[AfriNIC-rpd] Updated Version of the "IPv4 Soft Landing Policy"now Available Online

[Owen DeLong]

> 
> Owen, considering that this provision is ENTIRELY unenforceable, cannot be
> policed and creates the danger of limited global expansion, I fail to see
> what dangers the removal of this clause creates.  In the same way as you
> challenge me to provide a definition of an African company, which I will
> fully admit is not something I can do without much forethought, I challenge
> you to show me a way that we can enforce this provision and give it any
> meaning.  
> 
> A.) How exactly are you going to prove where the space is in use (Closest
> exchange point is meaningless, and I can give you concrete demonstration of
> why if you wish, latency is meaningless, and again, I can give you concrete
> examples of why if you wish, RDNS entries on infrastructure is useless,
> since you can set those to anything you wish, transit providers are
> meaningless in the world of long haul/long wire transport circuits), so,
> show me how you plan to actually enforce this.
> 
Show me any RIR policy which can be "enforced" or "policed". All of our
policies depend almost entirely on the cooperative compliance of the
community. If you want to use that as the bar for policy, you should
abandon the idea of policy altogether.

All RIR policies are merely statements of the intent and desire of the
community. All are subject to a certain amount of evasion and disregard.
For the most part, legitimate businesses attempt to follow the policies
and thus we achieve reasonably good compliance. In cases where
there are clear violations, the RIRs can and do revoke registrations
and issue new registrations for that space to other organizations.

> B.) Without RPKI in place (and that is something I will be fighting against
> with every breath in my body considering the hugely inherent dangers of RPKI
> implementation), how can we actually STOP a company announcing the space
> even if we DO come up with a way to prove they are using it off the African
> continent?  Sue them?  Does AfriNIC have that kinda resource?  Do we want
> AfriNIC to become a police force?
> 
You'll probably lose the RPKI fight and really, RPKI is no more dangerous
than anything else today. RPKI is only as dangerous as the ISPs that
run the routers choose to make it, just as any IRRDB, RIR Whois database,
or any other tool for authenticating address registrations.

> C.) As I pointed out, the clause restricts ISPs from using more than 10% of
> the space allocated UNDER THIS POLICY outside of Africa, pray tell what is
> stopping people from renumbering their African resources with the new space
> and then taking a previous allocation and using that overseas, in which case
> this clause is effectively NULL AND VOID.  I would also object VERY strongly
> to any policy modification that imposed such limitations retroactively on
> previously allocated space.
> 
As has been pointed out by others, the probability of African ISPs
making significant off-shore expansions in IPv4 is somewhere between
slim and none, so I tend to doubt that this is anything more than a
straw man argument.

> So, considering points A,B and C can you please either answer those or give
> me a demonstration of what danger scrapping this clause creates rather than
> some vague hypothetical statement.
> 
I would say the danger is no more hypothetical in my case than in yours.
I believe that with this clause, legitimate large foreign providers will
consider AfriNIC space off limits. Without the clause, they will consider that they
have permission from AfriNIC to plunder the resources as expressed in the
policies.

Owen