[AFRINIC-rpd] Latest version of the policy AFPUB-2013-GEN-001-DRAFT-03

[Guy Antony Halse]

Hi

On Tue 2013-06-25 (11:09), Alan Barrett wrote:
> Thanks, that's a nice lead-in to my comments.

Which are all logical, and provide a really nice summary.  However, I think
you've missed an important factor.

(TL;DR academic institutions aren't ISPs, and are disadvantaged by this)

> 3. I think that the same policy should apply to everybody; I don't
>    like different rules for different types of organisations.

In an ideal world, yes.  However, that assumes equivalence between
organisations.  The reality is this:

a) Many academic institutions are, from an IP addressing perspective,
   comparable in size to medium- to large- Internet service providers.

b) However, unlike such ISPs, providing network and Internet access is not
   their core business; their core business is teaching, research, and
   community engagement.  Instead, networking (and ICTs) is a cost centre. 

c) Many are reliant of government or other external funding, which often
   comes with strings attached which prevent them from being used for
   anything other than directly supporting the academic endeavour.

d) As a consequence of b) and c), they often (in my experience at least)
   have IT departments that are both under-resourced and under-funded.

e) Some academic institutions are further reliant on equipment donations and
   other in-kind services, for which there is often little paperwork.

f) As a result of d) and e) it is difficult, and sometimes impossible, for
   such institutions to produce the documentation required for:

> 5. I think that a detailed network plan, including estimates of number
>    of concurrent users of wireless networks, should be sufficient
>    to demonstrate a need for the space.  There should be no need for a
>    formula involving the number of students.

   to sufficient detail to meet AfriNIC's requirements.  In this regard they
   are, in my opinion, at a substantial disadvantage to ISPs for whom
   providing this sort of service is core business or part of a profit centre.

g) The combination of d) and f) -- read the perceived difficulty of getting
   number resources -- mean that such institutions perpetuate (or worse,
   introduce) the use of NAT.  In my experience, this is actually harmful to
   core business (many research projects depend on end-to-end connectivity).

h) In most cases, network plans as referred to above really just indirectly
   measure the number of customers or end-users an organisation has.

i) Academic institutions have a well defined, usually public, and often
   audited means of counting their customers (students).  This equivalence
   doesn't often exist in the ISP world, and is an advantage that academic
   institutions therefore have.

The proposed policy allows academic institutions to overcome their
disadvantage in applying for number resources by utilising an existing,
well-defined metric (student numbers) in place of the traditional metric
(network design).  This lowers the barrier to entry for academic intuitions,
and makes it easier for them to do the right thing(tm).  This is, in my
opinion, a reason to support it.

For the record, I don't see the academic institution I currently work for
benefiting from the policy any time in the foreseeable future; we have
enough address space to meet our needs, at least within our current planning
horizon.  But I do know a number of universities elsewhere in Africa (FWIW,
mostly outside South Africa) that will potentially benefit from this.

- Guy
-- 
Manager: Systems, IT Division, Rhodes University, Grahamstown, South Africa
Email: G.Halse at ru.ac.za   Web: http://mombe.org/   IRC: rm-rf at irc.atrum.org
*** ANSI Standard Disclaimer ***                                    J.A.P.H