Transcript

Disclaimer

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.

Tuesday 25 August 1100-1230

IZUMI OKUTANI: So welcome everyone to NIR SIG. Um, well, this is actually, no, a very small cosy significant, you know, exchanging information about NIRs' activities in general. So, as you can see, most of the attendees are usually NIRs, APNIC staff. But then, of course, it's also open to non-NIRs who are interested in our activities.

We have four presentations up for today. Which, the first one will be Miwa from APNIC. She will make a presentation about how APNIC will introduce about APNIC's IPv6 program and discuss how we can collaborate with APNIC as well as NIRs on these activities.

And then the next one will be an update from TWNIC.

The third, JPNIC - myself - will share IPv4 address transfer review currently done by JPNIC.

And then the last two will be based on technical information. One is from CNNIC that explains about IPv6 seen from DNS perspective. Dr Wang from CNNIC Labs will be making a presentation on this.

And the last will be Ji-Young from KRNIC. Is she will be sharing about the incident that happened in July on DDoS attack in Korea.

So that's it and before we make on to, you know, make presentations from each of the speakers, this time we have chair and co-chair election.

So I would like to hand over the microphone to my co-chair, Ching-Heng Ku from TWNIC.

CHING-HENG KU: Good morning, everyone. I am Ching-Heng.

The first item is the chair and the co-chair elections.

The currently nominations for chair is Izumi Okutani-san. We welcome her to have some speech to introduce her for the chair operation.

IZUMI OKUTANI: My name is Izumi Okutani from JPNIC. My responsibility is policy liaison and what I do is collaborate policy discussions within our own community. We also run very similar forum within Japan to discuss about policy with members and then collaborate our discussions with the APNIC forum.

And I've been engaged in this community since year 2000, so this is my tenth year. And I feel that we have a lot of issues we have in common, not just between APNIC and NIRs, but as a community as a whole we are facing exhaustion of IPv4 address, how we should deploy IPv6, co-existence of v4/v6 and moving on to 4-byte ASN.

And I actually had a chance to attend TWNIC's OPM this year. Thanks to TWNIC for inviting me and I really, you know, confirmed my feeling, that you know, we really share the same issues and instead of individually, um, discussing what we should do, I think we can really make a better contribution by sharing, you know, what each NIRs are, you know, considering to address these kind of issues, so I would like to make use of this forum to exchange this information and of course, we don't always have to, you know, only exchange information in public forum. We can, you know, offline, informally continue to exchange information.

So I hope that NIR significant can be like a queue on knowing what the other NIRs are doing and then, you know, ask further questions and exchange information regularly on a regular basis based on what we share over here.

And that's what I would like to do in my next term and I would like to request for your support for my position for NIR SIG chair. Thank you.

***APPLAUSE***

CHING-HENG KU: And so we have some vote to the SIG chair.

Do you support Izumi Okutani-san to be the NIR SIG chair? Please raise your hand.

(Votes are counted)

OK, thank you.

Do you disagree Izumi to be the NIR SIG chair? Please raise your hand.

(Votes are counted)

OK, so thanks, everyone. So we confirm Izumi to be the next term of the NIR SIG chair. Let's give her applause.

***APPLAUSE***

And next is the co-chair election. The current nomination for the co-chair is the Wei Zhao. So we welcome her to have some speech. Welcome.

WEI ZHAO: Thank you, everyone. It's Wendy Zhao from CNNIC for those who may not know me.

I was asked to make a short speech. I will keep this speech really short and save the time for the valuable presentations later on.

I actually am not going to go through my biography because my profile is on the website already. I'm not going to re-cover that. I would like to take a minute to clarify why intending to volunteer in this position.

We, as NIR communities, Izumi-san and Ching-Heng and earlier chair Billy Cheon have done a really great job sharing informations, providing informations, exchanging ideas in these specific communities but we are facing the changes. We moving to the next-generation Internet. So I think, actually, we are ready to take further activities together with other NIRs. I'm sure Miwa has a bit more about that later on for, you know, her presentation.

