[AFRINIC-rpd] New Policy Proposal: Inter RIR IPv4 Address Transfers (AFPUB-2013-V4-001-DRAFT-01)

[Andrew Alston]

Hi All,

I have sat and thought about this reply for the last 2 days and well, I've
never been one to keep my mouth shut so here goes.


In looking at this there are actually two questions, one, why is v6
deployment so slow, and two, why is our v4 allocation rate as low as it is.
Why are we sitting on so much V4 space that isn't being used, to the point
where we are discussing (again) letting the rest of the world use it via
whatever means.

I believe the two questions are highly related, we have loads of v4, so
people believe they will be able to keep getting it, so why bother going v6.
Short sighted in the extreme I will admit, but it's something I hear over
and over again.  So, that leads to the second question, why are our v4 usage
rates as low as they are...

Here is where I am about to get a little controversial.

Firstly, our v4 allocation process is cumbersome, slow, and quite frankly
bizarre.  Unless something has changed, AfriNIC currently has two
hostmasters, out of 40 odd staff, meaning these two individuals are
completely overloaded and your turnaround time to get new IP addresses is
multiple weeks.  Secondly, the process to get those addresses is complex
enough that people avoid it at all costs.  I know this for a fact, because
I've had institutions (three so far), ask me to help process their
applications and do anything they can to avoid dealing with AfriNIC.  In two
cases, we managed to justify the space and get the space in both cases in
around a months turnaround time (RIPE iirc was working on a coupla days to
get space turned around).  In one of those two cases, it took 3 months to
get a genuine invoice issued that wasn't a proforma as was required by the
clients accounts department.

In the third case, an institutions which we concretely proved was larger
than the first institution, the total amount of space allocated was
significantly less than the first two, due to complete inconsistency on the
part of AfriNIC with regards to their requirements for IP space
justifications, and the arguments went on for more than 6 weeks.  

When we do finally get v4 addressing, we end up with billing issues like we
have seen last week.

Quite simply put, unless AfriNIC gets its house in order, simplifies the
process to get address space, is a HELL of a lot more clear about the
requirements to justify that space, and starts being consistent and
providing real service, people will continue to avoid dealing with them.  In
the words of one of my clients "I would rather NAT than go through that
process".  So yes, we have loads of IPv4 space, yes, the rest of the world
is desperate for space, yes, we have massive infrastructure deployment and
development on this continent which SHOULD be using space.  BUT people I
have spoken to and dealt with across the industry are afraid to deal with
AfriNIC because of inconsistent application of policy, bad service, billing
problems, weeks to respond to queries etc.

I strongly believe that our slow allocation rates are can be heavily
attributed to an organization that quite frankly, is simply not delivering a
high standard of service.  This is of course not the only factor, but it is
surely one major one.

Andrew

-----Original Message-----
From: Owen DeLong [mailto:owen at delong.com] 
Sent: Monday, January 14, 2013 10:11 AM
To: SM
Cc: Andrew Alston; rpd at afrinic.net
Subject: Re: [AFRINIC-rpd] New Policy Proposal: Inter RIR IPv4 Address
Transfers (AFPUB-2013-V4-001-DRAFT-01)

Admittedly, I don't live or have operations within the AfriNIC region, so
take my position or leave it as you see fit on that basis.

I oppose the proposal. I won't repeat what others have said, but I think Nii
made an eloquent example regarding exploitation.

However, I do want to repond to Andrew and SM below because I think this
raises a tangential but much more important issue. Why IPv6 deployment is
slow moving in Africa and what can we do about it?

On Jan 11, 2013, at 12:33 PM, SM <sm at resistor.net> wrote:

> Hi Andrew,
> At 11:23 11-01-2013, Andrew Alston wrote:
>> move to v6 unless they are thinking long term.  Thinking long term
though, doesn't seem to be in the nature of this industry for some bizarre
reason.  ISP's want to make sure
> 
> Yes.
> 

I don't think "want to make sure" is the issue at this point. Anyone who
isn't sure about IPv6 by now isn't paying attention.

I think the issue is more insidious. I'll be talking about it in Zambia.


>> Why are our v6 deployment rates so chronically low on the African
continent?  Because the
> 
> There are speeches and there is reality.  The IPv6 deployment rates is
what is actually happening.  IPv4 deployment rates is not that better.
> 

I'm not sure about that.

However, the bigger problem, IMHO, is that due to economic realities, Africa
is often faced with the dichotomy between buying current generation gear at
unaffordable prices or buying the worlds previous or even two generations
old gear at 1/100th or so of the original price, rendering it much more
affordable.

In most circumstances, this means you get less bandwidth within the
continent, a few less features, and a generally poor, but better than the
generation before user experience. That's usually an acceptable tradeoff
when you don't have an economic model that will support buying current
generation gear and neither do your competitors. Especially when you take
into account the abysmal state and ridiculous cost of getting bits to and
from Africa from other continents.

Unfortunately, as the internet moves to IPv6 (and make no mistake, the move
is accelerating and once IPv6 is ubiquitous in the developed world, the push
to turn off IPv4 will be strong), this means that having equipment 1, 2, or
even 3 generations old may mean you aren't able to communicate with the rest
of the internet. It doesn't matter how many IPv4 addresses you have if all
the sites you want to reach are IPv6-only.

As a result, current and future purchases of equipment which is incapable of
full IPv6 support (for whatever definition of full meets the needs of your
particular organization) is, IMHO, foolhardy at any price.

I hope as many of you as possible will attend my talk in Zambia and I look
forward to seeing you all there.

Owen