APNIC 32 - Destination::IPv6

Transcript - Global reports


Due to the difficulties capturing a live speaker's words, it is possible this transcript may contain errors and mistranslations. APNIC accepts no liability for any event or action resulting from the transcripts.

APNIC 32 Conference
Monday, 29 August 2011
Global Reports
16:00 (UTC +9)

Sanjaya: This is the Global Reports session, where we will hear updates from global organizations such as NRO and IANA and the Regional Internet Registries.

As you can see from the agenda on the screen, it is organized from the global reports first, followed by the Regional Internet Registry reports in alphabetical order.

We have one and a half hours to finish the session and we have 10 presentations, so the reports will have to be brief, 9 minutes each.

We will start with the NRO EC report.  Raul unfortunately could not be here, so the presentation will be delivered by Axel.

Axel Pawlik: As you can see, I'm not Raul Echeberria.  He is busy overseeing the building of their new office extension, which the NRO EC was busy inspecting two weeks ago.  It's going well, a nice view of the ocean -- again, of course.  That's probably the most exciting news I can tell you from the NRO, but of course I'll go through the whole presentation.

What is it?  It's a Number Resource Organization, a vehicle for the cooperation and coordination among the RIRs and also external presentation of the RIRs speaking with one voice.

We formed the NRO many years ago, in 2004, which is at the bottom of the slide, as one vehicle to protect the unallocated address pool, to promote and protect and help the bottom-up industry self-regulatory process and the policy development process, and also acting as a focal point into the RIR system for, say, global entries into the PDP and similar.

The NRO has established the ASO within the ICANN framework supporting organization, with the MOU signed in October 2004, and it is going quite well.

The current office holders, we rotate the officers every year.  The current chairman is Raul Echeberria, secretary is John Curran from ARIN, treasurer is Paul Wilson from APNIC, and next year we will switch one further.

There are a couple of coordination groups, the engineering coordination group, the communications coordination group and an ad hoc group from the registration services managers, basically talking about common issues in the registration services areas of the RIRs.

ICANN and the ASO: for a long time we distributed the cost of our participation in ICANN among the RIRs, and other expenses as well among the RIRs in the NRO are split along the same formula and takes into account the number of resources distributed and the like and changes a little bit from year to year.

Our contribution to ICANN remains stable.  It is USD 823,000 for all the RIRs together.

We participate in -- it is slightly out of date -- most of the ICANN meetings.  Somebody from the RIRs is always there, whether it's the NRO or an RIR, it doesn't make much difference.  We also went to the Singapore meeting earlier this year.

We heard about the IGF a long time ago, and that is something the NRO has been very active in, from even before its inception through all the World Summit on the Internet Society summits and pre-summits and in-between meetings, and obviously we think that the IGF as it exists now as a non-decision making forum is really, really good.  It has helped us tremendously over the last couple of years in establishing the RIRs and the NRO as a name among the governments of the world, so it is a very good thing.

From that came quite a lot of other participations.

We heard about -- I won't go into much detail -- for instance, from the RIPE NCC, cooperation with the OECD is also quite important.  We have seen those slides before.

Ongoing activities this year and last year: Obviously, the cooperation at the engineering level is very tight, especially around resource certification, RPKI.  We did a couple of workshops and meetings as the NRO, together with other ISTAR stakeholders, people in the Internet industry -- our peers, basically.

Earlier this year -- you must have heard about it -- we are running out of IPv4 address space.  I think you know that.  In February, that was the event where we did the big do, together with the ICANN press conference and handing out the last five /8s.

Last week or two weeks ago we met in Montevideo, hosted by LACNIC, as an NRO EC retreat.  Basically we talked about the ongoing business with a little bit more leisure, leaning back and looking out of the nice windows onto the sea and also talking about important stuff.

What we talked about this time is again our interface with the governmental work; let's say public affairs basically.  Like I said, this has been quite a success story for the NRO over the last couple of years.

However, we decided to change it slightly so that it is a little bit looser internally within the NRO, so we have agreed to set up or re-establish the public affairs coordination group.

RPKI, or resource certification, is a very important topic among the RIRs, and I will talk a little bit about that later in my NCC talk as well.  Of course, Raul wouldn't have said that.

Technical coordination among the RIRs, implementation, making sure the various systems we have implemented are compatible with each other.  We have looked at some regional discussions, some concerns about certification as a whole, as well.

We have an ongoing review of the Address Support Organization within ICANN and we talked about the progress of that and also about legacy space, how different RIRs are looking at it and what we can do to have nice interregional coordination of that for the best of all of us.

On our website, nro.net, is a lot of material, all our communications and statements are there, and you are welcome to look at all of those.  That is basically the NRO EC presentation for now.  Did I make the 9 minutes?

Sanjaya: Yes, well done.  Thanks very much, Axel.


Sanjaya: The next one is Guangliang Pan, who will present the NRO statistics report.

Guangliang Pan: Good afternoon, all.  My name is Guangliang Pan, the registration services manager at APNIC.  I would like to give you a quick update on the number resource allocation status.  This report is prepared by five Regional Internet Registries.  The data in the report is up to 30 June 2011.

First, let's look at IPv4.  This graph gives you an overview of the status of all 256 /8s.  Out of the 256 /8s, 31 /8s have been allocated before the RIRs existed, this is called the central registry, and 130 /8s are formally allocated by IANA to the RIRs.

There were 35 /8s used for special service, like telco receiver, so it's not available.

In February this year, IANA allocated its last five /8s to RIRs, and then the IANA people became zero.  So this is the final state of this IPv4 space, this is the overview of how the 256 /8s have been allocated.

Looking at the yearly allocation by each RIR, this graph shows the allocation of IPv4.  Actually, it keeps going up, in the last few years, especially.  Even this year, only six months, in the first six months of this year, the total allocation of IPv4 from the RIRs is about 10 /8s, but this will probably change because APNIC has reached its final /8s.  Members in the APNIC region can only receive a maximum of /22 from APNIC, which is the final /8s.

This is the graph showing the total allocation since 1999.  In the past 11 to 12 years, APNIC actually allocated more IPv4 addresses than other regions.

