January 2010 - ipv6-wg - mailman.ripe.net

List of IPv6 measurements published on RIPE Labs
by Mirjam Kuehne 22 Jun '10

22 Jun '10

[I am sorry if you receive this more than once.] Dear colleagues, Please see two articles related to IPv6 Measurements recently published on RIPE Labs: 1. A list of IPv6 measurements various people and organisations are doing with a short description and pointers to the actual results of the measurements http://labs.ripe.net/content/ipv6-measurement-compilation 2. An interview with Alain Durand about the IPv6 measurements he is doing at Comcast together with the University of Pennsylvania http://labs.ripe.net/content/ipv6-monitor We would like to get feedback from you about the measurements presented. How would you chose to measure the growing deployment of IPv6? Have you set criteria for when you will deploy IPv6? Let us know what information and analysis would be helpful to assist you in the deployment of IPv6. And would measurements influence your decision to deploy IPv6 or not? Please go to the forum at http://labs.ripe.net/node/ipv6-measurements or contact me directly at labs(a)ripe.net. Kind Regards, Mirjam Kuehne RIPE NCC

1 1

IPv6 allocations for 6RD
by Jan Boogman 11 Mar '10

11 Mar '10

Hi, we are planning to offer IPv6 connectivity to our xDSL and FTTH customer base via IPv6-6RD. (see <http://tools.ietf.org/html/draft-townsley-ipv6-6rd-01> or for a working implementation Free's presentation at RIPE-58: <http://www.ripe.net/ripe/meetings/ripe-58/content/presentations/ipv6-free.p…>). We asked RIPE NCC for a larger than /32 allocation (because of the way how 6RD encapsulates the customers IPv4 address in his IPv6 address and also because we want to give the customer a small subnet). The IPRA now isn't sure how to proceed and after consultation with the Policy Development Manager asked me to pose the question to the community at large. What do you think, should RIPE NCC allocate larger than /32 blocks to LIRs for cases like this and not only based on the number of end-customers? Cheers Jan Swisscom IP-Plus - AS3303

12 27

Request for approval of the minutes for wg session RIPE 59 (20100207)
by David Kessens 10 Feb '10

