Wednesday-Amal-9AM
Wednesday, 29 October 2008 9:00
Address Policy WG
Main Hall � 9am.
The address policy session commenced on we had, 29th of October, 2008, as�follows:
CHAIR: Good morning, Address Policy Working Group. We will start in some five minutes or so, we are having some small technical difficulties that the webcast set up lost power and they are still reenabling everything, but this is the Address Policy Working Group and we will start soon.
Hello ops, do we have power to the projector? Good morning, everybody. We were having some technical problems with loss of power. This is why I am kind of late and the remote participation currently is not working and the network is not working, so you have to have� have to pay attention and not� now I lost the microphone. Welcome everybody to address policy, well, it's me and sander again,ed to is the day where we discuss the less formal things that need discussing, ideas that came up, well thoughts that we had and I want to get your feedback on.
Welcome here, everybody. This is the attendance� agendaed to. I have some issue with the letters of the alphabet. I have added one thing� there was the RIPE general meeting yesterday and, of course, 2007 01 played a big role in discussing the new charging scheme. Axel has agreed to give a short report what has been decided and what is the background of that. Then, I want to say a few words about, well, actually, timing for formal policy proposals. Then, we have a short presentation and some ideas about the wording, how the wording of policy texts in RIPE region could be improved. Then, we have sort of what I took from ARIN, the idea of the open policy hour, the only one that asked me to reserve time for him is Daniel, but something else comes up, feel free to bring it forward.
If we have time in the end, I'd like to spend a few minutes on brain storming regarding the policy process and whether we can increase participation, what your ideas are on that. If you have any other business, well, just bring it forward. Let's start right away, Axel.
Axel: Good morning. The few of you that are here. All right, general meeting yesterday, as you know, the general meeting in October sets the, supposed to set the charging scheme for the following year, and we had the interesting situation over the last two years that there was the policy 2007 01 which had obviously procedure and also corporate governance impact, speaking of fees. Basically what we have done over the last few months we thought about the charging scheme and how to integrate the PI holdings into that and what we did now, and what the members yesterday approved, is a soft start /slow start basically where we improve IP holdings into the usual charging scheme so there is no discrete fees per PI hold /SOG individual LIRs will barely feel the change here. However, made it very clear that is not what this Working Group originally wanted, it wanted a discrete fee per PI holding on the order of maybe 100 euros or so. We haven't ignored that wish, we have looked at a couple of options, different charging scheme there is and we found out looking at this one that a significant number of our members would be pretty hard hit with significant increases in fees from the RIPE NCC due to their PI assignments they did. And that goes up to a few tens of thousands euros. We thought that from a business operating standpoint, that is a bit of a risk. Now of course we could send out those invoices, some of our members would be rather surprised thinking that some of them might not have contracts with those assignies of PI space. We foresee or foresaw that that would create quite a lot of problems. We would send invoices, people would fall over in shock and say we can't get the money from our customers or ex customers so we can't pay that. Now we have somebody to invoice, we have to credit the in/SROEURBGS basically a big mess. What we foresee happening is we do the slow start and in let's say January, roughly, we will do another count of what we know of PI assignments and send out letters or e mails to our members who will be affected. We will tell them how much they are going to be affected, if nothing changes, and then they have quite a lot of time to get their books in order, their act in order and prepare for a possible change in the charges scheme at the end of next year, which then might, and depending on members' wishes, might include a discrete fee of, I don't know, 100 euros or something per individual PI assignment. Now, that will be a major change but our members then will be able to have prepared for that over the course of the year. So we think from a business point of view, that is a reasonable thing to do. It is not immediately doing what the Working Group wanted in terms of, well, charging for PI space directly and very, very openly, but this is something that our membership will be able to sustain. And that is basically my report. Guert, anything to add to that? You were there. Thank you.
CHAIR: Thank you. It's sort of gone away on me again but only the mobile one. Yes, thank you, Axel, for explaining the why and how and since this is also sort of falling in line with implementation plan on the registration services side that Andrew presented or said a few words about the day before yesterday, I think this is a reasonable thing to do. Personally, I wasn't fully in agreement with the NCC board before we had the discussion, but, well now I think this is a good way to move forward without too much disruption. In the end, there is what we wanted to achieve and it will not, well, cause too much trauma. So much for the report from the AGM. RIPE policy process, that line. Those of you that have been sort of following the meeting plan on the RIPE website might have noticed that originally that there was two slots planned for address policy and all of a sudden it happened to be four slots, and believe me it surprised me as well since this topic came up I had the feeling, oh we have some things to discuss but it's nothing we couldn't do in two time slots. In the four weeks before the meeting, then lots of things came up, and in the end we had about twice the amount of things on our agenda. So, the policy development process doesn't have a formal deadline like anything sent to Philliz or to us later than four weeks before RIPE meeting isn't going to be taken into account. So this is Moran informal request to those of you that work on new policy proposals. Since the weeks directly before RIPE meeting are very, very, very busy and especially the NCC stuff that supports us is very busy in the two weeks before a RIPE meeting, please ease our whole lives by sending in formal proposals sometime before the meeting and not on Friday before a RIPE meeting. That is basically the message I wanted to send out but I see someone standing on the microphone.
Daniel: If that is what you want to say, because I read cut off time, then the only thing you need to say the policy is first come first served. So if you are too late you don't get space on the agenda. It doesn't mean anything for the policy process� Daniel speaking for himself and nobody else� if you say as the chair the earlier you are the more sure you can of space in the agenda for the meeting, that is fair, right, we don't need anything new you just say that rather than only say please be early, and for the policy processes doesn't mean anything because the policy process doesn't happen at meetings, right, it's not� the meetings are not a formal part of the policy process. So anybody can do anything at any time but if you want space in the agenda at the meetings, the earlier you are the more sure you are that you will get space. Isn't that fair?
CHAIR: Well, that is technically correct. I think it's useful to have plenty of time in the meetings to speed up the process of hauling in feedback, and this meeting has shown that this really helps figuring out which way to proceed, so if you send in a policy proposal you wanted to� you want it to be on the agenda and so please send it on time. Randy and Philliz.
Randy: I think the question of policy goes first.
