[members-discuss] [ncc-announce] New RIPE NCC Charging Scheme Models: Feedback Required

On 1 aug 2011, at 17:17, Nigel Titley wrote:

> 
> The RIPE NCC Executive Board will present a new Charging Scheme to the membership at the RIPE NCC General Meeting (GM) to be held in Vienna on 2 November 2011.
> 
> The Executive Board is working with the RIPE NCC Senior Management on various options that could be included in the Charging Scheme. It would like to solicit feedback from members on two Charging Scheme models before deciding on a final version to present at the GM.
> 
> The two models are based on established RIPE NCC organisational principles. Both models, as well as an explanation of the models' features and differences, are presented at:
> https://www.ripe.net/lir-services/ncc/gm/november-2011/ripe-ncc-proposed-new-charging-scheme-2012
> 
> The RIPE NCC Executive Board welcomes any feedback on the proposed models. You can provide your feedback by using the online form or by sending an email directly to agm at ripe.net.
> 
> After considering the feedback from the membership on the proposed Charging Scheme models, the Executive Board will decide which model it will present to the membership at the GM.
> 
> If you wish to discuss the proposed new Charging Scheme models with other RIPE NCC members, please start a discussion on the members-discuss at ripe.net mailing list. Information about this list can be found online at:
> https://www.ripe.net/lir-services/member-support/info/membership-mailing-lists
> 
> We look forward to hearing your comments.

(My apologies for a rather lengthy reply)

So I have a few comments and observations. 

First of all, I understand that the the current model is complex, and that there might be a desire to make it easier to predict.

Secondly, I welcome that the RIPE NCC is trying to break out cross-subsidies of the various services the RIPE NCC provides, and that each service should support it's own costs. Making the suers of these services become members of the RIPE NCC and allowing a membership category where they can have their say, but without becoming resources holders is a very good one that I support fully. However, in line with this I don't' understand why the cross-subsidisation of DNSMON therefor is reintroduced for the larger membership categories. I believe this should be taken out of the proposed billing models. 

This brings me to my first observation. I think most of us in the RIPE NCC membership thinks of the RIPE NCC membership fees as the cost of maintaining the resource registration data. In reality however, we through the RIPE NCC membership fees pays for (I believe) the majority of the RIPE meeting costs, the RIPE RIS services, etc. Services that we might or might not feel are important that that we perhaps might not miss until they are gone. RIPE NCC, and the entire RIR system, is about to change. The changes are huge and will impact their roles and responsibilities going forward, and a natural result of that is to evaluate the revenue models as well as the cost. When the RIPE NCC board and management team proposed the changes, I believe they didn't plan to ask for an evaluation of their path going forward, but none the less I believe that is what is needed. The reason for this is my observation above - that the fees pays for so much more than just resource registration data - and I believe that we are the RIPE NCC membership community now needs to make a conscious decision of what we are willing to pay for, how and realize that we might miss some services as a result of some possible billing models. The discussion on the list has focused on fairness and absolute numbers, but I believe we have to leave the absolute numbers for now, and first make up our minds on what we actually are trying to cover with these numbers. 

That said, I think the billing scheme needs to take a few principles into account. 

§1. It should reflect the use of the resources of the RIPE NCC. 

With unbundling of some of current cross-subsisides for other projects, this becomes easier. I believe and hope, that all NCC members support some of the cross-subsidisation that is occurring to for example the RIPE meetings and the RIPE regional meetings - so there will be some level of extra costs covered. Measuring the real use of the RIPE NCC resources is hard, but one of the factors of the current model actually does this quite well - the time a resources holder has had a resource. Partly because the longer a resource has been in use the less like it is to go through major changes, partly because it creates a potential financial barrier on fast-flux trading of resources. I believe this is good, and also creates a potential barrier against fragmentation of resources. Not so much because of worries of the routing table (if you have drunk the Cisco coolaid LISP will save us all anyway) but because if measures the use of RIPE NCC resources. Which is fair. Further, the size of an allocation also gives a fairly good measure of potential use of RIPE NCC resources in that the larger the block, the larger number of possible  assignments are that needs to be registered or otherwise handled by the RIPE NCC. 

This is to me fairness in the model. Not just the size of an allocation with a price per IP. That is not necessarily fair either as has been argued by some. 

§2. The RIPE NCC budget needs to be balanced 

The RIPE NCC charges needs to cover the costs, and leave a health reserve. Jochen said that the current proposed models are based on the forecasts of member and cost. I believe Jochen in this, but the problem is that it's hard to have a view of the absolute numbers without being able to compare cost per service, forecast membership categories and associated revenues. I think it would be good if this was presented as well. 

Also, as for my second observation, my crystal ball tells me that as we run out of IPv4 and we possibly start trading in IPv4 addresses the work load of the RIPE NCC might go up and might justify higher revenues and larger numbers of members. However, as the market saturates the workload will either stabilize, and/if when IPv6 gets a stronger market presence it will go down. Most LIRs that have their /32 will not come back for more for a very long time. This will lead to lower costs and again a different model is needed that will keep the RIPE NCCs core business - registration services - operational, and with a model that reflects cost. AGain, this cost probably will include some additional services, the questions are which ones?


§4. The RIPE NCC budget and services needs to be inclusive

While I strongly believe that some services that are currently cross-subsidized shouldn't be - I also strongly believe that there are those that should be. Primarily RIPE meetings, RIPE regional meetings and the policy development process. The reason for this is that the current legitimacy is less and less coming from the historical support of the community for the RIPE process and the RIPE NCC and more and more from the proof of a functioning system, running code if you so wish, and the continued trust and use of the processes and the RIPE NCC. The RIPE region and RIPE have more than any of the other RIR regions (if I may say so) proven that we can adopt to new needs, to changing political, financial and operational realities. This has seen a continued support and trust in the model. This is only possible by outreach and interaction with the very widespread (both geographical and cultural) community that RIPE and the RIPE NCC serve. It's extremely important that this is continued in order to keep the current model, which I believe serves all of us better than any of the other models proposed. 

§3. Each extra service needs to carry it's own costs

The RIPE NCC has been very good at producing new and innovative services, many of very good and important operational use. However, I also believe that the process for developing them and gauging community interest in these services could be improved - and as a strong believer in market economy I think that having them self financed is a very good model for getting community interest input. This will of course not be black and white, but I think the proposed model is a step in the right direction. 


In a way I guess I am arguing for keeping the current model - but I also see the drawbacks. But I believe that some of the fairness in time of allocations needs to be captured. I also question the doubling in cost for PI holders without further explanation. And as said above, I think there first are some more fundamental questions we as members needs to ask ourselves. 

Best regards,

- kurtis -

---
Kurt Erik Lindqvist, CEO
kurtis at netnod.se, Direct: +46-8-562 860 11, Switch: +46-8-562 860 00
Franzéngatan 5  | SE-112 51 Stockholm | Sweden