Let's move on to the AS number assignment.  The AS number assignment seems a bit stable.  Every year we are assigning about 5,000 AS numbers globally.  This year, in the first six months, we allocated 3,000 AS numbers from the five RIRs.  So it seems similar to the previous years, just a little bit going up.

This graph shows the total AS number assignments since 1999.  You can see ARIN and the RIPE NCC region have allocated more AS numbers than other regions.

Let's look at the 4-byte AS number which has come up as a new slide.  It shows how we allocate the AS 4-byte numbers.  It is very interesting to see different regions adopt the 4-byte AS numbers it seems quite differently.  You can see RIPE NCC and LACNIC allocate or assign more AS numbers than other regions.  It seems they have the least issues to assign the 4-byte AS numbers.

However, in the APNIC region and also the ARIN and AfriNIC regions, it seems they still have a problem issuing 4-byte AS numbers.  Quite a lot of members still ask for 2-byte, they say they can't support 4-bytes.

That's why you see on the statistics, ARIN, APNIC and AfriNIC have assigned the 4-byte AS numbers at this stage.

This is the total AS 4-byte number assignment at this stage, since it started in 2007.  You can see RIPE NCC has assigned most of the 4-byte AS numbers at this stage.

Finally, let's look at the IPv6.  This graph shows the overview of the total IPv6 address spaces.  In October 2006 IANA allocated one /12 to each RIR.  Since then, no RIR returned to IANA to ask for more.  That's why this slide has not been changed since that time.

But this may change in the near future because the IPv4 is going to run out and more and more policy is coming in about IPv6.  The policy may change and the large allocation may also happen frequently, so RIR may come back to IANA to ask for IPv6 in the near future, I think.

This graph shows the allocation of numbers made by each RIR by years.  You can see in the last few years IPv6 allocation is actually picking up.  Since 2009, people are talking about the IPv4 exhaustion and the IPv6 allocation actually increased a lot.

Some people may think ARIN region has a lot of IPv4, they probably don't care about IPv6.  But look at the statistic.  It is not the case.  ARIN actually allocated more IPv6s than APNIC region in the last few years.  You can see the allocation number of IPv6 in ARIN region is still quite high.

This is the graph showing the total IPv6 allocations.  On the left-hand side this is the total allocation number and on the right-hand side the graph is showing the size, counting by /32, compared to the total size.

You can see RIPE NCC, ARIN and APNIC allocated more IPv6 than other regions.

Finally, let's look at the IPv6 assignments.  The RIRs also make end-user assignments.  You can see the assignment size.  RIPE NCC is looking at more portable assignments than other regions, and ARIN also has a lot of portable IPv6 assignments.  It seems in their region they are keen to develop IPv6, based on the statistics.

All the statistics are on the websites.  This report is updated quarterly, so you can go to the NRO website to see the report every quarter.

Thank you.  Any questions?


Sanjaya: Thank you.  That concludes the NRO reports.  Are there any questions?  None.  OK.

The next one is the ASO AC report, Fujisaki-san.

Tomohiro Fujisaki: Good afternoon, everybody.  My name is Tomohiro Fujisaki, APNIC-appointed ASO AC member.

Today I am reporting on the ICANN ASO activity.

This is about ICANN ASO AC.  We are the NRO Number Council and the number resource advisory body for ICANN.

We consist of 15 members, three from each RIR.  In history, we have two elected members and one appointed member by RIR board.  Our term is two years for elected members and one year for the appointed members.

Our responsibility is to advise the ICANN board on the IP address issue and to select ICANN board members and ICANN NomCom members.  Our activities is we have a monthly teleconference and one face-to-face meeting per year.

Our members this year are on the slide.  In the APNIC region, at the last meeting, APNIC 31, the NRO member, Dr Kenny Huang, became APNIC EC and resigned from the NRO, and APNIC EC appointed Andy Linton as the ASO member of APNIC.

Now I'll report on our recent activity.  After APNIC 31 we had six meetings, five teleconferences and one face-to-face meeting at ICANN 40 in San Francisco.  More recently, we had an ASO AC workshop on IP addressing activities in every ICANN to introduce our activity.

I can report that at ICANN 40 I presented the policy activities in APNIC and at the last ICANN in Singapore Andy presented the same issue.

Global policy management is our main job, and currently we have two global policies.  One global policy, IPv4-2010, was discussed at the last APNIC meeting and abandoned in APNIC but this policy is still under discussion in AfriNIC.  Two RIRs have abandoned this policy, so this global policy did not reach consensus.

Another global policy, IPv4-2011, APNIC adopted at the last meeting and in other RIRs this policy is now under discussion.

The next topic is ICANN board selection.  This year, we don't have the ICANN board election but we revised the ICANN board selection procedure.  At ICANN NomCom we selected Hartmut Glaser from LACNIC as the NomCom member for next year.

The last point, our activities, we will have the same ASO AC workshop at the next ICANN and I will join the meeting and present about APNIC policies.

If you would like more information about the activities of ASO, please visit our website.


Sanjaya: Thank you.  Any questions?  Kenny.

Kenny Huang: Kenny Huang, EC member of APNIC.

I find it is very difficult to create an interIR transfer policy, especially based on the existing bottom-up process, so is it possible for an NRO, either EC or NC, to create a harmonized interIR policy proposal and present in the five RIR regions, based on their local policy structure?

Tomohiro Fujisaki: At ASO, that is not our job, but at NRO we should manage to have a global coordinated policy, yes.

Randy Bush: Kenny, we had a scheme that would work, even though that doesn't happen, which is that the person in the selling RIR has to obey the selling policy of that RIR and the person buying has to obey the buying policy of their RIR, and the two policies do not have to be coordinated.

Sanjaya: Any other comments?  Thank you.


Sanjaya: The ASO has one more item to report, and that is an ASO review process that is going to be presented by Raimundo Beca.

Raimundo Beca: The ASO laws state that the ASO and the ACs have to be reviewed periodically.  The review started with the DSO in 2006, and the ASO is the last one that is being performed these days.  At the end of this, a new cycle will start with the DSO next year.

