internet governance, rir policy, and the decline of civilization
On (2014-09-20 13:32 +0900), Randy Bush wrote:
http://mailman.apnic.net/mailing-lists/sig-policy/archive/2014/09/msg00049.h...
Interesting quote from the paper. I've sometimes wondered if RIR's do too much, if there is inherent mission creep to justify increasing revenue due to increasing member base. While it is often useful to the community and at least mostly harmless, it may reduce competition in areas that are not strictly needed to be monopolized. I believe monopolies are good for many things, but the scope of each monopoly should be well defined and no mission creep should be allowed. I'm not sure for example, if 11MEUR is needed for number registry personnel costs, that could give you 100 hostmasters with 5500EUR/month salary, in good likelihood, we'd be able to run focused number registry with volunteers. Inside monopoly, there is always honest belief that you are operating as leanly as you can, because usually no organization realizes their inefficiencies until they must to survive. -- ++ytti
On Sat, Sep 20, 2014 at 3:56 AM, Saku Ytti <saku@ytti.fi> wrote:
I'm not sure for example, if 11MEUR is needed for number registry personnel costs, that could give you 100 hostmasters with 5500EUR/month salary, in good likelihood, we'd be able to run focused number registry with volunteers.
I think your math is off? 11,000,000 / 100 == 110,000 / 12 == 9,166 month right? Did you mean '200 hostmasters at 5500/month' ? you'd likely also have to put into the mix the cost of infrastructure, right? I'm not sure what current arin/ripe/apnic folk have deployed, I imagine some servers (100k of gear? replaced every 3yrs?) and routing/switching devices (2M replaced every 3 yrs), and link costs. -chris
I'm not sure for example, if 11GER is needed for number registry personnel costs, that could give you 100 hostmasters with 5500EUR/month salary, in good likelihood, we'd be able to run focused number registry with volunteers.
I think your math is off? 11,000,000 / 100 == 110,000 / 12 == 9,166 month
right? Did you mean '200 hostmasters at 5500/month' ?
you'd likely also have to put into the mix the cost of infrastructure, right? I'm not sure what current arin/ripe/apnic folk have deployed, I imagine some servers (100k of gear? replaced every 3yrs?) and routing/switching devices (2M replaced every 3 yrs), and link costs.
we could nit-pick saku's arithmetic to death, but what would we learn? it costs money and clue to run a good registry, news at eleven. in '92 or whenever, when the nic contract went out to bid, rick said he'd do it for free with some simple scripts. it's a long way from that to where we are today, and i doubt either extreme is where we should be. i suspect that if we threw out all the micro-management policies, restrictions on transfers, barriers to entry for legacy and newcomers, etc., we might be able to move significantly closer to rick's idealistic position. buy it would require a change of paradigm, and that usually requires a lot of folk retiring. so to repeat/paraphrase what i just said in the apnic forum, someone too shy to post here (yes, virginia, there are such people:) suggested i shill for them. i think their points are worth it. reasonable public resource governance practice would include at least the following: - term limits for board and committee positions (maybe 2-4 years?) - ten year employment caps on executive staff - members decide bylaws and budgets and as i suggested to arin, a gov/ops review consultation consisting of folk with some stature in these areas, and not having any members from board or staff. i would love to see some folk with enable on the board, such as you. but without the paradigm shift, it would just be pain and torture to no real avail. and good folk with enable are too busy enabling the internet as opposed to making careers as wannabe micro-policy wonks. randy
On Sat, Sep 20, 2014 at 9:32 PM, Randy Bush <randy@psg.com> wrote:
I'm not sure for example, if 11GER is needed for number registry personnel costs, that could give you 100 hostmasters with 5500EUR/month salary, in good likelihood, we'd be able to run focused number registry with volunteers.
I think your math is off? 11,000,000 / 100 == 110,000 / 12 == 9,166 month
right? Did you mean '200 hostmasters at 5500/month' ?
you'd likely also have to put into the mix the cost of infrastructure, right? I'm not sure what current arin/ripe/apnic folk have deployed, I imagine some servers (100k of gear? replaced every 3yrs?) and routing/switching devices (2M replaced every 3 yrs), and link costs.
we could nit-pick saku's arithmetic to death, but what would we learn? it costs money and clue to run a good registry, news at eleven.
sure, my point wasn't really that 'math is wrong', so much as 'running a registry likely costs some cake in gear/bw/admin-time' and that i'm not sure that 11m is off as a number close to the scale of the cost/problem.
i would love to see some folk with enable on the board, such as you.
can't other officer from same company already serving, phew! :) (see bullfighter turn stance)
but without the paradigm shift, it would just be pain and torture to no real avail. and good folk with enable are too busy enabling the internet as opposed to making careers as wannabe micro-policy wonks.