Philliz: Thank you. Queen it works now, right. OK. Just a clarification. We talked about this I think, Guert is also trying to help us here regarding the secretary, there is no intention to change the policy development process at all. Yes, in RIPE PDP, there is no binding towards the meetings and this is quite special (now) among regions because the RIPE region is the only one. There is no deline to submit proposals and this is into the suggestion so far as I understand to have a deadline, but rather, a recommendation because when formal� formal proposal we mean giving a number to it within the PDP and publishing on the website and making the necessary announcements on the mailing list and that takes quite a bit of work on the secretarial side as well because we need to prepare the pages, most of the time there is some recommendation, textual recommendations regarding the language of the proposals, even, from our side. That� we obviously do our best; however, if it can be not too close to the meeting then things cab bit better quality, that is what we are saying, I believe. Right.
CHAIR: That is just a point. It's easier on all of us if we have a bit more time to prepare.
Randy: 939th time
CHAIR: You miscounted you were at 42 yesterday.
Randy: No, I wasn't. The� my impression is, you know, I am APNIC policy chair so I have to read all the bloody policy mailing lists, the best thing about the RIPE one is it's not as bad as the ARIN one. But things seem to be more productive and rational face to face. And so I think this face time is important, and being plight about bringing up agenda items for it would seem to be a reasonable request. But I do want to note, and you alluded to it, I think, Philliz, is in the other regions especially APNIC and ARIN, the process of proposal being formal has all sorts of deadlines, etc., but you can discuss anything brought up at the last minute at the meeting. It's just not a proposal yet. I don't know if that is a worthwhile distinction to tell you the truth but I think being able to discuss things face to face is important.
CHAIR: I fully agree. Discussing it here is certainly� well I am certainly not aiming for sort of a hard deadline, if you don't send it in two weeks before the meeting we are not going to discuss it, so maybe the wording chosen here was not so ideal. I am not actually proposing to cut off anything; I just� I had the experience that too many last minute things mess up the agenda in big ways.
Randy: The IETF chair trick is you can put things on the agenda at any time, they go at the end, the later they come in and you start from the top and if we don't get to the end, well, you should have been earlier.
CHAIR: Yes. OK /WRUPB last comment and I think then.
AUDIENCE: Hey. The ARIN process currently has a deadline associated with it but it's not for any mechanical reason, it's not to give the staff time to do things or anything else like that; the intent is to allow it to have any given policy proposal to have a sufficient time or a time, a minimal amount of time of discussion on a mail list so those are not going to be at the meeting have an opportunity comment ton so that the chair of the meeting can use that as in the Advisory Council can use that as they are going through and looking at consensus building to get an idea of the broader community, what they think of it instead of the small sample of if you will, that is at the meeting. And I will give handy a pointer that when you are face to face there is a lot of different dynamics that occur on the mail list butty superb lien on a RIPE meeting where the total policy process is focused on list and not on meeting per�se, that is where all your discussions are made as far as discussion on the list goes, ample opportunity should be probably be your important criteria and not whether or not the staff needs to have extra time to get something done. As Randy says, if something comes in and needs to be discussed and you have time and room on the agenda to discuss it, it's used but because we put a little formal tag on it it doesn't get considered for consensus as it turned particular circle, wait until it goes through the next time.
CHAIR: OK, thank you. As I said, this is not something formal; it's just plight question to proposers to (polite) help us being organised better.
The next thing is about language used in the RIPE policy documents. During the process of discussing 2007 01 and also with some other proposals, people have remarked on the mailing list that the language of this document is so /HARable, this is no proper English (horrible) and well I am not ����
**
�����.Publishing on the website and making the necessary announcements on the mailing list and that takes quite a bit of work on the secretarial side as well because we need to prepare the pages, most of the time there is some recommendation, textual recommendations regarding the language of the proposals, even, from our side. That� we obviously do our best; however, if it can be not too close to the meeting then things cab bit better quality, that is what we are saying, I believe. Right.
CHAIR: That is just a point. It's easier on all of us if we have a bit more time to prepare.
Randy: 939th time
CHAIR: You miscounted you were at 42 yesterday.
Randy: No, I wasn't. The� my impression is, you know, I am APNIC policy chair so I have to read all the bloody policy mailing lists, the best thing about the RIPE one is it's not as bad as the ARIN one. But things seem to be more productive and rational face to face. And so I think this face time is important, and being plight about bringing up agenda items for it would seem to be a reasonable request. But I do want to note, and you alluded to it, I think, Phyllis, is in the other regions especially APNIC and ARIN, the process of proposal being formal has all sorts of deadlines, et�cetera, but you can discuss anything brought up at the last minute at the meeting. It's just not a proposal yet. I don't know if that is a worthwhile distinction to tell you the truth but I think being able to discuss things face to face is important.
CHAIR: I fully agree. Discussing it here is certainly� well I am certainly not aiming for sort of a hard deadline, if you don't send it in two weeks before the meeting we are not going to discuss it, so maybe the wording chosen here was not so ideal. I am not actually proposing to cut off anything; I just� I had the experience that too many last minute things mess up the agenda in big ways.
Randy: The IETF chair trick is you can put things on the agenda at any time, they go at the end, the later they come in and you start from the top and if we don't get to the end, well, you should have been earlier.
CHAIR: Yes. OK one last comment and I think then.
AUDIENCE SPEAKER: Hey. The ARIN process currently has a deadline associated with it but it's not for any mechanical reason, it's not to give the staff time to do things or anything else like that; the intent is to allow it to have any given policy proposal to have a sufficient time or a time, a minimal amount of time of discussion on a mail list so those are not going to be at the meeting have an opportunity comment ton so that the chair of the meeting can use that as in the advisory council can use that as they are going through and looking at consensus building to get an idea of the broader community, what they think of it instead of the small sample of if you will, that is at the meeting. And I will give handy a pointer that when you are face to face there is a lot of different dynamics that occur on the mail list butty superb lien on a RIPE meeting where the total policy process is focused on list and not on meeting per�se, that is where all your discussions are made as far as discussion on the list goes, ample opportunity should be probably be your important criteria and not whether or not the staff needs to have extra time to get something done. As Randy says, if something comes in and needs to be discussed and you have time and room on the agenda to discuss it, it's used but because we put a little formal tag on it it doesn't get considered for consensus as it turned particular /SEURBLG, wait until it goes through the next time.