The ASO is a particular SO, supporting organization, of ICANN, because according to the ASO MOU the ASO, the supporting organization, is provided by the NRO.  In accordance with that and with the ASO MOU, the review of the ASO is conducted by the NRO.

The NRO contracted an independent consultant, which is a French company called ITEMS International, and this enterprise contracted me as an expert in the ASO, to the extent that I was in the first address council installed in 1999, when I was appointed by ARIN, and then I came to the same address council, appointed by LACNIC, and then I went to the EC board, so it is something I know very well and I know very well many people in this room and in the community.

The point I want to make here is in the process we are making, which started in Singapore, we are conducting now interviews which are interviews of selected people as individuals -- about 80 people from the community were interviewed -- but we are also making a massive survey.  This massive survey is an online survey and we expect every person attending APNIC 32 to make an effort and fill in the survey.  It doesn't take very long, and in any case, if you have any difficulty or want to raise any questions, I will be around all the meetings.

Here on the slide you can see how it looks, the survey, and you have the link to it, and I fully encourage you as soon as possible to fill it in.  Thank you very much.

Any questions?

Sanjaya: We will help Raimundo by sending the URL to the participants if they want to participate in that survey.

Next up is Elise.

Elise Gerich: Hello, I'm Elise Gerich from ICANN, I am the IANA VP.  Never fear, I'm not going to do any more break dancing, although it might be more entertaining than my presentation, but you're safe.

First of all, some of you may know that the IANA contract with the Department of Commerce expires at the end of September of this year, 2011.  There was a clause in the contract that said that the Department of Commerce could extend it for six months, which they have done, so I'm not out of a job yet.  The contract has been extended until March 2012, while the Department of Commerce, NTIA, decides what they want to put out in an RFP or how they want to do the next procurement.

I would like to thank the ASO, the NRO and APNIC for sending in some of the 48 NOI responses that were received.  What the Department of Commerce did is they had a notice of inquiry, which is kind of like a marketing survey, and sent out some questions and asked people to reply.  Then they sent out a further notice of inquiry and got 48 more responses.  At this point in time they are reviewing those responses and will decide on what their next step will be.

Our guesstimate -- "our" being ICANN and IANA -- is that there will be probably an RFP published some time in mid-September and NTIA will make a decision as to who the contract will be awarded to, probably some time in 2012.  That is a guesstimate now.  Nobody knows right now, but we do have an extension until March 2012.

Business excellence 2010-11, what does it mean?

Some of you may have heard of total quality management or EFQM or some other quality management systems.  About two years ago, the IANA department within ICANN started a business excellence program based on EFQM.  For the last 12 months, we have gone through our documentation and processes, much of which was short narrative and old, and we said, okay, let's get a process work flow model and put our processes into this work flow model.

We have done that for the past 12 months and we are now setting some KPIs, which are key performance indicators.

Some examples of those are that we deliver reports on time that we owe to the IETF, IAB and NTIA; that we respond in timely manner to requests; and things of that nature.  This has been an ongoing activity that we do, along with the rest of our day-to-day jobs.

Finally, I am sure most of you have heard about the root zone automation, RZM.  ICANN and IANA have talked about this since 2005, I'm told.  That predates me, I'm in my 15th month, I wasn't here then, but we can claim success.  Whoo-hoo!  It rolled out on 21 July and now 23 of the requests have gone all the way from the beginning to the end of the root zone management system through the automation and we have 17 somewhere in that automation work flow.

What's happening now, those of you who are TLD owners will be receiving emails with your credentials.

We are rolling out the credentials in a gradual fashion.

The first 20 per cent of the credentials have been sent out and we have one person out of the approximately 60 to 70 credentials who have actually initialized their credentials and they have launched one of the requests that are in the system using their own credentials.

If you receive your credentials and you decide not to use them, that is OK too, you can still send email to iana.org, as usual, and a person will submit it into the RZM automated system for you.  However, we thought everyone would like to have their credentials because it gives you the ability to always query your ticket whenever you are ready to see what the status is and where it is.

This is a pretty big achievement for the team.

I would like to thank -- some of you know these folks -- Kim Davies, Simon Raveh, Nadia Sokolova and Naela Sarras for particularly being instrumental in launching this.

Cooperating with the RIRs; we do do this.  I even do break dancing for them!  The IANA and the RIRs are working together to clarify the process for IPv6 allocation requests.  You have heard some of this already in the NRO and the ASO reports, and basically we don't make policy but we have to implement it.  So when you engage with us and ask us if we can do something, we are happy to respond.

One of the conversations you have been having with us is the meaning of "reserved" and what does it mean potentially for IPv6 as far as allocation and the context.

Finally, there are a few other things that keep us busy too.  One of them is after the DNSSEC roll-out last year -- it seems like just yesterday -- we have received accreditation, the SysTrust accreditation.  I have been told that ICANN and the IANA department have been the first to receive it and now Verisign has also received it, so we both have accreditation for DNSSEC.  That is where an independent auditor comes in and says you have done things in a very good way.

Finally, we have had an MOU with the IETF/IAB since 2004 and they have SLA targets for the IANA team on handling all the requests for the registry and other things like that.  We have met or exceeded that and continue to.  The reports are posted on the IANA website, if you are ever interested, and we submit them monthly to them.  The SLA is 90 per cent, to meet the target 90 per cent of the time.  We have exceeded that, hitting 98 per cent some of the time.

Finally, mcast.net has been delegated to some new name servers and it's been DNSSEC signed, and this was done to improve stability and security.

Finally, I'm going to try to say this.  It's (speaks Korean).  Thank you.


Sanjaya: We are slightly ahead of schedule.  This concludes the global organization reports.  If you have questions to Elise or any of the previous presenters, please speak now.  Anything from the Jabber?

Thank you.

Next up will be the Regional Internet Registry reports in alphabetical order, starting with Hisham for AfriNIC.

Hisham Ibrahim: Hello again.  I will try to make this as brief as possible.  This is the AfriNIC activity update.