10 Feb '10

Please find below the minutes from the IPv6 WG session at RIPE 59. The minutes will be declared final if no objections, except for minor editorial corrections that will be fixed before publication, will be received by the end of Sunday February 7, 2010 in a timezone of your choice. Finally, I would like to thank Alex Band for the excellent minutes! Thanks, David Kessens --- RIPE Meeting: 59 Working Group: IPv6 Chair: David Kessens Date: Tuesday October 6, 2009, 16:00-18:00 David welcomed the attendees and opened the session. A. Administrative Matters o Welcome o Select scribe: Alex Band o Select Jabber Monitor: Rumy Kanis o Finalize agenda o Approval of minutes from previous working group meeting B. RIPE NCC IPv6 Update James Aldridge, RIPE NCC http://www.ripe.net/ripe/meetings/ripe-59/presentations/aldridge-ncc-update… There were no questions. C. Future of the Working Group: Charter Discussion Shane Kerr, David Kessens. Input from the audience http://www.ripe.net/ripe/meetings/ripe-59/presentations/kessens-v6-wg-futur… Shane mentioned that the initial idea for an IPv6 Working Group came from looking at the statistics Geoff Huston publishes. "Make IPv6 Happen", was the summary of the charter of the WG. Most of the discussion on IPv6 is actually in the Address Policy WG, so he proposed to shut down the WG and move on to a bigger and better thing. David asked the audience for input. Olaf Kolkman (NLNetlabs) asked how the activities of the WG tied into the IPv6ActNow website. Shane said that there is not really any tie in, and that he thought that is was a shame that something like IPv6ActNow is necessary at all. Rob Blokzijl (RIPE Chair) mentioned that the message that is coming out of the WG is heard by the community at large, however, it doesn't seem to be received and digested by the industry on a wide scale. He proposed that the WG is not shut down, because there is important work to be done, especially in the coming two years and when IPv4 has run out. Gert Doering (Spacenet) commented that the WG is not the place where people go nowadays to learn about IPv6. At the RIPE Meeting, those kinds of discussions are held in the Plenary. Rob gave an example: there were two presentations about the general implementations of IPv6 at customer premises. In one case, the CPE worked fine, in the other it didn't. He would have liked to have seen a technical presentation in the IPv6 WG session on why that is. Shane responded there are definitely opportunities for the WG in that capacity, and even other topics related to IPv6, like Multi-layer NAT. Denesh Babutha (Aexiomus//Cyberstrider) agreed that the WG had reached the end of its useful life as originally envisaged. He said that, as such, the WG needs to evolve, and he thought that this needed to happen through the the charter. Denesh added that the WG needed to move away from repeating the IPv6 situation in every single meeting and move onto more useful discussions. The new charter should be inclusive of other industry areas too. David said that there were many topics covered by the WG that were not originally in the charter. Bernard Tuy said that there is an enormous need for technical training and that kind of knowledge can come from the WG. David asked the audience how to move forward with the WG and whether it should continue in it's current form? Ruediger Volk (Deutsche Telekom) commented that the current form seems to be "quasi-plenary". Rob commented that the WG should be modest in its aims and not try to reach the end goal without some intermediate steps. He said that the WG should focus on real world implementations. Shane commented that the WG did not need to come up with a new charter here and now. Rob proposed that a new draft charter is written. David summarized the feedback for a new charter as more focus on education and outreach activities. Maarten Botterman said that not all of those activities have to be within the WG. David asked the audience if preparing for a hybrid IPv4/IPv6 world should be part of the charter. Rob commented that outreach should lie with the RIPE NCC efforts, such as the IPv6ActNow initiative. He said that training is a sensitive issue because the RIPE NCC does not want to compete with its members. Rumy Kanis (RIPE NCC Training Services Manager) commented that the RIPE NCC offers IPv6 training, but careful steps were taken to ensure the RIPE NCC does not compete with the membership. Marco Wertejuk (Binconsult) said that he believes the focus for the new charter should be "how to encourage conservation of IPv4 and deployment of IPv6". David said that he assessed from the audience feedback that there isn't much support for the WG to focus on non-technical outreach activities. David asked Rob for a time line. Rob said that the WG should take things slowly, and propose a new charter at the next RIPE Meeting and that further discussions should continue on the mailing list. D. Future RIPE Network Experiments - David Kessens. Input from the audience David asked the audience if there was further interest in IPv6 network experiments during a RIPE Meeting. Gert said that forcing IPv6 on people is actually a good idea. He said we should even go as far as making the RIPE Meeting network IPv6 only for the whole week. He added that we should do the experiment without NAT-PT. David said that he would like to see a completely new experiment. Marco Hogewoning (XS4all) said that although IPv6 connectivity in general is good, tunneling is needed in some places in order to get full connectivity. David Wilson (HEAnet) said that the IPv6 hour should be repeated, and that the WG should get new experiences because things have changed since the last time it was done. Izumi from JPNIC that for their experiments, they created different testbeds for different kinds of implementations. David asked the audience if the IPv6 Hour should be rerun at future meetings.. There is consensus that it should be. E. IPv6 Deployment: What are the Remaining Issues and Bottlenecks? - A panel discussion with interactive input from the audience. Moderator: Maarten Botterman, GNKS Consult http://www.ripe.net/ripe/meetings/ripe-59/presentations/botterman-towards-v… Gert Doering introduced himself and asked the question "What kind of Internet do we want to have in five years?". Geoff Huston introduced himself and in a slide set proposed the question "Is the IPv6 transition an example of market failure?" http://www.ripe.net/ripe/meetings/ripe-59/presentations/huston-ipv6-transit… Kurtis Lindqvist introduced himself and explained his own experiences with IPv6 deployment. He said that doing a slow and gradual implementation has been very beneficial. Maarten Botterman did a live survey, with the question "What would you gain by postponing, and what would you gain by stepping up to the plate, earlier rather than later?". 40 votes were cast, most go to: o By stepping up to the plate I have time and space to do the transition, which will allow me to do it gradually and well thought through o By stepping up to the plate I confirm my brand's image as 'state-of-the-art' or 'top-of-the-wave' Geoff commented that the respondents are most likely non-representative. Gert said that he is not surprised with the numbers. Kurtis said that looking at the responses, he is optimistic. Maarten said that some people responded with 'none of the above'. He asked three of those people to comment: Ruediger said that for all of the relevant players there is still enough time for doing a proper job. Bernard said that there was no vendor support, so ISPs will never start until that happens. Jan Hugo Prins said that there will never be end user demand. Deployment is not picking up because of security issues. Geoff reiterated that he thought that the situation we are finding ourselves in right now is no accident. It is in many companies' interests to delay IPv6 adoption. Shane agreed with Geoff. Gabriella Paolini (GARR) said that the biggest issue is that middleware doesn't support IPv6 and that changes very slowly. Maarten did another live survey with the following question: "There is perception and experience of/with a number of hurdles that seem to be difficult to overcome. Which are important to you?". o As there is very little customer demand, I cannot justify the investment to my management, even if I would like to move o I have the experience that my vendors cannot deliver what I need o None of the above (highest score) Ruediger said that the options are very extreme. Slow and gradual deployment is a real option. Geoff commented that in his days at a telco, the biggest cost turned out to be customer support. IPv6 could become expensive in that respect with end users calling in because of issues. Maarten asked people to comment: - Wilfried Woeber (Vienna University) commented that he would have to select 'none of the above' for all questions. His issue is that the questions are too black or white. The biggest problem from his point of view is in layer 7 or 8: infrastructure, applications etc. The backbone has been done for years. - Marco said that he agreed with Geoff, the cost is not in the hardware. IPv6 is not a product, don't pay extra for it. - An attendee said that in a lot of cases, even though the product spec sheet says that IPv6 is supported, there is no feature parity with IPv4. Gert said he agreed with most of the comments. Geoff said that any vendor is willing to give you what you want, as long as you put down the money. Kurtis added that there was also a big push from vendors for MPLS. Marco added that some IPv6 functionality used to be free, but after vendors realized that there was a market for it, they started charging for a license. Maarten did another live survey, the results with the most votes were: o The deployment of IPv6 is unavoidable for my organization in due time o Technically, everything is in place to deploy IPv6 Geoff commented that openness is the key in the end. IPv6 is the only way to guarantee that. Kurtis added that he was surprised that so many people responded "I will deploy in due time", because that is pretty close. Geoff re-stated that the market is well informed, and that this situation is intentional in order to maximize shareholder profits at large corporations. Ruediger agreed that the industry is flooded with information. But he asked how enlightened are people? Geoff responded that it's all about vendor lock in. Brian Nisbet (HEAnet) said that the WG keeps having the same discussion meeting after meeting. He said that he could not believe that people are not well informed. Kurtis said that CFO's cannot be convinced of the urgency of something that will happen two years from now. An attendee said that for mobile operators, the vendors of mobile terminals are the blockers. Maarten showed the results of the last survey. The panel shared some closing thoughts. The general thought was that the community should stop distributing messaging, and start moving packets. Maarten closed the discussion. F. Developments/Initiatives Regarding IPv6 in the RIPE Region and Beyond - Input from the audience No input was given. Y. Input for the RIPE NCC Activity Plan - Input from the audience No input was given. Z. AOB None. David closed the session. ---------