So I would like to be part of it, be part of taking actions, activities, to let all NIRs to contribute more to, um, lead people or helping people to go smoothly to transfer from v4 to v6. So actually that's my intention to volunteer in this position.

Thank you.

***APPLAUSE***.

CHING-HENG KU: Thank you, Wei Zhao. So currently we have some vote to the NIR SIG co-chair.

Do you support Wei Zhao to be the NIR SIG co-chair? Please raise your hand.

(Votes are counted)

Thank you. Do you disagree Wei Zhao to be the NIR SIG co-chair? Please raise your hand.

(Votes are counted)

OK, thank you. So we confirm Wei Zhao to be the next term of the NIR SIG co-chair. Thank you.

***APPLAUSE***

So the chair to be hand back to Izumi.

IZUMI OKUTANI: So, we'll move on to the actual presentation, and the first one will be from Miwa from APNIC, and she'll be presenting about the IPv6 program in APNIC and how we can collaborate together with NIRs.

MIWA FUJII: Good morning, everybody. My name is Miwa, Miwa Fujii from APNIC. I work at APNIC as IPv6 program manager. First of all, thank you very much for this NIR SIG to provide me the 20 minute slot to talk about collaboration among the NIRs and APNIC IPv6 program.

Here is the overview of my presentation. I would like to briefly introduce about the APNIC IPv6 program by talking about objectives and actives of this role. And I would like to briefly introduce about one of the most important projects we are currently pursuing as APNIC IPv6 program, which is IANA's final 10% project by introducing the background plans and impact analysis. And finally, I would like to talk about working together with NIR and I would like to hear your opinion too.

I would like to propose some of my ideas but I would like to have a good discussion with all of you in here. All right, so the APNIC IPv6 program: It was established in August 2008 by the APNIC member services outcome. APNIC stakeholders recommend to APNIC to expand this activity scope. Increase the outreach activities, not only for the network operators, but other multi-stakeholders. As you imagine, and as you know, the Internet is... the Internet operates within the environment with reasonably complicated multi-stakeholders.

OK, can you hear now?

So the APNIC IPv6 program was established in August 2008 to reflect the APNIC member services outcome. APNIC, and most likely NIR traditional audiences are network operators, but the stakeholders in the APNIC region recommended to APNIC to expand the scope of our outreach activities. Including the different layer of stakeholders, such as Government layers, content providers, application layers, end-users and vendors and so on. So the IPv6 program is established based the member survey outcome. We're trying to provide tangible support to all stakeholders in this region regarding v4 address exhaustion and v6 adoption. When we established the new role, we talked about within APNIC, we talked about several strategic aspects.

We defined through this program activity, we need to gather data and monitor technical developments and share information tailored to the individual stakeholder's requirements, and collaborate with regional and national organizations through the strategic alliances.

Based on the strategic points, so far APNIC IPv6 program has been very, very active. We represented APNIC's view on v4 address exhaustion and v6 adoption at various conferences and trade shows. Some of the conferences are listed in here. Global IPv6 Summit. Thailand, China and Korea this year. And we try to outreach to the APEC TEL economic corporations. The TEL is the telecommunications working group. We participated at 39 in April in Singapore and we continue our effort to involve this particular governmental group in the next conference, TEL 40 actually happening at the end of next month to provide, to participate and proactively be involved with the IPv6 workshop. And we also presented at various NOGs, network operator's groups and NIRs OPMs, open policy meetings.

We also set up a booth at CommunicAsia in 2009. This was the first approach to set up a booth, such a general ICT exhibition and conference. We delivered the information through the booth that we set up at CommunicAsia, but simultaneously, we visited other vendors who put their booth there, so we could actually outreach to the right audiences from different layers of the Internet.

We organized the round table meetings with Government people like the Singapore Government, the Indonesian Government and the Hong Kong Government. We are trying to disseminate information through www.APNIC.net/IPv6. And www.icons.APNIC.net/IPv6. And we developed an IPv6 brochure for multi-stakeholders and you may find the brochure in your APNIC conference bag.