As you may see through the charts, all the scales are going up for IPv4 allocations.  IPv6, we have had a huge jump in 2010 and 2011, compared to the other years, still going up into the upper right corner of the page.

We have the ASN allocations going up as well, meaning the networks are more resilient.  The membership growth; apparently the IPv6 training program, initiatives and our IRM training is paying off in the region.

However, what I want to spend some time on, other than the charts that everybody would have anticipated going up, is some trend analysis that I was playing around with yesterday.  I was trying to think whether there were any events during 2011 that impacted the IPv4 allocation.  Of course, there was: the global depletion.

So I wanted to see how that affected the AfriNIC region in particular.  I took the year from 3 February 2010 to 3 February 2011 and I compared that to the other year, which is 3 February 2011 to what I projected and predicted up to 3 February 2012.

So far, after seven months, AfriNIC has already allocated more than it did for the entire last year.  If the projection goes in a linear growth, with nobody requesting a large /10 or /12 or something, which is not likely in our region, we would have allocated this year twice as much as we allocated last year in terms of IPv4 after the depletion.  So there has been an impact of the global depletion in our region, people have felt a bit and people have been rushing to get their IPv4 space.

I also played around with the same figures for IPv6 allocations and I wanted to know, were there any events that happened this year also that would have an impact?

We already have the global depletion of IPv4; did that impact the allocation of IPv6 in our region or not?

There was also another global initiative, an ISOC initiative that a bunch of major content providers joined in, which was IPv6 Day.  So I wanted to do the same measurement after IPv6 Day and after the IPv4 depletion; was there more of a trend to go to IPv6 in our region or not?

However, the thing is the impact of both of them were not as high as I had hoped.  Actually, there is not much impact on both.  You see the chart on the left represents this year, since the depletion day up to IPv6 Day, that is the part in blue, post-IPv6 Day until now, which is almost three months, is the red part, and the projection is the rest of it.  As you may see, there is an increase, however it is not as visible as what was there in the IPv4 thing.

However, bearing in mind, as I showed in the previous charts of IPv6 allocations in the region, 2010 and 2011 had a huge jump, so although these two events happening this year did not impact dramatically, they increased from last year.  However, this year and last year we saw a large growth of IPv6 allocations.  Again, I am stressing, this is allocations.  I am not talking about deployment in the region or task forces being set up or about initiatives being held, either nationally or regionally; these are just the numbers that as a registry we are looking at.

The AfriNIC 14 report: we had our first meeting for this year held in Dar es Salaam, Tanzania, and it had three main themes, which I may have touched on in my previous presentations.  It feeds into everything we are doing -- our two meetings, our IGF proposals, everything we are doing.  We have IPv6, interconnecting Africa and cyber security issues, and these are the main themes we are going with, and they will feed into our next meeting coming up.

During our meeting we had three days of training, we had our public policy meeting, followed by the African Government Working Group, which I may have mentioned in a previous presentation.

Policy debates, we had six policies during the meeting.  Two of them reached last call, consensus and progressed to final call.  One was the global policy for post-exhaustion and the other was the IPv4 soft landing policy, we finally got that to last call.  Three policies did not meet consensus and one was withdrawn.

The one that was withdrawn was the limited out of reach region allocation of IPv4 resources, which the author suggested that when other RIRs run out, they can come and take space from AfriNIC in exchange for a fee, and the community did not go for that at all.

AfriNIC is launching a new website.  It is online now.  You can get to it through our website officially from the first Monday of September, which will be 5 September.  You will find the original page and you will find a link that will take you to the new website, so you can browse it and go through it and compare.  We will phase it out step by step until our next upcoming meeting in November.

There is a bunch of new cool features that we have implemented there, but one thing with our website is it was only in English.  With the new website, we are introducing French, so it will be a bilingual website, so we can reach more Francophone speakers in our region.

We have a bunch of other initiatives -- RPKI, DNSSEC copying root servers, all of that -- but I do not have time to go into that.  So I would like to invite you all to our next upcoming AfriNIC meeting in Cameroon, AfriNIC 15, from 19 to 25 November.  Registration opens today.  I don't know if you get a prize for registering early, but I would really like you to register and I would like to see you in our meeting.

Thank you very much.


Sanjaya: Any questions for AfriNIC?  Thank you, Hisham.

Next up will be the APNIC report, which I will deliver.

Because this is the first APNIC meeting after we run out, I am going to basically report our last /8 implementation policy rather than the general APNIC update presentation.  The general APNIC update you can see in our AMM presentation on Thursday.

This is going to be focusing on the last /8 implementation, how we landed.

The good news is it's not as dramatic as we anticipated, thanks particularly to the communications team in APNIC.  We communicated our stock status well and kept updating people, so no one was really surprised when we hit the runway.

The background, the three stages of IPv4 exhaustion, what we did in the stage 1 to 2 transition and what we did in the stage 2 to 3 transition, what life is like in stage 3, and then conclusion.

The background, you probably all know, we had a final /8 policy that was finalized in 2009, so it is just waiting for the last /8 to be triggered, which is two years after the policy is approved.  Then related to that is the ICANN board approval of global policy of the remaining IPv4 address space, if IANA is left with only five /8 blocks, they will evenly distribute them to each of the RIRs.

At the Gold Coast, Guangliang presented our strategy in meeting the final /8.  That is our three stages of IPv4 exhaustion.  Stage 1 is when everything is OK, IANA still had IPv4; stage 2 is when IANA has distributed the last five blocks to each of the RIRs; and stage 3 was triggered when we, APNIC, reached our last /8.  In a diagram it looks like this.  It is simple, really.

Let's look at the stage 1 to 2 transition.  That is close to the time when IANA run out of their IPv4.

I think the challenge at that time was to predict the timing of the next request to IANA that triggers that last five /8 condition, so we have enough time to coordinate the public communication with ICANN, IANA and the other RIRs.  Everyone is looking at APNIC, because they know APNIC is growing the fastest so we would be the most likely to get the next request to IANA.