2 5

Youtube over IPv6!
by Martin Millnert 03 Feb '10

03 Feb '10

Hello fellow IPv6 admirers, (I'm sending this here on behalf of Shane. :) ) As you might know already, Google has been so *tremendously* kind to, as of ~22 hours ago, initiate the enabling of their Youtube site over IPv6. It is currently available to participants in their Google over IPv6 program, http://www.google.com/intl/en/ipv6/ . The www.youtube.com user interface so far does *not* have any AAAA published, but the *much* more important (traffic wise) image and video servers do. Use tcpdump to verify. Eg., anticimex@hsa:~$ host s.ytimg.com s.ytimg.com is an alias for static.cache.l.google.com. static.cache.l.google.com has address 74.125.13.213 static.cache.l.google.com has IPv6 address 2001:4860:4001:402::15 anticimex@hsa:~$ host v1.lscache1.c.youtube.com v1.lscache1.c.youtube.com is an alias for v1.lscache1.l.google.com. v1.lscache1.l.google.com has address 74.125.97.80 v1.lscache1.l.google.com has IPv6 address 2001:4860:4001:402::10 The time when it began is clearly visible in our own IPv6 graphs, http://stats.csbnet.se/public/ipv6/csbnet-ipv6-traffictypes.html , and you can see similar correlated data at for example http://www.de-cix.net/content/network.html . To figure out all the mappings they give you, if you want to trace etc, something similar to: "for i in `seq 1 24 `; do for y in `seq 1 8 `; do host v$i.lscache$y.c.youtube.com ; done ; done | grep 'IPv6 address' | cut -d' ' -f 1,5 | sort -uk2" could be used - at least from where I sit. Adjust for correctness. I'm sure that as this matures and bugs are wiped out, there will no longer be any need for the Google over IPv6 program and everybody will be able to access this. Until then, join! :) I bow to the Google IPv6 team, their netops and everybody there that is working so hard to give the Internet true CONTENT on IPv6. Yippie! From 48514, much of love! Thank you guys (and gals)! Regards, -- Martin Millnert <millnert(a)csbnet.se>

