Re: Address space for individuals
It would help the debate if, as early as possible, other last resort registries could indicate their views, their current practice and their funding methods. If recommendations are to be made, they should be based on as wide a poll as possible.
For the record, the UK situation. Within JANET, we undertake two registry roles: 1. We assign numbers for those connecting to JANET. This is part of the general service we provide to customers connecting to JANET, and is effectively funded through the JANET connection charge and annual rental. Therefore we wouldn't unbundle a charge for network number assignment in these cases. 2. We're the NIC of last resort for the UK. This service is effectively unfunded -- we do it as part of the JANET registry activities. Our funding authorities in general think that doing this sort of thing is desirable (ie in the interests of networking in the UK), but (if we were to ask) wouldn't pay for it directly -- they'd expect us to fund it by charging those wanting numbers. Obviously, we don't bother asking, but the operation of this service is a constant drain, with the real possibility that we'll have to give it up (ie hand it back to the NCC) if it gets more onerous that at present. For information, I'd reckon that processing a "simple" request, billing etc would cost around 250 ECU. A more complex one, where there's a need to discuss the situation in detail with the applicant might cost 500 ECU, maybe more. With these sort of figures it'd probably be worth insisting on having the registration fee come with the application, rather than billing afterwards -- this should cut costs a bit. You'd still have to generate a VAT receipt, though. Bob
On Mon, 23 May 94 14:11:24 BST bob@informatics.rutherford.ac.uk wrote:
For information, I'd reckon that processing a "simple" request, billing etc would cost around 250 ECU. A more complex one, where there's a need to discuss the situation in detail with the applicant might cost 500 ECU, maybe more. With these sort of figures it'd probably be worth insisting on having the registration fee come with the application, rather than billing afterwards -- this should cut costs a bit. You'd still have to generate a VAT receipt, though.
Here are some techniques we are using. You might find them useful to lower the workload of the last-resort IR's: 1. Giving advice on how to set up networks is added value for Internet Service Providers; I usually turn down those questions quite quickly 2. Try to minimise your work, eventually making more work for the requestor. Don't build an addressing plan to see if their application is valid; ask them to provide one. Don't accept if they are providing data their own way forcing you to transform; ask them to provide the data the way YOU want it (we do some calculations based on machine-[012] and subnet-[012] as a first estimation; jokers not providing this information see their application returned to them) Based on some ideas that came up during the local-ir WG and ideas we had earlier (which I forgot to bring up during the meeting - sorry), we probably want to change ripe-107 a little - it times out in a month anyway. Anybody else having hot ideas? 3. During last RIPE, I have heard several cry-for-help from some IR's who have customers that do not accept the rulings of the IR and cost quite a bit of effort convincing. I'm willing to help - these are 'difficult cases' and you pay my salary to do those, how much I hate them as well.... ;-) Personally I don't like the situation that we have to bill for IP numbers per se (it opens the way for big-$$$ company to apply for a B for their 300-host network), but billing for consultancy around the registration process seems much more suitable to me. Also, what happens if a company pays you 500 or even 1000 ECU, expecting you to give them a B, but their network does not justify assigning that much of address space? A case could be made for 'services not delivered' with nasty consequences as well... Geert Jan
Here are some techniques we are using. You might find them useful to lower the workload of the last-resort IR's:
As I've pointed out before, the real problem lies in expectations the people have with respect to address allocation. These expectations are formed -long- before they actually contact a local-IR. What we need are rules that are: - stable (do not change every 6 months). - are published (that does -not- mean an announcement of yet another Ripe document on a Ripe mailing list) and announced to a large audience. - have support in the whole Internet community (with other words: are the same in the US). I would not blame people that have applied for address space over the last three years for coming to the conclusion that the policy is essentially random. Simon
poole@eunet.ch writes:
As I've pointed out before, the real problem lies in expectations the people have with respect to address allocation. These expectations are formed -long- before they actually contact a local-IR.
What we need are rules that are:
- stable (do not change every 6 months).
I contest that the general rules are changing that quickly. We are refining bits as necessary but we do not change the general rules.
- are published (that does -not- mean an announcement of yet another Ripe document on a Ripe mailing list) and announced to a large audience.
What can we do more than publish them as RIPE documents? We do not have the resources for PR and frankly I think it is not going to change much at all. A significant amount of requests is still received by SRI although the InterNIC has done quite some PR and there have been two different organisations doing it in the interim.
- have support in the whole Internet community (with other words: are the same in the US)..
We are working on that. But if we want consensus before doing anything, forget it. I strongly belive we should not be wasteful just because someone else is. Think globally, act locally. Of course I am also very open for the needs of *European* ISPs. We shouldn't damage our industry either.
I would not blame people that have applied for address space over the last three years for coming to the conclusion that the policy is essentially random.
I have never had that complaint. Many people commented that it was getting more difficult to obtain address space. That is not random! Also everyone I talked to understood the reasons and did not object to the policy as such. Of course many objected to it being applied to themselves.... .
participants (4)
-
bob＠informatics.rutherford.ac.uk -
Daniel Karrenberg -
Geert Jan de Groot -
poole＠eunet.ch