[AfriNIC-rpd] Definitions of LIR versus End User

[Owen DeLong]

On Jul 22, 2012, at 10:16 PM, Jackson Muthili wrote:

> On Sat, Jul 14, 2012 at 1:44 AM, Owen DeLong <owen at delong.com> wrote:
>> 
>> On Jul 13, 2012, at 2:26 PM, Seun Ojedeji wrote:
>> 
>> Hello,
>> 
>> On Fri, Jul 13, 2012 at 8:07 AM, Guy Antony Halse <G.halse at ru.ac.za> wrote:
>>> 
>>> Hi
>>> this creates exactly the sort of ambiguity I referred to previously.
>>> 
>>> There exists the potential for an organisation that neither uses its IP
>>> assignments *exclusively* for its internal use,
>> 
>> 
>> Such organization will then be referred to as an LIR
>> 
>>> nor *primarily* assigns
>>> address space to end-users.
>> 
>> 
>> I would think if an organization does not primarily assign IP to an end
>> user, then they be referred to as end-users.
>> 
>> 
>> Your previous two statements contradict each other.
>> 
>> A University, for example, primarily assigns addresses to its classrooms,
>> laboratories, and other IT infrastructure within the campus. It may also
>> assign addresses to customers such as students and/or faculty living in
>> on-campus or other institutional housing.
> 
> 
> A very interesting remark Owen.
> 

When I said his statements were self-contradictory, I was talking about
specifically this:

A. If an organization does not use its IP assignments exclusively for
	its internal use... "Such organization will...referred to as
	an LIR"

B.  "If an organization does not primarily assign IP to an end user,
	then they be referred to as end-users".

Either of those statements can be true, but not both.

Think of it this way... There are (at least) four possible types of
organizations...

1.	Like residential ISP, most of their address space is
	assigned to end users.
2.	Like backbone ISP, most of their address space is
	allocated to other ISPs.
3.	Like University, most of their space is used internally,
	but some of it is assigned to end-user customers such
	as residents in institutional housing.
4.	Like enterprise or other business, all of their space
	is used internally and they do not make external
	assignments or allocations.

In my opinion (which seems to agree with yours), 1, 2, and 3 are
LIR and 4 is an end-user.

This is in alignment with case A above.

However, in case B, only 1 would be considered an LIR. Cases
2, 3, and 4 would be end-users under that definition because
none of them "primarily assign IP to an end user".


> When looking at students as customers of an university, it does not
> matter anymore, a university is a LIR! All schools and colleges are
> LIR!
> 


Not necessarily...

I think that most universities are, in fact, LIRs.

However, there are two models...

One model where the university acts just like a residential ISP to the
students and/or faculty living in institutional housing. I believe this
is the most common model. It is certainly the most common model in the US,
Canada, most of Asia that I know of, most of Europe that I know of, etc.
Several people have commented that this is not the common case in Africa
and I admit I simply don't know the exact situation WRT Africa. (I suspect
neither do they as each cited only what happens in their individual
country as examples).

The other model involves the university assigning a single dynamic address
to each student end-device and not permitting the student to manage or
attach any router, switch, or other gear which is not an end-host to the
university network. In such cases, you can argue that the students
connection relationship to the university is more like employee/office
network than like subscriber/ISP and a legitimate argument exists that
the university is NOT an LIR in such a case.

Certainly, I think it is a far better scenario for the students if the
university acts like an LIR and allows the student flexibility to do whatever
they need with their residential network even if it is in on-campus
housing. However, I'm sure there are some university policy types who
would disagree with me and oppose such freedom of choice being placed
in the hands of students.

I also think that the current fee structure provides an incentive
to universities living close to the line on this issue to fudge
themselves into the end-user category if at all possible to avoid fees.

Rather than overhaul the entire fee structure, I think it would be
better for all concerned if AfriNIC were to develop a LIR fee structure
specific to universities and other not for profit entities which would
provided a deeply discounted LIR registration for them.

I hope that clarifies what I was intending to convey.

Owen