CHAIR: OK, thank you. As I said, this is not something formal; it's just polite question to proposers to (polite) help us being organised better.
The next thing is about language used in the RIPE policy documents. During the process of discussing 2007 01 and also with some other proposals, people have remarked on the mailing list that the language of this document is so horrible, this is no proper English (horrible) and well I am not a native English speaker myself so I am not really qualified to judge that but the way these documents have come to be by taking a basic document and then replacing single parts of it with policy process patching the document over and over again pretty much guarantees that the end result will not be sort of a whom /OPBLG news document that is sort of nice and well rounded inside. So we have discussed this with Phyllis, Phyllis has volunteered to say a few words about it and then we might have an idea how to solve this.
Phyllis OK. It will take a couple of seconds because� well, it was uploaded but since ROSIE is down due to network and power problems, it's not kept on the laptop so we will resort to last century's way to transport data, USB sticks. Which is actually quite a nice way. I will just give Phyllis time instead of jumping to another item on the agenda because that would just confuse things even more. Phyllis, RIPE NCC.
Phyllis: Guert made an intro, he said what I was going to say in my slides, anyway, but yes, we have received some feedback recently, we have seen feedback on the mailing list and face to face and taking that this� into account, this is just question to you to have your feedback on and instead of me giving the answers, hopefully you are going to be giving the answers to my question. /# Phyllis: All right. It's not a break. It's working now, I am going to go on. So, yes, here you go the first slide, just� background about how we document policies: You know when the proposals are published, the wording of that text, proposed text is basically owned by the proposer, so we� we do make as a secretary /KWRART some suggestions for improvement, however at the end of the day wording is really the ultimate responsibility of the proprocessors and if they want to keep some specific language there it will stay in the proposal.
Once this proposal hits the RIPE PDP, obviously community comments are taken into account as wellate resulting RIPE policy documentation, assuming that the proposal, the specific proposal is� has reached consensus and concluded with consensus, the agreed wording will be used. So the� just to make clear, a little bit more clear: We do not document things after discussing some rough idea, but during the RIPE PDP there is a mechanism where the community overall decides on the specific wording, all right? So this sometimes causes some problems, let's say; I wouldn't like to say call it a problem, but in time, it distracts the flow of the documentation the reason for that is, I just got some, you know, looked at the main policy documents which do have most updates on them, mainly they are the IP version 4 and 6 policy documents and both were updated back in 2002 and before that they received updates too, but in 2002 they got this core language, let me put it that way, and since then they have been updated, IP version 4 has been updated six times, each time� that means six proposals has been, they have been accepted and concluded and they basically affected the document. Similar situation we see in the autonomous system number document which is RIPE 389. Since we have this mechanism of updating the policy documents bits where the proposal is affecting, the flow and structure are obviously sometimes interrupted. That is what we are talking about, because the up dates on the documents are only allowed and we only do those updates regarding the proposal bit, so one thing, say there is a proposal, it is discussed, it's wording has been integrated through a draft document, community reveals it, but it's only for that part, the up date is done. Maybe, you know, there could be a smooth transition if would you change the previous title as well, but since the proposal is not affecting that previous title, the policy document is not updated for that.
Now, so the question is, is it necessary now to have a clean up in the language of the main RIPE policy documents? The idea is� if you think that is not necessary from now on, you can discard the rest of the presentation but if you think it may be an idea, the suggestion would be done maybe only cover the regional RIPE policy documents, so this doesn't involve the global policies, because that is� the wording of them requires other organisations involvement and other RIR' involvement but maybe we can look at the RIPE policy documents and the surgery is totally cosmetic, the inclination is not to change any meaning in policies, there is no policy change at all, it is just the language. Like I said now, if you think that is a bad idea, you can just, you know, go on reading your mail if the network is back but if your answer is yes, it might be an idea, then the next question is: Should the RIPE NCC staff draft a revised version and publish the draft document for comments? You know, we have a drafting mechanism for poll seep documents, we mark all the changes if, there is a policy proposal and it is affecting the� one of the documents, we can use the same method so we can draft something and put on our website, you can comment and realise the changes, and if you would rather, if you would like us to do that, we would like to do that in phases, not all the documents at once but just, you know, by� document by document and of course we are pay a lot of attention to the concurrent proposals because I am sure there will be some proposals, ongoing proposals which are affecting the original document, anyway.
So, this is all about it and we are looking for answers, actually, and comments.
CHAIR: Thanks, Phyllis, for sort of putting it in a nicely formatted way up there. OK. If you run away, give her an applause and then I go back.
(Applause) Marco: Phyllis, are you intending to formally use the PDP to publish this revised documents or not, because currently, and people might have seen it� it was on also on the address policy, there is some discussion going on whether or not the actual changes should involve the policy development process or not, and what you are now really creating some situation where some changes like ASPLAIN� there was some discussion, well, the ASPLAIN AS.notation which is basically also a textual change is moved into PDP process and it will take roughly 12 weeks without any discussion and you are now saying, well, we are going to do some changes without PDP and I would like to get some clear line into this one and maybe as a suggestion, not only now we advise the documents but keep on working as further proposals come to really work the complete text and not only make minor changes every time somebody makes a policy proposal.
CHAIR: Let me comment on the latest suggestion, first. I think it's� it will make the process more difficult if we sort of, for every formal PDP change we adjust the language of the whole document, because then people cannot see with a quick glance what exactly is changed in the way of policy and which is just cosmetic change, so I am actually not very much in favour of doing complete changes every time because that will make it more difficult to track. Regarding your other comments, you are right, the ASPLAIN thing is sort of borderline; we might have not wanted to put it in, it went in, so there it is right now. We will try at the same time, of course, to avoid things, avoid putting notation in the text that will require these sort of changes, so this is not about technical details; it's really about the overall wording document structure and so on, so this is� even less impact than actually changing the notation for technical aspects. Personally, I think we should accept the NCC's help on this because they have very skilled technical writers, they have the mechanics to display old version, new version on the web quite nicely so you can look at it for�
AUDIENCE SPEAKER: To be clear, I really in favour of this and rewriting some stuff is pretty good. It's just about the process in use and the different processes in used to to get changes to documents, which is� should be� might be more, a bit more streamlined.