1 1

IPv6 Act Now news update
by Chris Buckridge 28 Jan '10

28 Jan '10

Hi all, In what I hope will be a regular addition to the IPv6 WG mailing list, I wanted to provide a quick round-up of some of the items we've featured in the IPv6 Act Now newsfeed over the past week. These are provided for your information, but also, we hope, to provoke some discussion. You're obviously welcome to discuss on this list, or you can leave comments on the new pages themselves at www.ipv6actnow.org. I'd also be keen to hear feedback on whether the WG feels that this is a useful activity! Regards, Chris Buckridge, RIPE NCC IPv6 interview: Maria Häll, Swedish Government (28 Jan 2010) Maria Häll, Deputy Director for the Ministry of Enterprise, Energy and Communications in Sweden, and Co-chair of the RIPE Cooperation Working Group, talks about the role governments can play in encouraging IPv6 deployment in the latest IPv6 Act Now video interview. <http://www.ipv6actnow.org/2010/01/maria-hall-interview/> Comcast to start a series of public IPv6 trials using 3 transition mechanisms (28 Jan 2010) Comcast announced plans to conduct production-network trials of IPv6 technology in the coming year. The trials are aimed at helping identify and solve any areas of difficulty involved in the transition to IPv6, and determining what approach will be the easiest and most seamless for Comcast's customers. <http://www.ipv6actnow.org/2010/01/comcast-to-start-a-series-of-public-ipv6-… > Blue Coat introduces industry’s first IPv6 secure Web gateway solution (25 Jan 2010) Blue Coat Systems announced that it has introduced the industry’s first IPv6 secure web gateway solution to enable the secure and seamless migration of applications, services and content between IPv4 and IPv6 environments. <http://www.ipv6actnow.org/2010/01/blue-coat-introduces-industrys-first-ipv6… > 2009, the year before IPv6 took off? (22 Jan 2010) Emile Aben, of the RIPE NCC Science Department, talks about some of the signs of IPv6 adoption seen from the RIPE NCC, in terms of traffic to RIPE NCC websites and services. [links to the RIPE Labs website] <http://www.ipv6actnow.org/2010/01/2009-year-ipv6-took/> BT Diamond IP awarded U.S. patent for integrated management of IPv4 and IPv6 (21 Jan 2010) Shane Kerr mentioned this briefly on the list earlier in the week: BT Diamond IP announced the awarding of U.S. Patent 7,623,547 for "Internet Protocol Address Management System and Method", an IP Address Management (IPAM) system which supports both IPv4 and IPv6 address space using a unique hierarchical container structure and block allocation techniques. <http://www.ipv6actnow.org/2010/01/bt-diamond-ip-awarded-u-s-patent-for-inte… >

1 0

Comcast IPv6 trials
by Sander Steffann 28 Jan '10

28 Jan '10

FYI: Jaap Akkerhuis just brought this to the attention of the Dutch IPv6 task force list: <john_brzozowski(a)cable.comcast.com> wrote: > Folks, > > I am emailing you today to share some news that we hope you will find > interesting. > > Today we are announcing our 2010 IPv6 trial plans. =A0For more informatio= n > please visit the following web site: > > http://www.comcast6.net > > We have also made available a partial, dual-stack version of our portal > which can be found at: > > http://ipv6.comcast.net

1 0

[Fwd: IPv6 deployment scenarios]
by Brian E Carpenter 26 Jan '10

26 Jan '10

Please excuse the interruption if you've already seen this. -------- Original Message -------- Subject: IPv6 deployment scenarios Date: Fri, 22 Jan 2010 10:37:08 +1300 From: Brian E Carpenter <brian.e.carpenter(a)gmail.com> Organization: University of Auckland To: nanog(a)nanog.org Hi, Sheng Jiang (Huawei) and Brian Carpenter (University of Auckland, research consultant to Huawei) are currently running a questionnaire on IPv6 deployment, addressed to every ISP. The purpose is to provide facts for a document about deployment scenarios that we are drafting for discussion in the IETF. We will keep your reply strictly confidential and we will publish only combined results. We will not identify information about individual ISPs in any published results. If you request it, we will not mention you or the ISP in the acknowledgments. Please find the questionnaire at http://www.cs.auckland.ac.nz/~brian/ISP-v6-QQ.html Answers are requested ASAP. (Hmm. As I write, the IPv6 access to that server is broken. Hopefully it will be back soon.) btw, we know that the questionnaire is imperfect; all questionnaires are imperfect. Also, this is not a marketing survey; we are after technical information. Regards Brian Carpenter

1 0

Call for nominations for co-chair position ipv6 working group 20100202
by David Kessens 25 Jan '10

25 Jan '10

I would like to invite everybody who is interested him/herself, or would like to propose somebody else for the position of co-chair of the ipv6 working group. Nominations should be received by me or on the list by close of business in a time zone of your choice by Feb 2, 2010. I already received two nominations for Shane Kerr <shane(a)time-travellers.org> and Marco Hogewoning <marcoh(a)marcoh.net>. The procedure that we plan to use to select our co-chairs is outlined in: http://www.ripe.net/ripe/maillists/archives/ipv6-wg/2010/msg00028.html As mentioned in the above mail, the exact selection procedure is still a proposal and subject to change. David Kessens ---