What happened was on 19 January we qualified to receive subsequent IPv4 allocation from IANA and we submitted a request, but we immediately asked Leo to hold the allocation just yet, because we really need to coordinate the public communication.  So there's a bit of a rush in making sure that we talked to other RIRs, NIRs, ICANN and IANA to prepare for the public communication.

On 31 January we received our next allocation from IANA, which was two /8s, and that actually triggered the last five /8s distribution to the RIRs.  Basically, APNIC got three /8s and the other RIRs got one /8 each, and we still finished our three /8s faster than the other RIRs.

Then on 3 February, IANA NRO announced the press release that you have read, so 3 February is the date where IANA runs out.

Stage 2 and 3 transition, the challenges are how to ensure to set up a really strict first come first served mechanism, so we are not accused of playing any favouritism to any economies, any members, any NIRs; we have to apply the same first come first served mechanism for NIR members as well.  The NIR members and the APNIC direct members need to be treated exactly the same, and that was the challenge.

We set up this rather complex diagram, which you can study from the website if you want to look into it in more detail.  It is self-explanatory really.

I want to spend a bit more time explaining it.  It consists of two different queues.  One is the evaluation queue and the second one is the approved queue.  The evaluation queue, the request will stay in the evaluation queue and keep pushing back to the bottom of the queue every time there is more information needed from the member, from the requester.  So that helps make sure that people who submit some very good requests have advantage over those who just submit a very low quality request.  So it turns out to be quite effective in making sure that those who provide us with good quality information, so that both the NIR Hostmaster and the APNIC Hostmaster can evaluate the request, get an advantage there.

Once the Hostmaster has enough information to make a decision on the request, then it goes into the second queue.  That is the approved queue.  Now, that is the strictly first come first served, and we actually look at the last response received from the requester, including the day and the hour and minute and second, because some of the communications getting back to us could be just 2 seconds apart and we want to make sure that those who are in the front are kept in front.

Then we delegate based on that sequence, the second queue sequence, the approved queue sequence.  Because of that, we are also able to estimate when we hit the last /8, the red line over there.  It went well.  It worked.

That method worked.

On 15 April, we announced that we have reached our final /8 of available IPv4 space and we approved 328 requests out of those queues.  181 requests missed their allocations, so there are 181 requests under the red line.  But most of them accepted the last /22, basically.  I really thank our members for being really understanding of this.  Out of the 181, we only received six complaints, and they have been dealt with.  That was really a good outcome.

What happens in stage 3?  This is just a quick statistic we compiled from 15 April, the day we hit the last /8, to 27 August, which is just before we started the meeting.  People are still coming for their IPv4 delegation, their last /8.  583 requests coming from 30 different economies, so that is approximately six requests a day for IPv4.  IPv6 also increased, 242 over 26 economies, which is 2.5 requests per day.

Membership also slightly increased.  Normally we get one new member a day.  Because of this, we are now about 1.8 members a day, on average.

Of course, we implemented the new policy that says that people can also ask for /24 instead of having to go for /22, so if they want /24 now and another /24 later, they can still do it, up to a maximum of /22.

The most common question we receive on our help desk is the members ask us, "We actually still need more IPv4 address space, and we appreciate your /22, thank you very much, but what can I do now?"  The only way out for us is to tell them that the IPv4 transfer policy is available, to allow transfer of unused space from one member to another, and we suggest they contact their network operators, mailing lists or use apnic-talk mailing lists to find transfer sources.

We also have approved interregional transfers, so we can receive resources from other regions where the other region permits it, basically.  We are actively talking to our colleagues in other regions to see if there is a matching policy over on that side that allows transfer into the APNIC region.

There is also still a possibility of IANA receiving more address space, returned address space, be it from legacy or whatever.  I think IANA has stated that they are willing to take custody of these returned address spaces and then redistribute them according to the global policy.

In conclusion, I think as a community we work well together.  We own this process together.  It's not just the APNIC Secretariat; the whole community worked well together to set up the policy so the APNIC Secretariat can manage the IPv4 exhaustion.  We successfully implemented it and carefully set up the plan and executed it and kept the community well informed throughout the stages.

Here are some policy references for you, and that's it, that's my presentation.  Thank you.  Any questions?


Sanjaya: Next up will be ARIN, Paul Vixie.

Paul Vixie: Good afternoon, Paul Vixie, ARIN Board of Trustees.  I will make this brief.

Our focus at ARIN for 2011 has been the depletion and transition.  As is true of all our RIRs, it has required a lot of re-engineering of internal processes -- again, like other RIRs.  There is a fair number of increased IPv6 activity, which we like, that is taking some staff time.  We are doing a lot of IPv6 outreach, which I will get to later in the presentation.

ARIN Online is completely functional now.  We are now adding new features to it -- I'll get to that.  We are working hard with the other RIRs on RPKI deployment.

Of course, we have got the DNSSEC rolled out, and we continue to participate in the global IGF process.

Some details: these are the v4 requests that have been received quarter by quarter.  What you should notice is that there was a big uptake in March.  We have all, in the RIR community, every RIR, benefited from the outreach of the panic.  It doesn't matter how many times we all tell our members, "Gee, they're running out of space, we've got some policy work to do," and so forth, it's when they read it in the newspapers that they call in and say, "Hey, we need more space."

That's been good, because it's caused a lot of people who weren't getting the outreach message any other way to hear about IPv6 and to start investigating that, so that has been a side benefit to this cloud of exhaustion.

Obviously, the requests that have been received since June are lower than the ones during the big press roll of February and March.  We are not out of space in the ARIN region, we have six /8 equivalents in our pool.

We are not so much seeing fewer requests because people think we are out; we have really tried to make sure people know we are not out yet.  We are back to normal, as the press roll is mostly finished.

These are the /24s that have been issued.  What you should notice here is a dramatic drop in the second quarter.  The reason for this is a policy change whereby ISPs can get only a three-month window when they come to us.  They used to be able to get more than three months of growth, now they have to keep coming back every three months, as part of the soft landing system and a bunch of interrelated policy proposals in the past few years that have caused this to drop off.  What this should mean is a nice smooth decline, rather than going straight to zero.

v6 requests received, again peaked in March when a lot of people heard we were running out of v4 and they had better get v6.  We certainly hope all the v6 allocations now get routed and used, which is kind of a secondary problem.