CHAIR: Yes, thanks for that. I don't think we should run this one through the PDP, actually, just put it up for review for the Working Group of course, the Phyllis and the NCC is not proposing to change any text without getting approval from us so maybe it will be like review period for couple of weeks and then everything that is not been debated goes in for the difficult ones we see how that works out.
AUDIENCE SPEAKER: The other one of course but that is more for the plenary would be to get more formal textual change procedure or procedure on� well, which doesn't touch policy as such and are only formatting changes.
CHAIR: Do I see Rob in here? Yes. Well Rob is sort of father of the PDP and I think he might want to comment on whether we want to� to permit textual changes have thank have no policy impact.
Rob: I think the PDP stands for policy development process and beautifying text, changing it into English English, correcting typos, that has nothing to do with policy, so the PDP should not be used for that. I think we are all grown up, sensible people and give it to somebody who has been hired by the RIPE NCC to write proper English, technical documents in proper English and publish the revised version not the policy but the wording and if people agree that it's not the policy change but better way to express the policy, then it's done.
AUDIENCE SPEAKER: I completely agree with you and I understand it's strange that we actually allow� I had some discussion with Wilfred, he isn't here, and I am really not sure if he is on line right now, but we had some discussion on whether or not it can be done and it happened in the past, but if formally we are sneaking in changes because the only formal way to change a document is during the PDP. I think /TPHAS that is the way I read it or understand T the PDP this is the policy, this is the procedure to change the document.
CHAIR: Well, it's the procedure to change the policy. It's not.
AUDIENCE SPEAKER: I think we should not spend too much time on this. I would propose that Phyllis and her technical staff take one of the documents, rewrite it and then we can see how that works. It will be published as a draft, not as a revision. And you can� then people can have a better understanding what it means to rewrite the document.
CHAIR: If we end up being in big disagreement on the changes later on, then we might really need to formalise this in a stronger way but I think we should do a light wide approach now and see how this works out.
AUDIENCE SPEAKER: Yes, I mean, yes, formally, we don't have procedure for this because it has never come up. And I think before even discussing whether we need the procedure let's first see what we are talking about, so Phyllis if you could take one of these documents, have it rewritten and send it to the relevant mailing list and then people can see what it means and then people can decide, oh, yes, this is so, so much has changed, I don't know whether it still is the same policy or not.
CHAIR: OK, Ray is next.
Ray: Extending what Rob just said what you want to be careful about doing, when you are in the editing process is that you don't actually change the substance of the policy, and so someone in a position such as the chair should be someone that says wait a minute, now this is materially changing the policy it's no longer correcting English or making it more gram atically correct or whatever and so you have to have some time of editor's control over the entire process from a quality merit standpoint. That is just an observation.
CHAIR: Curtis.
Curtis: I am going to� I think it's a good idea and I think I agree with Rob and Ray just said but there is different stories also at one point that I didn't really quite follow was the� when we do the re route of existing documents, future this doesn't exist, but when do you re route there is a problem how do you reference these after they have been rewritten. I had some trouble with Guert said it was the difference between the actual policy and the text, because even if the text means the same and it's rewritten in a better language it's still rewriting the policy route. You need to have some sort of referenceability and if I stumble across an as key file on the net how do I know if this is the old or new version. The text somehow reflects this. I don't think we should solve this now but there is impact on how you reference this that we need to keep in mind.
CHAIR: OK. I think we we have some disagreement on the fine print on the textual change approach, but we seem to be in agreement that we should see what this is going to� what the changes are going to look like and then see how the way ahead is going to be.
AUDIENCE SPEAKER: So, I agree that there should be no policy development of the rewriting the text. However, I would like to make sure that whenever you rewrite a document, you create a new document and not revise the same document with the same number.
CHAIR: It will get a new number and it will be in the RIPE document store, I think it will be very clear that this is obsoleting the old one. So numbers in the RIPE document store are never revised.
AUDIENCE SPEAKER: If I shall if it gets a new number and I am going over the revisions how can I see it's only textual changes and not policy change without reading the whole document? Will there be some form of demarcation there saying no changes were made to the contents of the policy just textual changes, so maybe introduce minor document number on the RIPE documents, calling it.0,.1,.2�
CHAIR: As we currently do not put in there which changes we have made to come from one revision to the other one, I am not sure whether this is so much relevant. Actually, if you want to know about the policy you go there and read the latest document. If you want to know about the process, how the document ended up being what it is, I think we have the mailing list which this is going to be very, very clearly announced and very open about and, well, there is no� currently there is no mechanism to tell why there is a new document, whether it was a policy change, wording change or whatever, so I am not sure we need one, but if you just want to look at the actual policy as it is right now, you will just find the latest document and that is it. Ray and Daniel and then we move ahead. Otherwise Daniel will not have a enough time.
Ray: The way we do this in the ARIN region is that we actually, as you know, we replace text in one manual because we have a organised by section and paragraphs, is that we have an company change log archive and cross reference system so anyone can do his research back and forth and if you look at the IETF index, it will talk about obsolete or whatever and perhaps consideration should be for the staff to create some sort of an index type document or a reference document which would do that cross referencing or so forth so that when a person wants to see what has happened to a policy� policy that has been replaced or changed or whatever could you go to that document instead of having to read through two or three different documents that all relate to the same topic.
CHAIR: Well, we have the RIPE index.text which says 389 is the AS number policy and it obsoletes 266 or whatever the old number one. So we sort of very similar to the RFC index document.
AUDIENCE SPEAKER: Right but that is a more generic type thing. You may want to go to more in depth discussion particularly when you are talking about a new document that only changes like a small portion it, didn't obsolete the preach one; it changed a portion. Example to use global v6 document that was used in 2002.
Daniel: Speaking only for himself. I think you have to be very careful that there is only one� one single document, that is the current policy.
CHAIR: Yes, I agree with that.
Daniel: That means that even when only text sort of� you are beautifying revisions are done, that this Working Group has to sign off on it and we have to be very conservative, if some people feel that there is a material change we either have to take it out or to run it through the formal process. But we will have to be very careful that we don't end up a situation where we say (end up) that is beautified version that have because then we open ourselves to this attack where somebody says there is an inconsistently and I will choose whatever is more favourable to me.