We are quite happy if you take a look and if you think this is useful in your community, in your economy, to use maybe through translation, feel free to use any content that we published in here and please feel free to link those to URLs as well. We are very happy to share our effort. There is no reason to reinvent the wheel. If there is anything available, readily available for your effort of information dissemination regarding v4 and v6, please use them.

OK, so I'd like to briefly explain about one of the most important projects we are currently working on, which is IANA final 10% project. As you all know, the remaining IPv4/8s at IANA is 28, as of today. If you do a simple calculation, the 10% point of 256/8s is 25.6. So we think that if the remaining goes to 25, then we can tell that that is below 10%. The reason why we focus on 10% so much is that 10 is such a nice round number. It may be easy to get the media's attention and we should revel on such media attention and use that media attention positively. So, this is the chart showing the IPv4 allocations from IANA to the regional Internet registries year by year. So, with using the previous pie chart and this bar chart, what we can tell is that the RIRs allocated 13 /8s in 2007.

9 /8 it's in 2008, and there are 28 /8s remaining as of today, basically. And APNIC received 2 /8s this month, and other RIRs can request more /8s at any time, and this means that the remaining IPv4 pool at IANA may hit 10% sometime very soon.

So many use the new source in a negative way and create fears and anxieties amongst people. We would like not to be caught in that kind of negative flow. We want to ride on that wave positively, and take advantage of the 10% event. Be proactive in disseminating information to support IPv6 adoption. We would like to use this occasion to sell the idea of the necessity of IPv6 adoption. Selling and marketing and those terms are there for the regional Internet registries and NIRs for many years, but I think we reached the point where we need to modify our activities and actions, like Wendy mentioned in her speech. A more productivity to sell and market the idea of the importance of adopting IPv6.

In this project, I set up a couple of strategic points. I would like to collaborate with the NIRs like Izumi and Wendy mentioned in their speech. We are going through the changing period and we can collaborate together and maximize our effectiveness to communicate with the region. I would like to establish a point 2 point communication link with each NIR member. I would like to co-ordinate activities with other RIRs and possibly this can be done through the collaboration with NROs, Number Research Organizations. There are some people were ARIN here. Obviously every region has a very big media such as CNN, ABC, and hard copy newspapers like the 'New York Times' or 'International Herald Tribune'. Once the major media catches the major information and starts publicizing the 10% IANA remaining IPv4, it could create a big impact, a big wave throughout the globe.

We are not ready, we need to be able to handle the impact to be generated by major international media corporations. We need to be ready for that. We can not be reactive. We need to plan now how to handle that. To do so, I would like to have a close communication with the other RIRs through the NRO and probably co-ordinate a joint announcement if possible, and we would like to utilize the public relations agent to maximize information dissemination.

So the possible impact of this 10% PR campaign may lead to some enquiries to APNIC and NIRs, it may increase. Some of the questions could be, what do we need to do now? Can you please reserve IPv4 address blocks for my company now? Or can APNIC or the NIRs help us? What will happen after the exhaustion? Or some of them may ask, why have you not done anything? Although we have done so many things up until now, people want to find a blaming point, so they may come up and say, "Why have you not done anything?" We need to be able to handle the negative questions and we need to be able to provided pro-vied, APNIC and the NIRs together need to provide a clear, focused and unified messages to the community to avoid any unnecessary confusions.

At the right timing, the right message needs to be delivered. And we need to channel this anxiety or fear of the community into the positive energy so that they can adopt IPv6 more smoothly.

So, working together with the NIRs... I want to propose that through this presentation. As I mentioned earlier, I'd like to establish P 2 P links between the APNIC IPv6 program and the NIRs. So, at the end of this session today, I want to have at least one business card from each of the NIRs, and that person will become my contact point, probably our contact point. Probably the mailing list, setting up a mailing list could be a good starting point, but I am quite open to hear your opinions, your thoughts as well. You may have better ideas as to how we can achieve this collaboration. I have a discussion time at the end of this presentation and I would like to hear your opinions.