This shows the split between end user and ISP; no real change there.

What you should notice here is that about one third of our subscribers have dual stack.  This is, I guess, good.  We have a lot of subscribers; 33 per cent is much better than it was a year ago.  It is this chart which more than anything else justifies our continued outreach at all kinds of industry forums and gatherings and conferences and what not, to let people know they really need to be running dual stack.

In the last five or six years we have recovered quite a bit of address space.  What is not shown on this chart is that in the last year we have received 16 or so /16 equivalents which have been returned voluntarily.

Just to forestall the question -- yes, in every case we let them know, "Did you know that there will be an IPv4 market soon, because this is a scarce resource, and if you don't return it you will probably be able to sell it to someone soon?"  And in every case, at least in 16 or so /16 equivalents worth of cases, people said, "Yes, we know that, we don't care.  We got it free.  We are giving it back to you.  Do what you want with it."  In the last four years, about 30 /16 equivalents have been returned.

I don't expect that behaviour to continue when we are out of space, but at the moment anybody who wants space and qualifies for it can still get it from us in our region.  When that is not true, I expect there will be a lot of outreach by people who are not us toward the people who are not using their space and then we will get a different result.

We recently implemented policy 2010-14, where we standardized the IP reassignment registration requirements.  This amounts to extending some of the looser policies that cable ISPs were enjoying to the other DHCP related home facing ISPs.  We don't require quite the same utilization threshold but we require more information to be given to us in order to qualify for your next block.

Very interestingly, you are required to SWIP every /64 that you allocate, and that is a very different thing.  It used to be that only /56 sized allocations needed SWIPs, but now our members are SWIPing every LAN they assign.  That is increasing the utility of Whois to the extent that IPv6 is getting rolled out at all.

In 2011-3 we have loosened a lot of the v6 allocation.  Eventually, our community woke up to the fact that it is routing table slots that matter, not address space in v6, and we should just give large tracts to make sure people have enough so that they will never have to come back, or it will be a long time before they have to come back, to avoid the de-aggregation effects.  This will be beginning implementation in February.

There are four active draft policies.  John Sweeting, who is the Chair of the ARIN Advisory Council, is here with us.  If you have any questions about ARIN's policy process, John is your man, see him at the breaks and at the bar.

There are nine policy proposals in the mill at the moment.  They are all well described at this URL.  Some of the proposed policies are interesting.  2011-1, you will see, is pretty much our version of APNIC prop-097.

For a global policy to take effect it has to be approved in compatible form across all five RIRs.  Whenever there is a global policy, you will see the same thing being introduced everywhere but having a different number in each RIR.  This is the number we give it.

The point is this is a globally coordinated transfer policy, so that if someone out of region wants to exchange an address block one way or the other with one of our members, that will be possible, as long as we and that other region have compatible policies.

At the moment, the controversial part about that is that ARIN has a needs-based policy, so this would only work in the case of other RIRs that also have a needs-based policy.

We have a shared transition space.  This was controversial, 2011-5, because there were all kinds of transition technologies -- many of them discussed here today and yesterday -- that have required little bits of IPv4 space, or even a big bit of IPv4 space.  They did not need to be uniquely assigned.  In other words, it could be like network 10 or 192.168, where the same block is used everywhere and in order to avoid allocating unique space for a non-unique need, someone came to ARIN and said, "Can we set this aside and let everybody use it?"  It turns out we didn't have a policy for that.  In other words we only allocate unique space.

IETF is where network 10, 192.168 and some of the other non-unique space has been allocated.  The ARIN board pushed that policy proposal on to a sidetrack and said, "Could you go and work with the IETF and get them to ask for this?  Then we will find out exactly how it ought to be allocated and whether it comes from an RIR or came directly from IANA, as network 10 did."

I'm pleased to report at the last IETF meeting a lot of consensus appears to have been reached on this topic, so that is going to go forward.

Prop-137 will instruct IANA to accept returned address space.  Elise is highly cooperative, she would accept it even without this, but we felt it was necessary to have all five RIRs, through their community developed policy process, instruct IANA to accept it and also instruct IANA as to exactly how to dole it out.

That's pretty much the function of NROs, to be the steward of this global pool, even though we all accept IANA as our secretary for the global pool.  That, I think, is probably going to go forward.

Prop-151 would remove in most cases the needs-based evaluation for transfers to specified recipients.  We have a transfer policy, as APNIC does, and right now it requires that the recipient be able to demonstrate need in the same fashion that they would have demonstrated it had they been getting the space out of our pool.

Prop-151 proposes to change that and allow a recipient to get it for no stated reason.  We will see where that goes.

In prop-155, we are essentially trying to open a discussion.  As written, it specifies that an IPv4 address can only be used in our region.  This would keep, for instance, a multinational company who is a member of every RIR just as a normal fact of their business, from getting space from ARIN and then using it in a non-ARIN network.

I'm not sure that it will end this way, but you have to begin the debate somewhere.  Since we have a community-based policy development process, we are beginning here.  In Philadelphia I think this will -- John, did this make the docket for Philadelphia?  Yes.

So this will be discussed there and I think it will be a very interesting discussion.

I mentioned that ARIN Online has been drastically improved in recent years.  We are now integrating billing, so that you can see your invoices and pay them online.  You can do the specified transfers through ARIN Online, there are no more templates for that.  Of course, you can still use templates, but most people who have ARIN Online prefer it.  Once they see it, they switch to it.

We are going to be rolling out more IRR functionality.  I'm not sure how relevant that is in today's world, but we are doing it.

I mentioned earlier, we are working on RPKI in cooperation with the other RIRs.  So hosted is pretty much done, it's ready to be done in Q4; and delegated, which is the up/down protocol, will be done the following quarter.

I'm excited about this for any number of security-related reasons.  Just the idea that YouTube, for example, was able to get hosted out of Pakistan for a couple of days last year or the year before last should be a wake-up call to everybody, we can't let BGP continue to be unsecure.