I also agree that 'lots of policy' hasn't really gotten us anywhere :(
I also agree that 'lots of policy' hasn't really gotten us anywhere :( < cheap shot > this is not exactly true. we just don't like where it has gotten us :) that's a fair cheap shot.
On Sep 20, 2014, at 9:32 PM, Randy Bush <randy@psg.com> wrote:
in '92 or whenever, when the nic contract went out to bid, rick said he'd do it for free with some simple scripts. it's a long way from that to where we are today, and i doubt either extreme is where we should be.
i suspect that if we threw out all the micro-management policies, restrictions on transfers, barriers to entry for legacy and newcomers, etc., we might be able to move significantly closer to rick's idealistic position.
It is true that we have accumulated several decades worth of policies and assumptions that originate from the baseline that "IPv4 is a very limited resource"; whether these make sense to carry over to a world where RIRs are not doing IPv4 issuance but merely keeping track of the present address holder is an excellent question. There is nothing that inherently prevents a change in approach other than the policy developed by the community. Given that anyone can participate in the policy development process (with results that are based on participant support for various proposals), does the lack of change reflect just a general lack of interest in making that happen or is it a reflection of the hysteresis built into the system? (i.e. the perceived need to learn the policy terminology and the policy development process, write up a proposal, educating others in the problem your trying to solve, participating in the discussion, etc.) I have had folks tell me that it doesn't appear worth the effort to change policy when all they really want is to get some address space - that a fairly hard situation to address; how does one have policy which truly serves the needs of entire community, when the actual participants are volunteers and thus a self-selected subset by definition? In the ARIN region, the presumed answer to this question is the ARIN Advisory Council, which shepherds the development of policy proposals into draft policies and proposes them for adoption before the community (which does reduce the amount of process issues that someone with a good idea needs to know in order to raise it for consideration...) It still ultimately comes down to the show of support for a given policy proposal at the ARIN Public Policy Consultations (PPCs) which are held during NANOG and the twice annual ARIN Public Policy Meetings. Those who participate (generally between 50 and 100 folks depending on meeting) determine via their show of support whether draft policies ultimately get abandoned or adopted. If there is an "unserved" segment of the community out there of any size, it simply takes attending either on-site (or remotely) to be included; making significant policy changes does not require membership, agreements, seats on the AC or Board, or anything else other than actual participation in the process.
buy it would require a change of paradigm, and that usually requires a lot of folk retiring.
I'm going to disagree since most of concerns you cite above (e.g. transfer restrictions) are set in policy, and as I indicated, that can be readily be changed if even a small number of people got involved with a clear intent to change them.
so to repeat/paraphrase what i just said in the apnic forum,
someone too shy to post here (yes, virginia, there are such people:) suggested i shill for them. i think their points are worth it. reasonable public resource governance practice would include at least the following: - term limits for board and committee positions (maybe 2-4 years?) - ten year employment caps on executive staff - members decide bylaws and budgets
and as i suggested to arin, a gov/ops review consultation consisting of folk with some stature in these areas, and not having any members from board or staff.
Despite the fact that I do not believe that current policy development is encumbered by the practices you cite above, I do believe that any member- based organization should periodically look at its accountability to the community served. There is some commonality of belief in that principle among the other RIR; in fact, the RIRs (working via the NRO) recently completed a RIR survey and published a matrix providing an overview of the governance frameworks of the RIRs. It is designed as a reference for the global Internet community, and provides a structured overview of various aspects of RIR governance, with links to the source documents on the respective websites of the RIRs - https://www.nro.net/about-the-nro/rir-governance-matrix I do not believe that the current RIR governance matrix gets to the level of detail necessary to show each RIR's compliance to your "reasonable public resource governance practices" listed above; I will propose that it be updated accordingly as a first step in this process. While everyone may not agree on what constitutes best practices for RIR governance, there is no reason not to have clear documentation of the current state of affairs to aid in the discussion. Thanks! /John John Curran President and CEO ARIN
On (2014-09-20 21:12 -0400), Christopher Morrow wrote:
I think your math is off? 11,000,000 / 100 == 110,000 / 12 == 9,166 month
right? Did you mean '200 hostmasters at 5500/month' ?
No. I live in socialist country where every 1 gross EUR paid to employees costs about 1.7 EUR to employer. -- ++ytti
participants (5)
-
Barry Shein -
Christopher Morrow -
John Curran -
Randy Bush -
Saku Ytti