5 5

Two new IPv6 related articles on RIPE Labs
by Mirjam Kuehne 25 Jan '10

25 Jan '10

Dear colleagues, Please see two IPv6 related articles published on RIPE Labs recently: 1. REX - the Resource Explainer - is now also supporting IPv6 addresses. See a description and some examples here: http://labs.ripe.net/content/rex-supports-ipv6 Please check out the tool and let us know if you have any comments or suggestions. 2. We have been measuring IPv6 usage on RIPE Labs and other servers and found some encouraging results. You can find more details here: http://labs.ripe.net/content/2009-year-ipv6-took Kind Regards, Mirjam Kuehne RIPE NCC

1 0

Proposed procedure for ipv6 working group cochair selection process
by David Kessens 22 Jan '10

22 Jan '10

All, As mentioned earlier, I would like to write down our approach for co-chair selection. This would help us to have an open and transparent process to get a cochair appointed, but also help in the future for other working groups as either an example or at the minimum that they can use the same set of principles. I believe that this request already has resulted in a quite bit useful discussion on a few things that we should, or shouldn't do and many of you mentioned that it would be useful if we don't wait too long as our new charter would also be a good starting point for the appointment of (an) additional cochair(s). Below, I will summarize some of the key principles that I came away with and my initial idea on how to start the process. Since these are initial ideas, I very much welcome your input or if you would be interested in volunteering to work out the text further. At the same time, there doesn't seem to be much disagreement about the first step of the process so I would like to start with that step right away: an open call for candidates. I realize that there is a bit of risk involved on starting this before we have agreed on the full process: however, there has been already quite a bit discussion on the list on the general direction and I have confidence in our community that we will be able to find a workable solution along the way. I will send a seperate mail with an official call for nominations. David Kessens --- Date: Thu Jan 21 21:36:39 PST 2010 Draft Proposal on handling of chair position selection within the RIPE community RIPE (Réseaux IP Européens) is a collaborative forum open to a parties interested in wide area IP networks in Europe and beyond. Work is carried out within a variety of working groups. Each of these working groups has more than one co-chair. The selection process for a chair position is deliberately left to the working group, so that at any given time, the community is free to select the organisational structure that best reflects the work they want to do at that time. Working group chairs are responsible for organizing the working groups activities as defined in their charters. In addition, they are required to participate in the policy making process to judge and ensure that the RIPE policy making process has been properly followed after new policy proposals have been accepted by a working group (http://www.ripe.net/ripe/docs/pdp.html) While the actual implementation of the chair position selection process is on purpose not codified in any process document, it is expected that such a process follows the same principles as outlined in the RIPE policy development process: 1. It is open to all and follows an established, bottom-up process of collaboration. Everyone interested in the well-being of the Internet may propose a policy and take part in the discussions that follow on from the proposal. 2. It is transparent. All discussions and resulting actions are documented and freely available to all. 3. Conclusions are reached by consensus. 4. All policies are documented within RIPE Documents and placed in the RIPE Document Store The same dispute resolution process as used for the policy development applies to cochair selection process or results of the process. The Process As there is no standardized process for working group selection, this section serves as an example only and describes on how the IPv6 working group is currently selecting its co-chairs. The following is a proposal only and it has not been decided on how to continue after the nomination step. - Call for nominations of more than 7 days (self nominations being ok) - During the same time period: 1) The nominees are given a chance to determine whether they want to be considered or withdraw as they decided that there are other good candidates already available 2) The community may express opinions that could help the candidates to come to a decision on whether they want to be continued to be considered as a candidate - the community may consider appointing more than one co-chair in case of several good candidates - If the discussion period seems to result in a consensus decision, a last call of more than 7 days is issued or - Each nominee writes a short statement to the WG Mailinglist - Election using single transferable vote (we need to figure out whether we can do this online or wheter this is better done at the meeting) ----

5 6

Ars Technica article about the last 20^H^H10% of IPv4 space
by Shane Kerr 22 Jan '10

22 Jan '10

All, I hate to interrupt our normal process discussions, but I liked this article by Iljitsch van Beijnum on Ars Technica yesterday: http://tinyurl.com/yaqu3f9 Maybe not news for folks here, but a nice explanation of different views of how much space is actually available. Sadly, I think the summary is correct: "it's now too late for a smooth IPv4-to-IPv6 transition, so prepare for a bumpy one where we run out of IPv4 addresses before there is widespread IPv6 adoption." Fasten your seatbelts everyone! -- Shane