And before I go into the next slide, I would just like to share the APNIC IPv6 program stance, because that would create the important foundation with which we are standing on in the future. APNIC - it is for a nicks IPv6 program. It is not the objective of APNIC to make operational decisions for our members and other stakeholders. We want to be very careful about this. We will talk about IPv6 a lot. But we will choose carefully what words we're going to use. We rarely use the word "promote". Promote is like promote for the sake of IPv6, and it doesn't make sense to have that. The final decision has to be in the operator's hands and other stakeholder decisions. But we need to provide a required support to make the right decisions as much as possible, so we focus on this point.

And we need to tread very carefully too and increased awareness is important, but we should not create unnecessary panic or anxiety. Especially among the non-technical stakeholders. How to distribute information is important? What sort of tone should we use? What sort of material would be useful? We need to be unified and synchronized in our effort at this time. So based on that, the basic stance, I would like to share the APNIC IPv6 quarterly objectives in the future if you agree upon establishing some sort of mechanism to support each other and to collaborate amongst the NIRs and APNIC.

For example, for this quarter's IPv6 program objectives, it is provide maximum effort to develop high quality IPv6 workshop content to be delivered at APEC TEL 40 happening next month in Mexico, Cancun. And also to maintain the relationship we have so far developed with regional bodies and national bodies, and we would like to expand our scope to the content providers as well. So those three are important quarterly objectives and I will come up with the next quarter's objectives and I will share them with you, and I would like to talk about the quarter's objectives with IPv4 exhaustion and IPv6 adoption. And I would like to maintain close communication with NIRs so that we can provide a synchronized message to the APNIC community.

And APNIC will provide some message in regard to messages, content and PR and so on. We will happily share, openly and happily share this information to all NIR members, and NIRs can save resources for developing PR messages by utilizing or recycling the available messages developed by APNIC, and I also would like to know any specific information which is more suitable for your particular economy that would help our understanding as well, and we can customise our messages as well when we go to your region. I would like to organize a round table meeting with relevant NIR staff and also board members too to share the strategies and activities of the APNIC IPv6 program, and also I would like to learn from your NIR's activities as well.

So, basically, that is my presentation proposal. And now, I would like to... if it is OK with Izumi-San, I would like to open the floor and hear other people's opinions. Thank you.

IZUMI OKUTANI: So, do you have any questions to Miwa on APNIC's activities or possible ways you feel like you can collaborate? I'm sure more details will be discussed individually, so people can talk and come up to you and let you know the contact point.

MIWA FUJII: That's the minimum point I want to achieve today. I would like to get at least one business card from each NIR, and I hope we can establish the mailing list based on that information. Maybe you want to assign someone else. That's OK too. But I would like to have a particular contact point so that I don't lose track.

IZUMI OKUTANI: Any questions to Miwa?

SPEAKER FROM THE FLOOR: Thank you for your presentation. Just a comment - we already know the collaboration of the IPv6 deployment with the information by another RIR such as an example for the IPv6 Act provided by RIPE. Do you have any idea for collaboration with another RIR?

MIWA FUJII: As I mentioned in my presentation, it's probably such collaboration can be achieved through the NRO. The Number Research Organization. That's representing all of the RIRs and particularly like when we run the 10% event public relations event, we need to unify messages amongst RIRs too, so yes, we are on that track of making some possible way to collaborate. Thank you.

IZUMI OKUTANI: So, thank you Miwa, and as Miwa said, I think it is really important that the NIR economy also keeps consistent levels of awareness as much as APNIC. So please make sure to talk to Miwa after this SIG.

***APPLAUSE***

IZUMI OKUTANI: So the next speaker would be Sheng-Wei from TWNIC... TWNIC's activities and...

SHENG-WEI KUO: Good morning, everyone. My name is Sheng-Wei Kuo. Today I will introduce TWNIC's update.

This is my outline today. First, I will introduce the status of 4-byte AS numbers in Taiwan. Second, I will introduce the IPv4 address exhaustion and IPv6 adoption. In this topic, it will include two items. One is the project of the measurement of the IPv6 readiness in Taiwan. Another is the IPv4 address exhaustion solutions survey.