DNSSEC is done.  If you have DS records for your in-addr blocks, you can upload them to ARIN through the ARIN Online portal.  Highly recommended.

Outreach events: the thing you should notice about this is that we went to the Game Developers Conference.

This is not your usual Internet Governance-type conference, and yet there are people there developing code that benefits in the presence of peer-to-peer non-NATed, non-UPnP sorts of networks, and we thought we would go there and tell them, "If you are suffering from having all your customers being on 192.168 then perhaps you would like to start helping us put some pressure on all the providers in the world by having your gamers want v6."  That's the kind of thing we are doing.  We are going everywhere and talking to everybody, and we are going to keep doing that, because ultimately it's the applications that will drive the deployment of any new feature, and to the extent that more address bits is a new feature, this is where it will come from.

Please come to Philadelphia, if you can.  Our policy meetings are open to anyone.  I hope to see as many as possible of you there.  The meeting following Philadelphia will be in Canada, in Vancouver, in April.

These are two great North American cities.  The food is good, the weather is good in these places at those times.  Please make all headway you can to come join us and help us discuss things and watch what we do.


Thank you.


Sanjaya: Thank you, Paul.  I noticed that World of Warcraft, probably the biggest online gaming in the world, actually participated in the World IPv6 Day, probably as a result of the outreach from ARIN.  Thanks for that.

Next up is LACNIC, Sofia.

Sofia Silva: Good afternoon.  My name is Sofia Silva, I work at the registration services area at LACNIC and I am going to present the LACNIC update.

First of all, about membership, as you can see, we have been growing.  As of the end of July there were almost 2,000 LACNIC members, compared to 124 in 2002.

Regarding resource, in February we received our last /8 block, which was 179.  As of the end of May we had more than four /8 blocks, which is equivalent to approximately 73 million IPv4 addresses.  According to our projections, we will be able to allocate and assign IPv4 addresses until May 2014.  You can find our daily report at the first link, which is just text, but it is updated daily.  If you want a complete report, you can find it at the second link.

Regarding allocations, as you can see in the first chart, the amount of IPv6 allocations and assignments has been growing and it will be growing because a new policy has been ratified recently, by which any organization requesting an IPv4 block for the first time must also request an IPv6 block.

On the second chart you can see the amount of IPv4 addresses we have been allocating and assigning through the last 12 months.

Regarding ASNs, the amount of 4-byte ASNs has been growing.  Since January this year we are assigning 4-byte ASNs by default.  Sometimes people contact us saying that their equipment or their providers' equipment doesn't support 4-byte ASNs, so we exchanged the 4-byte ASNs to 2-byte ASNs, but fortunately this is not most of the time.

Some of the important projects we are working on: first of all, customer service.  We have a new area which is a customer service area and a new manager.  The idea is that processes are more focused on the customers.

We are also working on RPKI.  Our system was launched in January this year.  We are working on hosted mode and we are implementing the delegated mode and we hope to have it in production by the third quarter this year.

We also are working on the DNSSEC, signing parent zones and planning to sign child zones.

On IPv6, we are organizing workshops and virtual seminars and also working on investigation and development.

Other important projects: AMPARO, which has the objective of providing training and promoting the creation of computer security incident response teams; FRIDA, which is the Regional Fund for Digital Innovation in Latin America and the Caribbean; Raices (Roots), which has the target of installing anycast copies of root DNS servers in the region; and SARA, which is the automatization of the resource management system used by members.

Our last meeting, LACNIC 15, was held in Cancun, Mexico.  There were almost 300 participants.  Seven policies were presented and three of them reached consensus and have already been ratified.

The other policies which reached consensus, the first one is about allowing the IPv6 de-aggregation for end users.  The second one is the one I mentioned before, which adds a new requirement for those organizations requesting an IPv4 block for the first time; the new requirement is that they also have to request an IPv6 block.

The third policy that reached consensus was the global policy for post-exhaustion IPv4 allocation mechanisms by the IANA.  This global policy completed the PDP at our region, now we have to wait for it to complete the process at the other regions.

Our next meeting will be LACNIC XVI in Buenos Aires in October; LACNIC XVII in Quito, Ecuador, in May next year; and LACNIC XVIII, will be in Montevideo, Uruguay, in October next year.  It will be LACNIC's 10th anniversary, so you will be very welcome.

Any questions?  Thank you.


Sanjaya: Thank you very much, Sofia.

Last but not least, Axel, we started this with you and we'll end with you.

Axel Pawlik: According to my watch, we have 9 minutes left.

I am the Managing Director of the RIPE NCC and this is the usual update.  What I want to do is not to go into relative details about the number of addresses we have allocated and the policies that are open -- there are people who know much more about that than I do.

I would like to do a high level overview of things that I find particularly interesting in our region that might have some repercussions here or that you might find particularly interesting too.

Geoff has said that the shortening of allocation times in ARIN, he has seen and observed, has led to having less space until 2018 or something like that.

That is interesting.  Since July we have a similar thing going, we call it the run out fairly policy, basically to avoid that one or two huge requests will clean out our coffers and many, many people will not have anything left.  That is in place right now.  So our members do get allocated space for the coming three months only.

In servicing those requests, basically what we did before is that you would put in a request and then if there were things to clarify, you would go back straight to the person that was talking to you before and go through your stuff.  We thought about that and we thought, that's not entirely fair because if you are not well prepared and you have to come back, you should go back to the end of the list, which is something we have implemented right now.  It is a little bit awkward in the change, people don't really like it, but I think it's fairer, so that's what we do.

We have seen a relatively high ticket load before the summer.  We do have a pronounced summer hole, we usually get that.  This time it is extremely low, but people will come back from holidays and continue their work.

We see that membership growth is much higher than anticipated.  We did look at it carefully at the end of last year and we wanted to be relatively conservative, but that is what you get and people do come.

The little graph shows the descent in the addresses we hold, IPv4 addresses.

We have had a couple of people leaving us.

Occasionally that leads to interesting changes in the company.  This one basically is about Mark Dranse, who left earlier, we looked at information services and you know what they were doing.  We merged them into what was the DNS group and the science group and now have something called GII, global information infrastructure, which is the guys running remote boxes, more or less, and we have research and department, headed by Wolfgang and Robert.