1 0

working group co-chair(s)
by David Kessens 22 Jan '10

22 Jan '10

All, During the charter discussion, some of you suggested to appoint a co-chair. Most RIPE working groups have more than one chair person. In the past, I have never felt the necessity to do this job with more than one person. However, in the spirit that it is always good to try something new or different, we can certainly add a co-chair if the working group believes this to be beneficial for achieving our goals. Considering that we don't have any formal process for appointing a working group chair, there is no standard on how this is achieved in a transparent and fair way. This might have been acceptable in the early days of RIPE when there were far fewer people involved and governments weren't watching us. Given that there is an interest in appointing an additional cochair, this seems an excellent opportunity to achieve something more than just appointing a co-chair: if we define an open process and are happy with the results, we can propose the process as a RIPE policy proposal to the wider community for the appointment of chair people within RIPE. At the same time, I don't believe we should overregulate and loose one of the key benefits of our loosely organized community. Therefore, my personal preference would be to formulate a set of principles that we need to follow to appoint chair people while the actual implementation can vary depending on the position that needs to be filled. I would first like to hear what people think about this. If there is support, I would like to form a small task force to write a first version of the principles and how we plan to implement them in the ipv6 working group. David Kessens ---

15 22

New charter approved
by David Kessens 14 Jan '10

14 Jan '10

The Last Call for our new charter resulted in a very large number of positive responses and nobody at all speaking up against the new charter. I therefore conclude that we have consensus on the charter. I did receive a few comments from the working group chair list that the first sentence mentioning the 'IP protocol' is a bit awkward as 'IP' already stands for 'Internet Protocol'. It was suggested that we simply use the words 'Internet Protocol' instead of 'IP protocol' to avoid this issue. As this is really only an editorial change, I will go ahead and make this change when I request the RIPE NCC to post the new charter. Please let me know if you have an issue with this editorial change. I would like to thank Shane as the author of the charter text and everybody else who contributed with comments and suggestions for the charter discussion. David Kessens --- New proposed charter for the IPv6 working group: (first sentence was changed from: IPv6 is the next generation IP protocol to: IPv6 is the next generation Internet Protocol) ----------------------------------------------------------------------- IPv6 is the next generation Internet Protocol. The IPv6 working group exists to promote IPv6 adoption. The working group activities may be anything useful in helping people to deploy IPv6, and to manage IPv4/IPv6 co-existence. These activities include: * Outreach * Education * Sharing deployment experiences * Discussing and fixing operational issues The working group will cooperate with operators and others, both inside and outside the networking industry, to share resources and combine efforts. -----------------------------------------------------------------------