CHAIR: I think want to have it up for review�
AUDIENCE SPEAKER: I think what you are proposing is only sort of not to go through all the motions of the PDP but still have had discussed in the policy Working Group and establishing consensus that in fact no material changes happen and to say very clearly then the new version is the one that counts. If we overlooked anything, that might be interpreted later as a material change, we say the new one is the one, because otherwise we have this attack and also we had recently we had a procedural audit of the registration services department and I am quite sure if the auditors who had to audit against the policy the procedure against the policy, if they found something like that they would just laugh at us and charge double and treble because they would have to look for the inconsistencies.
CHAIR: It would have to be very clear that the Working Group has consensus on the proposed textual changes. Folks with more English skills than I have need to look into the fine nuances of the text, where they can be interpreted this way and that way, and then the Working Group needs to agree, otherwise this is not going to be� Phyllis, then Marco and then we go ahead.
Phyllis: I want to summarise a bit what I hear now, because it just comes to us I want to do what this group thinks the right thing is to do. What� obviously the reason we talked about this here is to get community's mandate on this practice. If you want to have the language improved, then we can do that, we are saying we can do that, if you want us to do it, we will do it but that doesn't mean we will just publish something and say this is RIPE policy document, obviously. The intention was to, like I said, publish a draft, open it for review and use the same drafting mechanism that is we used for proposals within the formal PDP. Although this is into the proposal, still mark the changes so people can see that, where the changes are. I must warn that you it may be the case, since it's a long document, most of it will be marked as changed, language wise but at least you will know where the stuff; and we can also publish� you can also do review for those who want to read the whole document, there will be a full version for the draft, not the� you know� fine, there is no final RIPE documentation.
Marco: If I� Daniel, if I understand you correctly, you are basically just published revised documents and moved it into lost immediately, that is more or less what you are saying, use the process but scape few steps.
Daniel: I wasn't solution earing, I was saying what the properties of the solution should be.
AUDIENCE SPEAKER: As a solution to what the properties does that fit your description.
Daniel: Personally I think we should just use PDP but I don't want to� I don't want to rain on this parade and, say OK if we can all agree at least as one� at least there is consensus and there is only one instance of the policy, then that is fine, but I think Rob made the most sensible suggestion here, let's try it and beautify one document and see whether we can live with that and then make the decision on how we want to treat T it's a much better.
CHAIR: That is basically wrapping up, I think things very nicely so we go ahead. Daniel is here anyway.
Daniel: So the open policy hour has now become open policy 25 minutes and I will try to do it in 25 minutes. I would like to use up a lot of credit with you folks and actually open something that I think we all re� reopen something that I think we all closed, so but I have serious concerns, that is why I am doing it. So you have been warned.
The disclaimer is a whole slide. I am not an expert at this stuff. I am no longer an IP analyst, IP resource analyst, I have done this stuff more than ten years ago but I by no means I know the details of what happens in registration services. I am really not an address policy expert any more either because I don't regularly read even only the RIPE policy mailing list, never mind the ARIN or APNIC ones. And lastly, I am not express ago formal official or any other opinion of the RIPE NCC as such. So, this was the disclaimer (opinion of). You can do me a favour here because we have this discussion in the RIPE NCC from time to time on whether, you knee, this is confusing or not so in order to test whether you are awake and to give me some feedback I would like to have a show of hands, and the question is: Do you feel, everybody in the room, do you feel that you understand when I stand up here and I am employed by the RIPE NCC, you know, I have been among the RIPE's founders and I am actually the oldest employee on payroll of RIPE NCC, when I say I am not expressing anything as the RIPE NCC, we haven't agreed, it's not an official position, do you feel you understand the difference. A show of hands please, who feels they can understand that concept. Who feels that I don't? OK. All those who were awake raise their hands. So the� the second question, indulge me one more question, since you seem to understand it; if this is happening, and suppose this might happen more in the future, especially in this Address Policy Working Group, so suppose that some people in registration services or some people in policy development like Phyllis would stand up at some point and say, there is an item of discussion here, we haven't had time in the RIPE NCC to actually form an opinion on this but my personal opinion, as you know, me, Daniel, me Phyllis, me I think grid, me ago that, whoever, I have this thought; what do you think that makes the RIPE NCC stronger, would you feel better about the RIPE NCC or would you feel worse about the RIPE NCC because there might be some contradiction later? And I I don't mean the hypothetical case that Axel stands up being the RIPE NCC and I think the world is is a sphere and then somebody from the� from the RIPE NCC getting up and saying "I think it's flat." I don't mean it's flat. I just mean constructive stuff in discussions. Would you feel that the NCC was stronger or less strong? Can we have the poll. Who would think� would make the NCC� give you a Bert feeling about the NCC. (Show of hands) and who would think you it would give you a worse feeling because they are inconsistent? Nobody thought worse but not very many people thought Bert.
Leslie: I want� I didn't vote either way because I think that it's not possible to comment CRISPly. I think that part of the challenge in your hypothetical case at some point a difference in number may become a difference in kind. The situation� I personally value the people who are in the thick of this stuff as their day job having the ability to stand up in this room and share their perspective, I think that does indeed make things stronger but what you want to generally avoid is is a situation where it starts to be the case people feel fur not in that context your opinion is somehow not as informed or that there are things that are going to go on between meetings because there is this unknown RIPE NCC space that is going off and doing their thing. It's not true, but this is the perception that you want to avoid and that is the difference in kind you can fall into, depending on the difference in number.
Daniel: That is clear.
Rude Kerr: I think there is a space for formal discussion and there is a space that is very important for informal discussion. At least for informal discussion, it seems to be clear to me that censorship always exposes weakness of an organisation.
SPEAKER: So in this sense this is informal discussion. So actually the talk is not about that, I wanted to check everybody� who is reading mail and all that stuff. The reason for this talk I looked at some IPv4 allocation data over the last two years and it caused me personally concern and some people shared this and the RIPE NCC hasn't formed an opinion about this but I wanted to share it with you. So I want to share this concern with you and at the tend, I will ask questions of whether, in principle, we should take policy action.
