualification correctly, we will help eliminate the problem of people attempting to abuse this policy to get free space for commercial use.

I like the thought behind this because money in NGO's and non-profit organizations (and community based initiatives) is often extremely tight, and promoting IPv6 should be high on the list of things to do, I just think we need to very carefully hammer out how we define who this policy applies to, and how its controlled. As for, would I be happy to see a small commercial ISP pay and see a large non-profit not pay? Yes, I would, because size doesn't really come into this, what comes into it is what the space will be used for in the long term. IPv6 space used by non-profit, research and other such entities should be being used to uplift the African people, where as the small commercial entity is still using the space to generate income, thats the fundamental difference, not the size of the entity involved.
Just my thoughts

Andrew Alston
TENET - Chief Technology Officer



[email protected] wrote:
Hullo David,

Thanks for the perspective, my comments inline:

On 1/14/09, David Conrad <[email protected]> wrote:
On Jan 13, 2009, at 8:26 AM, Bill Woodcock wrote:
     On Tue, 13 Jan 2009, Graham Beneke wrote:
IPv6 Allocations to Non-Profit Networks
----------
Many community and non-profit networks exist on the African continent
and around the world...  Many of these organizations provide the
services free of charge and do not have any kind of revenue stream.
I strongly support this policy.
So, you'd be happy for (say) the Gates Foundation to not pay AfriNIC
while a small two person commercial ISP serving some rural area in the
bush would be subject to full freight?


I certainly wouldn't, and while this isn't the intent of the proposal,
it might certainly be the effect.

Also, in my (perhaps dated) experience, few countries had a useful
definition of "non-profit corporation".  This is why, for example,
APNIC is actually a for-profit company, at least according to the
Australian government.  If this remains the case, this policy would
appear to disadvantage ISPs in countries without a non-profit legal/
tax status.

I'm curious: what percentage of the total cost of doing business for
ISPs in Africa is the annual AfriNIC fees relative to the cost of
connectivity, staff, rent, electricity, etc.?


small, but this is for community networks, "ISPs" as we usually think of them.

This is the model under which the Internet was initially built...
No it isn't.  The model under which the Internet was build (at least
in the context of address registration) was government subsidies.
Before the RIRs existed, the US taxpayers (via the US government) paid
for _all_ address registration services.  That time is long past and
we probably don't want to try to revisit it.


Agreed

To be clear, this policy is asking one set of AfriNIC members (for-
profits) to subsidize another set ("non-profits", whatever that
means).  Long ago, APNIC looked at the same policy and decided against
it since (a) it was felt yearly APNIC fees were such a tiny proportion
of the total cost of providing Internet service in the AP region that
it was nonsensical to think that waiving those fees would have a
significant impact on the growth of the Internet in the region and (b)
the APNIC executive council had mechanisms by which fees could be
waived.  Whether or not these considerations are relevant for the
African region at this point in time is obviously something you all
will need to determine.

This seems to be a corner case, and while I am all for getting IPv6
into the hands of community networks and their ilk, I wonder if this
can't be done via tunnel brokers making /48 assignments
or by other means.  In general, I think it is not helpful to write many
corner cases into policy.

Like Badru, I am on the fence on this one, but currently leaning on
the "no" side.


_______________________________________________
rpd mailing list
[email protected]
https://lists.afrinic.net/mailman/listinfo.cgi/rpd

Reply via email to