3 2

2009 IPv6 Address Use Report
by Iljitsch van Beijnum 02 Jan '10

02 Jan '10

[ (Non-cross)posted to NANOG, PPML, RIPE IPv6 wg, Dutch IPv6 TF. Web version for the monospace font impaired and with some links: http://www.bgpexpert.com/addrspace-ipv6-2009.php ] 2009 IPv6 Address Use Report Since 2005, I've been compiling an IPv4 address use report every year. With the start of the new decade, this is a good moment to start doing the same thing for IPv6. http://www.bgpexpert.com/addrspace-ipv6.php shows the amount of IPv6 space given out by RIR and by year, while http://www.bgpexpert.com/ipv6addressespercountry.php shows the amount of IPv6 address space by country. Both these pages are updated weekly from the delegation data that the RIRs publish on their FTP servers. 2008 saw a huge increase in the amount of IPv6 space given out and then a big drop in 2009 (amounts of IPv6 address space in the equivalent of /32s): 1999 10.88 /32s in 17 blocks 2000 19.75 /32s in 32 blocks 2001 33.13 /32s in 61 blocks 2002 156.75 /32s in 271 blocks 2003 261.38 /32s in 290 blocks 2004 13340.63 /32s in 295 blocks 2005 26985.00 /32s in 245 blocks 2006 9798.00 /32s in 243 blocks 2007 6687.01 /32s in 491 blocks 2008 81012.02 /32s in 886 blocks 2009 1091.03 /32s in 1280 blocks However, this is not the complete picture. The large number for 2008 is the result of two unusual events. The first one is LACNIC's delegation of the 2804::/16 block to the Brazilian national internet registry (NIR). At some point in the future, the delegation records will not show such blocks as "used" in their entirety anymore. Also, ARIN delegated 14 /22 blocks within the range 2608::/13 to the US Department of Defense. With these two artifacts removed, the amount of IPv6 space given out per year looks like this: 1999 11.00 /32s in 17 blocks 2000 20.00 /32s in 32 blocks 2001 33.00 /32s in 61 blocks 2002 157.00 /32s in 271 blocks 2003 261.00 /32s in 290 blocks 2004 13341.00 /32s in 295 blocks 2005 26985.00 /32s in 245 blocks 2006 9798.00 /32s in 243 blocks 2007 6687.00 /32s in 491 blocks 2008 1140.00 /32s in 871 blocks 2009 1091.00 /32s in 1280 blocks So the number of blocks given out keeps increasing, but their size is going down. There are two reasons for this: roughly between 2004 and 2006, RIPE and APNIC gave out some very large blocks to some very large ISPs. They mostly stopped doing that. And provider independent blocks started to be allowed and are getting more and more popular. These are the /32 - /35 allocations. /32 is the minimum block size given out to ISPs, this used to be /35. So this view shows the numbers of small-to-medium sized ISPs obtaining IPv6 address space: 1999 11.00 /32s in 17 blocks 2000 20.00 /32s in 32 blocks 2001 33.00 /32s in 55 blocks 2002 157.00 /32s in 254 blocks 2003 223.00 /32s in 251 blocks 2004 235.00 /32s in 241 blocks 2005 217.00 /32s in 217 blocks 2006 186.00 /32s in 186 blocks 2007 351.00 /32s in 351 blocks 2008 734.00 /32s in 734 blocks 2009 1011.00 /32s in 1013 blocks These are the blocks larger than /32 that go to large ISPs (excluding the BR NIR and DoD blocks): 2003 38.00 /32s in 3 blocks 2004 13106.00 /32s in 9 blocks 2005 26768.00 /32s in 13 blocks 2006 9612.00 /32s in 14 blocks 2007 6336.00 /32s in 8 blocks 2008 406.00 /32s in 7 blocks 2009 80.00 /32s in 4 blocks And these are the blocks smaller than /35, which are now mostly provider independent blocks, but also "critical infrastructure", such as root servers get a /48 block: 2001 0.00 /32s in 6 blocks 2002 0.00 /32s in 17 blocks 2003 0.00 /32s in 36 blocks 2004 0.00 /32s in 45 blocks 2005 0.00 /32s in 15 blocks 2006 0.00 /32s in 43 blocks 2007 0.00 /32s in 132 blocks 2008 0.00 /32s in 130 blocks 2009 0.00 /32s in 263 blocks So after a small dip in 2006, the number of small-to-medium sized ISPs obtaining IPv6 address space shows a steady upward trend, but apparently the very large ISPs either already got their IPv6 address space, are not focussing on IPv6 right now, are starting with a small block (or several small blocks), or a combination of all of these factors. Even with the BR NIR and DoD blocks included, the (equivalent of) 39395.56 /32s given out in 4111 blocks is only 0.026% of the 536870912 possible /32s in the currently defined global unicast space (2000::/3). For comparison, the number of IPv4 blocks given out is 99562.