So, if we look, and this is about the run out, when we really run out of IPv4 addresses. So I think in our minds, if we think about the IPv4 request queue, we have a picture like this in mind, and the individual cubes represent requests and the size requests the A address space that is being requested and I think when we talk about these kind of policies, our mental image is something like this. Now, when we run out, at some point, there will be, you first come first served, some requests that will be fulfilled and some requests that will no longer be fulfilled because the pools run out. And I am not going into the details on whether, like you know, after the red line, some of the smaller ones they might be still some little bits left to satisfy the smaller ones. I don't want to go into that detail but I want to say what we imagine and what I always imagine is something like this, without even thinking about it. So at some point we run out, it's first come first served, the people to the left get something, the people to the right of the deadline get nothing. So what my concern is about, that it might actually look like this. And this is not drawn to scale; this is way off scale because if I drew it to scale, the small ones would be dots, you know. So what happens is like you know, we have the same kind of pattern that we had here to the left and then there is a real big request and there is nothing wrong with this big request it, fits our policy and it's duly evaluated like all of them by the IP resource analysts and the policy says you have to fulfil this request, it came in at this point in the queue and it was fulfilled. But what that means is that the stuff to the right side of the queue that we actually thought would look like this becomes to look like this and that is not run to scale because there is a much longer tail here so it's just� it's not a real representation. This is worse than it's been represented here. So our deadline that we thought sort of was here, is actually moved here and all these people who were maybe because of previous patterns expecting still to get something, don't get something, and actually these things are about four thousand times the smallest things here. So there is four thousand there or between, like, a thousand and four thousand there.
That is very unpredictable, and when we had the discussion before about the soft landing stuff and so on, we all said well but it's fair so it can be unpredictable. It's in the matter of nature. But it's going to be perceived as very unfair, because what happens is that we, in fact, is that we here fulfil this request, which according to the policy can have planning horizon of, Phyllis, what is it now, a year, two years� Phyllis: 12 months
SPEAKER: It can have a planning horizon of 12 months while we were fully aware that it would be the last one we have to give out, so the perception, and I am not talking about the real unfairness, I am talking about the perception, when this happens and it will attract attention, obviously, we will have to say we run out now, there will abperception of unfairness and there will be� and if it's� you know, if it's� if it is raised in the consciousness of various groups, be that the general public or the Internet governance crowd or the regulator crowd, it will be perceived as sort of optimum. It feels unfair. Well, actually we think rationally it's fair, it will feel unfair and it will discredit our policy process because I think the public sentiment will be, how stupid were they making this kind of thing, this unfairness possible? And I am only talking about the perception; I am not talking about sort of anything real here. What also might happen is something like this and again that is conceptual picture, don't get me on the details, but what what this is supposed to represent there were two or more of these big requests right next to the deadline and one of them gets fulfilled and the other doesn't. And there is stuff here in between and before and after but I don't want to go there. I want to show the concept. I think that would be perceived as even more unfair because here are people with you know same size of legitimate requests, according to the policy, and this one gets another year's worth of stuff and the dotted one gets nothing. If both requests come from the same market, say they are big ISPs in the same market, I think the people who are actually having to explain that to their management will have a hard time and if we then go back to them and say yeah well you were members of the RIPE NCC, you were members of the right community, you could have paid attention when we made the policies, it's not going to cut it very far, and it might actually be that the one didn't get anything runs to the regulator and it might not not be telecoms regulator it might be economics and say there is a cartel here that prevents me from doing business. Things like that. They could try to use public opinion and it would really reflect badly on us.
Now, suppose these two come from antagonistic societies. We all know in our region there are some countries, societies are not really friendly with others and groups of countries and societies and things like that and I am not going to fall into the trap of naming names, but you know, think a little while and you will think of some. Then, it can be even more embarrassing to us, and as my first history teacher always used to tell us when we were flabbergasted about the causes, not the real causes but the little causes of actually armed conflict, he used to say "oh, wars have started over less" and tell another story. So I am not entirely joking here. But I am not worried about the war� I am worried about the war but I am more worried about the fact this really would discredit the policy process because they would say how could they make a policy that causes all this turmoil, and it would discredit the RIPE NCC as an institution. And actually, sort of as a side effect, this perceived unfairness may cause Internet breakage because whenever this is raised beyond a certain awareness level and positions get entrenched, you know the one who didn't get anything might decide to stat on the stuff that the other guy did get, or even worse, the one society might say we don't recognise this RIPE NCC and this kind of address allocation process, we will make our own. It's not entirely hypothetical, I think, because when this gets heated, right, unpredictable stuff happens and that is my concern.
So let's look at some hard data that prompted this, and bear with me for a couple of minutes while I go through it. Somebody else did all the regular work here so we actually used our tools to get me these graphs; I didn't do them myself. This is on the X axis you have the years from 1990 to 2010. And on Y you have the number of allocations and I didn't like at assignments but assignments are quite similar and a little bit smaller so I only look at allocations, these are the allocations we gave out so this line here is 1,200 and what you see is that from 20 O2 on the absolute number has been increasing, the last one here is not a real decrease because that is 2008 and we are not finished with that year yet so I guess it will be somewhere like here but that is just a guess, so it's not like we are declining there. And we are at 1,200, 1,300 in total. (Route). Now if we look at the size distribution for last year, it's something that is not really surprising, so here on the X axis you have the size of the allocations and here you have a/21, the smallest, and you have a/10, the biggest that we gave out this year, and you know, it's all part of two arithmetic so there is a difference of two to the 11s so a difference of 2000, so the thing here is 2000 times� suppose it's one, I am not really sure but I think it's one, the one thing here is 2000 times as big as the one here. Now, if we look at this, it were less than� there were about 1,300 assignments in general so we know, already, that this one here took more than all the rest, at least took more than all the small ones. If we look at� and the distribution is not really all that surprising because we have a lot more smaller members than we have big members, and these kind of distributions where a lot of the stuff is on one side, one like, and then it trails off to one end, is quite common in these kind of statistics so it's not really shocking. Everybody follow so far? Any questions? Thank you.
