Re: 1996 Charging Scheme

Daniel writes:
> 
> Local IRs are the entities who directly generate most of the NCC work. 
> The system of local IRs was consciously designed to take assignment work
> off the NCC in order to avoid the growth problems now so evident in the
> US.  On the other hand the NCC has to review big assignments and audit
> the local IRs in order to guarantee a fair process.  Of course the NCC
> also provides support for the local IRs.  This is a major activity of
> the NCC directly associated with the registries.  If you think the local
> IRs are a problem with the charging model, please amplify. 

I'll get back to this later.
 
> 
> NB: I expect that non-paying last-resort IRs will not exist anymore in
> 1996.
> 
>   > As Peter more or less pointed out, usage should be measured now and the
>   > results published. How otherwise is the commitee supposed to come to
>   > reasonable conclusions? 
> 
> We are not at that stage of detail in the discussion yet.  I also do not
> think it is practical to develop charging schemes from raw data on this
> list.  Rather I plan to propose concrete schemes with assotiated data
> *and implementation costs*.  So I think it would be premature to enter
> in a discussion about figures. 

We do need some idea on the granularity of the work done in the NCC. Just
for a example: are block allocations frequent enough and have enough
"value added" that we could charge on these allocations (I doubt it)?

> Nonetheless most of the data you request is -or will be again-
> available:
> 
>   > I would like to see at least:
>   > 
>   > 	- address block allocations per local IR and total
> 
> This is in the quarterlies and will be available again once we have 
> enough resources to resume them.  At this point I plan to resume publishing
> quarterlies at the end of the month.
> 
>   > 	- reassignments per local IR and total 
> 
> Will probably be in the quarterlies.

I consider this to one of the more "chargeable" services.

>   > 	- time spent working on requests per local IR and total
> 
> This is impossible to measure at the moment because the overheads are
> prohibitive.  We simply do not have time for it.  We barely cope with
> the load.  The cost of setting up a proper billing scheme with different
> levels of service was already much higher than expected.  Keeping
> meaningful records of time spent will be a significant overhead.

But this is probably the most expensive activity of the whole NCC, if
we -can't- charge this back in some way, we will end rewarding the 
local IRs that simply immediately punt to the NCC and punish those that
try to educate and consult their "customers".


> Using it for charging will probably be more expensive. 
> We are currently starting to implement a new request tracking system
> which will provide some data. This will take time.
> I will certainly look seriously into that possibility but it will not
> be before the middle of the year that we have useful data.
> 
>   > 	- Ripe database usage on a per domain/address base
> 
> What do you mean? Queries? Updates?

Talking to people this would seem to be a service that actually
has a value and for which people might be willing to pay for.

Simon