1 0

2009 IPv4 Address Use Report
by Iljitsch van Beijnum 01 Jan '10

01 Jan '10

[ (Non-cross)posted to NANOG, PPML, RIPE IPv6 wg, Dutch IPv6 TF. Web version for the monospace font impaired and with some links: http://www.bgpexpert.com/addrspace2009.php ] 2009 IPv4 Address Use Report As of January first, 2010, the number of unused IPv4 addresses is 722.18 million. On January 1, 2009, this was 925.58 million. So in 2009, 203.4 million addresses were used up. This is the first time since the introduction of CIDR in 1993 that the number of addresses used in a year has topped 200 million. With 3706.65 million usable addresses, 80.5% of the available IPv4 addresses are now in some kind of use, up from 75.3% a year ago. So the depletion of the IPv4 address reserves is continuing in much the same way as in previous years: Date Addresses free Used up 2006-01-01 1468.61 M 2007-01-01 1300.65 M 167.96 M 2008-01-01 1122.85 M 177.80 M (with return of 16.78 M to IANA) 2009-01-01 925.58 M 197.27 M 2010-01-01 722.18 M 203.40 M These figures are derived from from the Internet Assigned Numbers Authority's IANA IPv4 Address Space Registry page and the records published on the FTP servers of the five Regional Internet Registries (RIRs): AfriNIC, which gives out address space in Africa, APNIC (Asia-Pacific region), ARIN (North America), LACNIC (Latin American and the Caribbean) and the RIPE NCC (Europe, the former Soviet Union and the Middle East). The IANA list shows the status of all 256 blocks of 16777216 addresses identified by the first 8-bit number in the IPv4 address. http://www.bgpexpert.com/ianaglobalpool.php is a graphical representation of the IANA global pool (updated weekly). The RIR data indicates how much address space the RIRs have delegated to internet service providers (and sometimes end-users). The changes over the course of 2009 are as follows: Delegated Blocks +/- 2009 Addresses Used Available to/status (in millions) AfriNIC 2 33.55 14.89 18.66 APNIC 34 +4 570.42 540.36 30.06 ARIN 31 520.09 486.58 33.51 LACNIC 6 100.66 79.77 20.89 RIPE NCC 30 +4 503.32 450.11 53.21 RIRs subtotal 103 +8 1728.05 1571.71 156.34 LEGACY 92 1543.50 1413.88 129.62 UNALLOCATED 26 -8 436.21 436.21 Totals 221 3707.76 2985.59 722.17 The RIRs requested an unusually small number of /8s from IANA: only eight. As a result, APNIC is well below the nine months working inventory threshold, so it should be getting no less than six additional /8s soon to get back to 18 months working inventory. Similarly, ARIN should be getting five additional /8s soon. This would bring us to 15 /8s remaining in the IANA global pool, and should allow for regular operation for about the rest of the year. Then in early 2011, the next round of delegations will have to happen, which may or may not hit the magic fifth-to-last /8, after which the remaining four will be given to the other four RIRs and then each RIR will run out of IPv4 space at its own pace. The total number of available addresses is slightly higher than the previously mentioned figure at 3707.76 million because the table above includes 172.16.0.0/12 and 192.168.0.0/16, which are set aside for private use. Networks 0.0.0.0/8 and 127.0.0.0/8 aren't usable because of special uses and 10.0.0.0/8 is also set aside for private use. 224 - 239 are multicast addresses, and 240 - 255 is class E, which is "reserved for future use". The 2985 million addresses currently in use aren't very evenly distributed over the countries in the world. The current top 15 is: 2010-01-01 2009-01-01 increase Country 1 - US 1495.13 M 1458.21 M 2.3% United States 2 - CN 232.45 M 181.80 M 27.9% China 3 - JP 177.15 M 151.56 M 16.9% Japan 4 - EU 149.48 M 120.29 M 24.3% Multi-country in Europe 5 (6) DE 86.51 M 81.75 M 5.8% Germany 6 (9) KR 77.77 M 66.82 M 16.4% Korea 7 - CA 76.96 M 74.49 M 3.3% Canada 8 - FR 75.54 M 68.04 M 11.0% France 9 (5) GB 74.18 M 86.31 M -14.1% United Kingdom 10 - AU 39.77 M 36.26 M 9.7% Australia 11 - BR 33.95 M 29.75 M 14.1% Brazil 12 - IT 33.50 M 29.64 M 13.0% Italy 13 (14) RU 28.47 M 23.18 M 22.8% Russia 14 (13) TW 27.10 M 24.01 M 12.9% Taiwan 15 (17) NL 22.84 M 21.67 M 5.4% Netherlands The reduction in address use by the UK is because net 51.0.0.0/8 is now registered as country "EU" rather than "GB". In 2008, the United States was the biggest user of new address space with 50 million new addresses put into use. China was second with 46.50 million. In 2009, they swapped places: China used up 50.65 million addresses and the US a mere 36.92 million. The US now holds 50.1% of the IPv4 address space in use, down from 52.4% last year. The total for the top 15 excluding the US is now 38%, up from 35.8%. The rest of the world gets the remaining 12%, up from 11.8%. The size of address blocks given out was increasing until 2005, and then started decreasing. The table below shows the number of delegations for a certain range of block sizes (equal or higher than the first, lower than the second value). 2003 2004 2005 2006 2007 2008 2009 < 1000 745 1022 1309 1507 1830 1896 1747 1000 - 8000 1009 1516 1891 2265 2839 3235 3185 8000 - 64k 1014 1100 1039 1192 1015 1129 1169 64k - 500k 215 404 309 419 395 410 403 500k - 2M 46 61 60 57 62 82 70 > 2M 6 7 18 17 24 18 21 In millions of addresses: 2003 2004 2005 2006 2007 2008 2009 < 1000 0.25 0.35 0.44 0.51 0.63 0.65 0.59 1000 - 8000 3.45 4.49 5.07 5.83 6.93 7.75 7.55 8000 - 64k 14.00 15.99 15.46 18.01 15.67 17.40 18.01 64k - 500k 25.51 42.01 34.23 50.86 50.83 52.58 50.50 500k - 2M 31.98 44.63 41.63 46.69 45.50 57.41 49.28 > 2M 12.58 20.97 68.62 52.43 67.37 54.00 54.12 Numbers of delegations and their average size: Year Blocks Addresses (M) Average block size 2000 2794 78.35 28043 2001 2824 88.95 31497 2002 2463 68.93 27985 2003 3035 87.77 28921 2004 4110 128.45 31252 2005 4626 165.45 35765 2006 5457 174.32 31945 2007 6165 186.92 30320 2008 6770 189.79 28035 2009 6595 180.06 27302 (The numbers of addresses given out are lower here because ARIN often attributes the delegation of new addresses to a previous year, this view doesn't correct for that.)

1 0