First, I will talk about the status of 4-byte AS numbers in Taiwan. You know the timetable for moving from 2-byte only AS numbers to 4-byte AS numbers is as follows:

1st July 2009, APNIC assigns 4-byte AS numbers by default. APNIC assigns 2-byte AS numbers if a 4-byte AS number is demonstrated to be unsuitable.

1st January 2010 - APNIC ceases to make any distinction between 2- and 4-byte AS numbers. APNIC assigns from an undifferentiated 4-byte AS number pool.

So we will - so we had requested a block of 4-byte AS numbers in 2008 and assigned one 4-byte AS number. We also announced formal letters to ISPs about AS number policy adjustment. Since 2006, ISPs discussed 4-byte ASNs on 6th, 8th, 11th and 12th TWNIC IP Open Policy Meeting.

In TWNIC's 11th Open Policy Meeting, TWNOG provided '4-byte AS number migration suggestion' report. For more information, you can see this website.

In TWNIC's 12th Open Policy Meeting, APOL shared experience of ISP migration 4-byte AS number.

TWNIC invited major ISP and router vendors such as Cisco, Juniper and extreme to discuss the migration issue of 4-byte AS number.

So we will hold 4-byte AS number workshop in October 2009. It is a hands-on workshop. Cisco and Juniper will provide lab to support this workshop. We expect that this workshop will help ISPs Migrate to 4-byte AS numbers.

And this is our 4-byte AS number workshop agenda.

We will provide - to include 4-byte AS number peering and trenching and a schedule to include 2-byte and 4-byte AS number peering and trenching. And we will talk about the relevance. And we will talk about 4-byte AS numbers from a BGP neighbour and the BGP configuration.

Now I will introduce another topic - IPv4 address exhaustion and IPv6 adoption in Taiwan.

In order to measure the status of IPv6 deployment, we will do the project of measurement of the IPv6 readiness in Taiwan.

This project will define the IPv6 metrics set as the measurement of the IPv6 readiness. It will establish the method of analysing data using measurement. And it will publish the result of the measurement.

There are six classifications. One is the address allocation. In this item, we will check IPv4/IPv6 address advertisement in the BGP routing table.

In the section of the DNS query analysis, we will compare the amount of IPv4/IPv6 connections and in particular the distribution of DNS query by resource record type.

In the DNS deployment, we will look at the deployment rate of DNS server on the base of whole .tw domain name and in particular the deployment rate of mail server on the base of whole .tw domain name and the deployment rate of web server on the base of whole .tw domain name.

In the web server access, we will take a look at the total amount of IPv4/IPv6 traffic from/to web server. We chose the www.ipv6.org.tw.

In the section of IPv6 traffic, we will choose the major IPv6 ISPs in Taiwan. We will take a look at the total amount of IPv6 traffic from/to ASIX, in particular the total amount of IPv6 tunnel broker traffic. Because, you know, TWNIC and five major ISPs had established the IPv6 tunnel broker, so we will take a look at the amount of IPv6 tunnel broker traffic.

In the section of IPv6-ready products, we will take a look at the number of products certified by IPv6-ready lo Logo Program phase 1 and phase 2 certification program operated by IPv6 Forum.

We will finish the measurement in December 2009.

In order to understand the level of awareness about IPv4 address exhaustion and the measures of TWNIC's members are considering, we will do the project of IPv4 address exhaustion solutions survey. The survey period is from 24th August to 15th September 2009.

In these survey questions, there are 11 survey questions, including the requirement of IPv4 address before IPv4 address exhaustion, facing IPv4 address exhaustion, what are their plans, strategies? If ISP adopts IPv6, how much expense and what are the bottlenecks of the IPv6 deployment?

That's it. Are there any questions or suggestions?

TERRY MANDERSON: Hi. Terry Manderson. How many TWNIC members do not have IPv6 allocations? How many TWNIC members do not have IPv6 allocations that potentially should have IPv6 allocations?

