[AFRINIC-rpd] Academic IPv4 Allocation Policy Second Draft (AFPUB-2013-GEN-001-DRAFT-02)

[Owen DeLong]

On Feb 6, 2013, at 14:46 , Nii Narku Quaynor <quaynor at ghana.com> wrote:

> Hi Owen 
> 
> On Feb 5, 2013, at 22:49, Owen DeLong <owen at delong.com> wrote:
> 
>> Nii,
>> 
>> [..]
>> 
>> In my opinion, the question is not whether or how people design networks or document them.
>> 
>> The question is how much effort should AfriNIC staff be required to put into reviewing a request from a university? How much scrutiny do we need to apply to such requests? Are there rules of thumb that can be used to provide guidance to the staff which will facilitate a simplified application process with greater fairness and better outcomes for all with less labor required on both sides?
>> 
> Same uniform effort for university as to serve every other member. Perhaps requests should be smarter and the scrutiny more efficient but this seems an operational matter and not in realm of policy
> 

For general improvement, I would agree. OTOH...

> If you believe application process has been unfair send information to CEO attention 
> 

I don't know. I don't have first hand knowledge here. However, I do think, having been involved in the address planning and design of networks for several universities, that one can actually apply a rule-of-thumb about the size of IP block the university actually needs being pretty linearly related to the student count in most cases. IMHO, that ratio is very close to 5:1 in most of the developed world and that is why I have suggested it would be a good rule of thumb here as well.

If other applicants had such a rule-of-thumb available that would support simplified applications and greater objectivity while reducing processing time and effort, I would support applying those rules as well.

The bottom line is that we have here a proposed policy which can make life better for a significant subset of AfriNIC users without disadvantaging anyone. I see no downside.

>> I believe that the amount of effort being expended on both sides of this equation (which translates into costs as well) is excessive compared to the task at hand based on documentation provided by both AfriNIC staff and others throughout this thread and in other discussions I've had with people.
>> 
> I'll prefer to leave afrinic management to do this job more efficiently and universities to submit same information for service as any other member

Noted. I think it is fine if we agree to disagree on this point. We are both men of good faith and men of good faith will on occasion have differing opinions on the best course of action.

>> Further, I believe that Universities and other institutions of learning in the region represent the best hope for advancing various goals of development in the region. ICT investment and internet capabilities are huge levers for education, if applied properly. As such, I believe it is strongly in the best interests of the region to maximize the ease of implementing ICT in educational institutions and facilitating student use of ICT and internet capabilities.
>> 
> Agree but this is not the issue; evidence shows afrinic supports ICT in educational institutions

But that support can be improved while simultaneously reducing AfriNIC costs.

>> I believe that the discussion thus far has well established that a fully developed ICT infrastructure with students that are not financially or otherwise disadvantaged in their ability to possess devices can readily justify a rule of thumb around 5:1 IPv4 addresses per student or more. This includes not only the student devices, but also addresses for faculty, staff, university owned equipment, network infrastructure, etc. As such, I believe providing a policy which allows a University to apply to AfriNIC and use their student count with an agreed upon ratio provides a good rule of thumb which can allow a simplified justification process while still preventing abuse. AfriNIC expends less resources reviewing the request. The university spends less time going back and forth with AfriNIC staff over the justification. It's better for everyone.
>> 
> Afrinic would not be an impediment to a university request for a 10:1 network. University could be developing tools to make everyone's situation better instead of avoiding the work altogether 

The proposed policy would not make AfriNIC any more of an impediment than it is today. It would make AfriNIC less of an impediment for any ratio up to and including 5:1.

>> Are there universities that don't need such a large ratio just yet? Sure there are. Especially in the region. However, I would argue that it is far better if such universities are allowed to get the address space even if they can't use it today because it will support and potentially encourage them to develop better student ICT facilities which will improve the quality of their educational experience.
>> 
> 
> Are we encouraging hoarding ? 

I would not call it hoarding in this case. In fact, one could argue that what is hoarding is the AfriNIC free pool is being hoarded with overzealous attempts to make it last significantly longer than IPv4 may be relevant.

> Universities in Africa are not starved of address space. One recalls afrinic and AAU had a program with funding for universities to access address space for free and we also know the uptake of that arrangement 

I can only go based on what has been stated in the discussion by various representatives of Universities. You are the only one so far that has said it is not hard for universities to get space from AfriNIC. All of the other university representatives that have said anything on the subject have discussed various difficulties.

> 
> ...there is ipv6 that universities may use so they can chose what ratio they wish

You and I both know that host ratios are irrelevant in IPv6 and that only network counts matter.

Obviously I encourage universities to implement and train IPv6 as well.

>> I do not understand the argument in favor of preserving address poverty simply because other forms of poverty prevent taking full advantage of address wealth at the moment. Let us, instead, first resolve address poverty because it is easy to do so. Then we can continue to work on the other forms.
>> 
> I have not made that argument. 

Than I have misunderstood some of your earlier statements.

> Afrinic would not turn away member requests. I don't see address rich or poor as appropriate description 

That is not what other members have said earlier in this discussion.

> One only wants to administer resources going into real networks in Africa

Actually, at this point, because of the leverage provided by ICT training in universities, I believe that somewhat speculative address issuance in order to encourage ICT development at universities is actually warranted.

> A posture is that the universities should set example and do the requests clearly and cheaper rather than create special cases by policy

I can respect that position, but I do not agree with it.

Owen