So if we only look at the left hand side of this, so here we look at only the/16s or bigger, and remember, we are getting bigger as we go to the left, it's not really surprising, either. It shows the short of same curve but there is some pretty big ones here, yeah? Let's not forget it. So they are all in that queue so it's not� the queue is not like the first one I showed but more like the second one I showed. And this is actually consist sent over the years. If you look at 2006 we actually gave out the/9 so that is 4,000 times the initial allocation size and it's 8,000 times the initial allocation size in the� in some other regions, so and there is nothing wrong with that again, there is nothing wrong with that, according to our policy the RIPE NCC IP resource analysts have done their job, we have actually done additional checks and balances for the big ones, they are all according to the policy it's all fine, there is nothing wrong with them, but they are here we are doing them. We go to 2005, about the same picture, 2004, similar picture; 2003, similar picture; 2002; 2001. But if I run this movie backwards you will see a trend. Let me do this. Watch the trend. 2004, 2005, 2006, 2007. So they are getting bigger, right? It's quite obvious, right. Let me run again, 2001, 2002, 2003, 2004, 2005, 2006, 2007. So what I am trying to say with all these nice charts is basically, it exists, you know, the I think that concerns me is actually you know, is actually there. So, repeat, you know, this is not about� this is about the perceived unfairness, you know, the up roar that might happen. It's not about the real unfairness because we had that discussion and we said FC F S is fair enough, is really fair; it's not about allocation efficiency to say let's do something more efficient. It's not about making the pool last longer and it's not about dealing with all the previous con� contradicting all the previous landing proposals and the discussions that were going on there.
Curtis: So while I probably agree with you, the graphs would make more sense and prove your point better if you told us where the allocations were going. Are they being assigned to people who already had a lot of address space, are they being assigned to new entrants, are large blocks going to� I am guessing that you are right, I am not necessarily questioning it but I am just saying there is some more data would have backed your point up a bit better
SPEAKER: I am quite happy to provide more data. I personally think that is irrelevant because I assume, according to our current policy.
Curtis: I also aassumed according to policy but when you are talking about fairness I assume fairness in some sense in the perceived sense equal distribution, equal access to resources and in that case, what you are trying to argue all the large allocations were given to people who had massive amount of address space.
SPEAKER: No into no I am not saying that. I am saying that this situation� no matter whether it's a new entrant or it's an old one, this situation is perceived to be unfair. One gets a year's worth of a lot preempting the other guy or that one, but can I finish or is any question, Randy, or can I finish first. I have four more slides, we will have the questions at the end that I have.
Randy: Clearly you can finish. ARIN actually provide that had analysis for the last end years 75 percent of the address space went to ten holders. And we, as a community of people established in the market, have placed a barrier to entry with the excuse of preserving the routing table by you can only enter if you are big enough to save our router resources.
SPEAKER: Guys we have five minutes. I would like to finish my presentation and ask the questions I want to ask. So I repeat, this is just about the perceived unfairness at the end and how it reflects on us. There are lots of responsibilities to deal with this. I don't want to go into them, right? I don't want to go into solution earing before we know that we really want to get this done. The only reason for the slide is to say, there are ways to do it; some may be better, some not so good. I want to discuss whether we should do something because I think some of us at least feel that we have been there, we have decided not make changes and I think there is new data here and there is new concerns that weren't raised then.
So the pros for doing something about it is that, to my mind, this avoids some really bad outcomes. For our credibility and the credibility of the RIPE NCC and for actually the maintenance of ourself governance process. You can agree with that or you can't, but that is my� to my mind, that is pro, and that is really the cause here. There is some additional pros, it improves the predictability of the end game which is always goods in these type of things, often surprise in a crunch and might provide an additional signal of urgency if we say we will have to� you will get less, you will have to� your planning horizons will have to go down at the end, it actually might provide� give additional push, in which direction, you would hope towards IPv6 but that is another discussion. The cons and that is mainly the feedback I got when I talked to people, is like you know Daniel don't go there, we have been there, is to� it could be interpreted as deviating from the needs based principle and we have said OK we will do it needs based because that is the only defendable way. I would argue that is not quite true because it depends on what the horizon of the needs is. If in the end it's just a game of vabonk where somebody gets a year's worth and other nothing, you could also argue if you say towards the end we can only consider immediate needs or more immediate needs, which still be needs based. And then of course, what everybody said immediately is you will never get consensus about doing anything about this. So those are the cons.
The issues to be addressed, so these are some things we have to keep in mind, when and if we want to address this, is the codeification or language of the proposal itself needs to be practical so we cannot impose or tests or something like that on the host� or the IP resource analysts that, they cannot perform or would cause the NCC to need an infinite amount of resources and I see our chairman of the board nodding emphatically and of course we need do it rather quickly which seems to be somewhat contradictory with what many people's expectations were about the ability to reach consensus.
So, my question really for the discussion, and please read the bottom first; I don't want to have a discussion about policy details, you know tweak this, that and all that kind of stuff, given that we are already out of time, it's not a good idea and it's not a good idea in general. My questions for you that I like to focus the discussion on those are, are the scenarios that I am presenting realistic? Will they cause, or will they like competitors to fight in courts, by regulators, by government, things like that? Which has a fallout on us. And with us, I mean both our policy process, us here, but also the RIPE NCC, I am a little bit attached to it. And another question is, are governments concerned about the scenarios and the perceived unfairness? Do we think it will impact on them? And then obviously, the penultimate question: Do we need to develop policies to address this? I am done. Thank you for indulging me. Ray was a tenth of a second faster.
Ray: ARIN. The answer to the first three questions are three. The answer to the fourth question is up to the community to decide.
SPEAKER: Thank you, Ray.
Randy: This is an example of, you know, the policy for the last/8, especially APNIC's version slighting it into many pieces, you are just moving this forward I agree with it. Are there scenarios realistic, I am not going to do planning on fear based and attack me thing in courts. It's fair, is the point. Keeping /TK B making our marketate ability for people to enter� get some of what is a sincerely diminishing resource, so the network can keep running. So, I think doing something to see that we� the last� we have 39/8s left� 75 percent of that does not go to ten entities.
SPEAKER: So you are saying, basically I am arguing for the last/8, give everyone a small piece? Randy: No, I am saying that is the end example, of the last/8 and I am saying you are discussing it moving it forward into the present tense and I agree with that. I am not sayinged to you should chop it that fine, right? But maybe this kind of�
SPEAKER: You are basically saying I am saying the same thing as last/8 but maybe smear it out more Randy correct but maybe
SPEAKER: The last/8 doesn't address this case Randy: I understand.