Now, certification: we all talk about number resource certification, or RPKI.  This is a very interesting thing to deploy, especially if it gets adopted by the operators and maybe even used for secure routing or maybe automated secure routing and stuff like that.  Of course, we have seen the potential a long time ago and we started very carefully to try to talk about it in the RIPE meeting in Lisbon, which was in October 2009.  People were focusing very much on business impact: what happens if the idiots at the RIPE NCC by accident throw me out and my part of the Internet is off the net and my customers are all very upset?  So, please, what are you proposing RIPE NCC do about that, and some policies around that.

That is not quite what we thought was really interesting.  We have, of course, deployed the certification system earlier this year.  About 600 of our members have requested certificates and a relatively astounding number of them are working with rollouts already, so there is some interest in the system.

However, at the last RIPE meeting in May, some people from the community got up and said, basically, "What are you doing?  You are implementing something that comes close to the red button to switch off the Internet, or parts of the Internet, at the behest of the bad guys in the world, the governments and the police, law enforcement and all those people."

Yes, that's what we tried to say some time ago.  Of course, there is real concern there.  It is not a black and white thing.  As we say, the RIPE NCC, basically we are tasked with running the registry, the registry must be correct, and certification is just a reflection of what is in our registry.

Whatever the operators do with it -- and of course they can go ahead and use the system to make their lives easier, to look into secure routing and stuff like that -- that is all very well.  It's not so that we would demand that they do it, it's not that they could do that either.  We are not the routing police, we will not be the routing police in future.  But people say, "OK, but what if the police think you are and the police force you to withdraw certificates, stuff like that, what would that mean?"  So it's all very interesting.

There was pronounced, let's say, resistance to that and people have said, "Stop this immediately."  That led to the retraction of what we had running for two years as a policy proposal on certification.

We have talked to the community further since May, and our board -- and I said this at the ICANN meeting recently as well -- has resolved to ask RIPE NCC staff to continue the development of the system.  I don't know it by heart yet -- saying, "Do be very aware of what the operators are doing and that the operators are independent."  They guide what the RIPE NCC is doing, of course; focus on tools and all that.

So basically what I'm saying is that this whole effort is not black and white, there are all interesting shades of grey in there.  What we want to do at the next RIPE meeting is to talk more about it, discuss this thing more.  We have a general meeting, at our general meetings the members usually tell us what to do.  They sign off on the charging scheme, and that gives us the budget and the budget allows us to do things.  So we will ask them, one way or the other, what they think about all this -- very interesting.

Also interesting is the financial stability, obviously.  RIPE NCC is a longstanding organization.  It is doing lots of things and also allocates numbers.  For a very long time, our way to do what we are doing was based on a charging scheme that is in turn based on the allocations of address space of Internet numbers basically over time.

So the old stuff isn't important any more, the new stuff is more important.  What do we do if there's not much new stuff left, as in IPv4 allocations?  We have to think about that.

We did that.  Basically we thought we must find a way to indicate what amount of benefit it is that you derive from our services.  Obviously, a big telecom doesn't get as much benefit as a small mum and pop ISP around the corner.  We thought, strangely enough, maybe the amount of addresses that you are holding is some indication.  If you are really big, you probably have a lot of addresses; if you are really small, you have less.

Basically we played with that, published it to the community, our members discussed the mailing list, and the main things that came from that are that there are still jumps between small and large, those jumps are too big, we want more jumps; or, actually, we don't want any jumps, we want a smoother sliding scale possibly.  So we are taking all this back in and talking to our board, and whatever comes out of this we hope will be finally and formally adopted during the general meeting at the next RIPE meeting.

Surveys -- always fun.  You all love to fill them in.  So I do.  We do them occasionally, roughly every three years -- actually, every three years the big ones.

We have done a new one, basically to look at what our members think of our services, what we are doing well, what we are not doing so well, what their needs are and what we should be doing in the future.

We have had focus group meetings and they were attended and run by Desiree Miloshevic and John Earls, all long-time members of various communities around the world.  The analysis is currently ongoing by the Oxford Internet Institute, where Desiree is working.  We waited for September.  We have a couple of extensive piles of paper that go very much into the details, the rough data, and from that we see there are no great big terrible surprises in there.  Of course, we have to look at what the Oxford people show us.  Ninety graphs, I think.

Membership and community development: as you know, we are working with the regional communities for a long time, especially in Russia, so far always in Moscow, and around the Middle East.

Next week, for the first time we go to the Balkans, to Dubrovnik.  Those things have been very successful.

People love us coming to their areas.  MENOG and ENOG, the local operators groups have been set up, and some board members returned from some of those regions and said, "We have got to do more."  OK, we do more.

I'm sending Paul Rendek to Dubai, to live there and schmooze with his friends around the corner and also to look after the Russian area.  That's what he does best and that's what he has to do.  He leaves a bit of a hole in Amsterdam, so we fill up the hole with Serge Radovcic, who you probably know from the Exchange Point area.  He will start with us in October, taking over our comms department and looking after the RIPE community, RIPE meetings and the like.

We have more regional meetings and we have the big RIPE meeting coming up in Vienna and I hope to see most of you there.

If there are questions, if we have time for that, I will be with the coffee out there.

Sanjaya: Thank you.


Sanjaya: That concludes the global reports session.

Before you leave, there are some administrative announcements.

The NRO NC election to appoint the APNIC ASO representative will be held on Wednesday from 11:00 am to 4:00 pm during the Policy SIG.  This is open to all meeting attendants.

For members, online voting is open now, since last week, and it will close at 9:00 am tomorrow.  If you are an APNIC member and you have not cast your vote, go into APNIC and cast your vote.  If you need any help, contact the service desk outside the room.

The next session will be at 6:00, and that is in this room on network security discussion.  Thank you.

Key Info


Paradise Hotel,
Busan, South Korea


28 August -
1 September 2011

Program included:

AMM, Policy SIG, IPv6 plenary, APOPS

Please wait...