How many TWNIC members do not have IPv6?

OK, so about 30-something? About half? Thank you.

SPEAKER FROM THE FLOOR: I was quite interested that you have allocated, sorry, assigned one 4-byte AS number. Do you have more information about that and which organizations or ISPs you have signed on to have the test or use that formally? And do they have feedback for you?

(Pause)

Alright, OK. Thank you. So those organizations, they are using the 4-byte?

SPEAKER FROM THE FLOOR: Planning to, um, publish the result of IPv6 measurement, you're going to finish in December 2009. Through your website? Would it be available in English too?

(Pause)

IZUMI OKUTANI: So you're saying you do plan to publish the Chinese version but not sure about the English version at this stage? Maybe. Oh, OK.

Maybe we can use Google Translation.

OK, yes, I think a lot of the activities are very similar to what we do and, you know, lots of people, you know, ex- - express interest. I think it's something of common interest to everyone here. And we actually conducted a similar survey in JPNIC about IPv6 deployment as well and then our communities working on ASN.

So I really hope that we don't just end exchange of information here and after the meeting, maybe I don't know if it's better to do it, like, on a public mailing list or individually contact NIRs, but, um, I think we can, you know, like, based on your presentation, each NIR can input, "OK, in this area, we also do this," or, you know, "You can get this information through here." I think it would be useful to, you know, do something like that.

So, as an action item, I will, you know, put it for NIR SIG, based on your presentation, each NIR will give input on their activity on this area of 4-byte ASN and exhaustion of IPv4 and v6 deployment.

Thank you very much for a very useful presentation.

***APPLAUSE***

I will hand over the chairing to Ching-Heng from here, because I will be speaking next.

CHING-HENG KU: So the next item of presentation is by Izumi about JPNIC's review of IPv4 address transfer.

Let's welcome Izumi.

***APPLAUSE***

IZUMI OKUTANI: So I'll be putting JPNIC hat on now and, um, just to... well, I'm making this presentation because JPNIC has expressed quite a bit of concerns over implementing the transfer proposal and people might be wondering, like, what exactly is JPNIC's concern and, you know, probably don't have a really concrete idea.

So I'd like to share what they were and then the kind of reviews we've conducted based on these and then hopefully maybe some of it is a little bit too much in detail for you, because Japanese tend to go into details. But then I hope that there's something in common that we could share, you know, when you try to deploy your own transfer policy in your economy because some kind of, like, legal checks that we did or some issues we found might be of an interest to you..

So I'm sure you're very much aware of our situation about the transfer proposal, but a very quick summary - we've actually been discussing about the transfer for two years since September 2007. And then initially people actually raised concerns over it but gradually people started to be supporting the idea itself.

And in Manila we did reach consensus at the meeting itself and it's just that we had a slight disagreement over the specific conditions on the transfer so the conclusion at the moment is continue mailing list at the moment, which will be proposed on Thursday, and discussed.

This is just like more days so you can take a look at it. I wouldn't really explain to it in details so please take a look at it in the slides later if you're interested in how, you know, each meeting, the community's opinion has changed, both in APNIC's forum and JPNIC's forum.

And actually, JPNIC's position in situation of regarding this proposal is we've acknowledged that there's wide community support over this proposal in the region as a whole as well as support from ISPs in Japan.

And we will be actually proposing this in our own OPM in November based on our own policy development process to consider how we should implement it in Japan.

But, as expressed in Manila, JPNIC still feels that before we implement it within Japan, it's really important to ensure what kind of impact this transfer proposal will have on outside address management. We should review those and at the same time ensure there's no large confusion, concerns, before we are able to provide service to a community. That is our basic idea. So we're working on the issues we think we have over the transfer proposal and then, um, doing analysis on what would be the possible issues and what are the measures we can take.

And four major questions we have.

The first one - it's very difficult to do an accurate speculation of course because we never know what's going to happen in the future. But we want to have a very brief broad idea of how is transfer going to happen. Is it really large and everywhere or just limited to some type of address space? How long is it likely to last? How are people likely to transfer address space? These kinds of questions. That's the first question that we have.