SPEAKER: Thank you.
Sander: I want to agree with your research. I looked at some numbers myself and it seems that, actually, more than half of all the address space that RIPE NCC distributes is actually going to a few big blocks, so I made some (distributes) pie charts of that. I think they are on the desktop.
SPEAKER: Let's see whether they are up here. So sander, this is quite good� wrong one.
Sander: It's on the desktop.
SPEAKER: Actually what sander did, he looked at my presentation, I uploaded yesterday and said, well, really what we should do is look at not at the number of� can I talk to this four� sander: Sure.
SPEAKER: He walked over to me in the general meeting and said I did this not in the number of of allocations sense but in the sense of how much address space is it. And I like that very much and it's often public data so what he did is he made a pie chart not how many allocations are there but what is the number of addresses in each, and I will just show them. They are a matter of record, you can download them from ROSIE. In 2002 it looked like that and 50 percent boundary was at a/14 in 2003 it already moved into the/13 business N 2004� 2005, suddenly/12 and bigger is half of it. 2006 is similar. And 2007 is again similar. So this means like, this doesn't even have anything smaller than a/16 but what it shows you is that in 2007, actually more than half of the address space went to the allocations that are/12 or larger. So that is just an additional piece of data that I too lazy, frankly, to do, and thank you, sander, for doing it. And here is 2008 which seems to be an opposite trend but we are not finished with that year yet. So this is is� on ROSIE so if you are interested, you can download it.
Curtis: So I am actually not so surprised by the data, I was surprised if the data showed something else because we are also seeing a lot more people coming into the market. The data here is not necessarily all that surprising because you are leaving out lots of stale stuff that might make it interesting. I think the interesting thing to discuss who is the address space going to and how many people getting access to. That is much more interesting than size and how many allocations you are doing, that doesn't say anything. I agree with everything you said, I think the four points yes, I think people are worried yes, we are running out of resource, yes, I think the government will get interested because we are talking about infinite resource that have affect on allocations and how economies can grow so yes to all the questions. Yes, I do think we should adoctors policy we are out of time to discuss how. You didn't want to talk about it which is fair. I do worry of the fact we are running towards, we still have only two RIPE meetings a year so we can't respond to policy needs as fast as other regions can which might be a nice observation to make. So I do think we have to address this, we are out of time to discuss how.
SPEAKER: OK. So, I don't think with� I don't agree with one thing you said. We can actually make policies between RIPE meetings. The PDP doesn't say that we absolutely need a RIPE meeting. Curtis: I agree but there is nothing to generate progress as much as a meeting.
SPEAKER: I agree with you. My question is: When I pose this to some people in this room and outside this room they said "Daniel don't go there, you will get no consensus about it and you will just waste a couple of days of your life trying to actually write this policy." So all I heard so far is, like you know it would be useful. Then, I am quite encouraged, actually.
Randy: If you want I will tell you it's absolutely horrible. Now, are you happy?
SPEAKER: No Randy: I think it's a good idea. We have let the incumbents dominate entry to the market. Not fair. There is one of my closest in/WUPL bent buddy and of course I am, too.
Rude Kerr: Personal opinion. Well OK, I think it would be stupid not to consider things here. One of the potential threats, Daniel, that I missed to see in your list is, well, OK, if things go into� well OK, in the end we will have� we will have pain and bad public relations, anyway
SPEAKER: It's about this�
AUDIENCE SPEAKER: What I am seeing is that if people, at that time, ask a question, well, OK, even assuming that the policies have been set up fair and so on, but are there any compelling reasons why, knowing we are running into the wall, nothing has been adapted to smooth it? And if there are no compelling reasons, we will be looking really bad. On the other hand, of course, hastily creating a bad policy is going to do even more damage.
SPEAKER: So your first remark I interpret as support, basically say let's think about it. Yes. And let's try to do it. And the second remark is obviously we don't want to do something stupid and to look even worse and I agree with that, as well. Now, OK, then I have a last question: Would you guys sort of� in this instance Curtis is so pessimistic about progress between meetings. What is the feeling of the room? Would anyone have a really big problem with the NCC actually taking a stab at this? Because that is always regarded as something that we shouldn't do, we should be neutral in the policy process, but if we� I am quite willing to work ton and I am paid by the NCC. Of course I am asking for volunteers as well.
Randy: As one of those people that always objects when the NCC takes too strong role, I was going to say even though I am not of this region I will volunteer to help.
SPEAKER: OK.
Randy: But you can always kick me out for being not of the region. I do think members should seriously step up.
SPEAKER: I would love for the membership to seriously step up and I think�
Randy: But I am putting my money where my mouth is.
SPEAKER: OK. Thank you very much Randy.
Cathy: I would like to say this is supposed to be an open policy process, and not just the members can participate in that. And so anybody who� I mean we have people in the ARIN region, people who present proposals in the ARIN /RAOEPBLG /TPREPB all over the world and they are not necessarily an LIR in our region and that is the point of an open policy process.
SPEAKER: Well said. Good point. So I am waiting for all the volunteers to hit me in the coffee break. Thank you for letting me run over.
CHAIR: I think it was quite important. Thanks Daniel for bringing this up.
(Applause)
CHAIR: I didn't say my personal feelings on this yet because Daniel was nicely steering the discussion, anyway. I think this is sort of scary, to see how the queues really look like and what might happen and I am happy to help Daniel with trying to figure out a good way forward. Well, it's my job, anyway, but, still, I am interested in moving forward.
We have lost ten minutes of your coffee break. There is one item left on the agenda that is how to increase participation. I had no really specific proposal to make so I think I will just let it drop off the agenda and we will bring it back to the mailing list and try to get some feedback on that, but that wasn't sort of the most important item on the agenda, which is why it went left and I am not so sorry to see it drop. Thanks for attending. Thanks forgiving us good feedback and enjoy your coffee. And there is a social tonight, a RIPE meeting. It's starting quite early for RIPE dinner because, well, we are being sent to the desert so we need to leave early.
(Coffee break)