And then if we have any concerns over the confusion of the transfer process, is there any action we can take in advance to prevent that kind of thing?

The second point is that since transfer of address space would imply monetary transactions, do we possibly need a new legal framework within Japan or maybe the APNIC region to address such cases as frauds, disputes, you know, if these concerns are serious?

The third question is that it's more of a procedural concern - like how do we draw the lines of responsibility between the actual transfer transactions and the registration, especially once the transfer is completed and APNIC register or each NIR registers the transfer record? And then people start, you know, disagreeing. You know, "We did register this information, but then, you know, I didn't really intend to, you know, give this address space." How do we handle these cases? I think that's something that we should be, you know, prepared for in advance. And I think in APNIC's proposal inter-registry transfer is also included, and especially with our region, it's very likely that transfer between NIR's member and APNIC's member might happen so how should we be prepared for this? Are we going to have, like, customs duty or, like, how do we split this responsibility between registries? I think these are things that should also be considered.

And the last thing is that since IP address implies to have monetary value, what will be the impact on taxation or finance? Not just on registry's side but on ISPs who actually transfer space? I think it's very likely that, you know, it will be taxed if you actually sell the space but then, could it be possible that you could be taxed for just holding the space? That's the kind of questions that we have.

And the current review status is that we did, like - we tried to, you know, get opinions from experts in different areas. For example, simulation of possible methods of address transfers. We felt that ISPs will know what they would like to - how they would like to transfer space the best. So we did round-table discussions with ISPs and we also heard experts - well, I don't know if you can call it expert, like, we also did, like, a review with an economist on Howe it's likely to happen from economic perspective.

And in terms of legal framework or procedures concerns, we did some hearing from lawyers and in addition in domain name, actually transfer is allowed in reality so we actually heard from a DRP dispute resolution policy expert, a specialist on handling issues when they have disagreement over, um, the right holder of a specific domain name. We did a hearing about this.

We also had discussion - APNIC arranged a teleconference with APNIC and NIRs and we discussed about possible procedures.

And we tried to hear details about taxation and finance issues from a contact chartered accountant. I think they have a different name in the US. I can't remember it very well but an expert in accounting anyways. And we also shared the status with the government to make sure that they don't have concerns from a social perspective or possibly from, like, Internet governance perspective.

And, um, so I'd like to briefly share how our ISPs in Japan speculate to transfer.

First, um, they think that, well, obviously no strong motivation to undertake transfer before APNIC pool runs out because people can simply get more space from APNIC so they don't need to buy and purchase it from another organization. And there was a feedback that maybe if we start charging registries, start charging historical resources, it might help in, um, you know, maybe some organization historical address space, return address space to APNIC or transfer space to another organization. So it might help in fluidity of unused historical space.

I don't know if it's applicable to anywhere in the world. People don't feel very comfortable about buying address space by auction if you're a company. If you're a person, then maybe OK but if you're a company you want to buy it from someone you feel you can trust.

So they think that they're likely to purchase address space from, like, a company that they regularly do business with. For example, system integrators or consulting firms. So, as a part of, like, consulting the network construction or maintenance, they, you know, talk to - like, what can we do about address space and then these people will try to find, you know, unused space from another organization and get hold of space.

They think that's the kind of thing that's likely to happen.

In Japan we feel there is probably no strong motivation to run an open market such as like an IP address version of a stock market because it's only going to last for a few years so no strong motivation for market operators to, you know, actually set up this kind of market and, you know, they can't really make enough money to recover the cost. So they think that - I don't think we will have, like, public kind of market, formal market to address this kind of thing, unless registry, you know, proactively try to provide or operate this kind of thing. I don't think private organizations - we don't think private organizations would try to run this kind of thing.

And, again, not too surprisingly, the fifth point is that most of the transfer address space is likely to be from historical range, not from what we call portable allocations - that's allocated to our LIRs and members, because we think most of this space is already efficiently utilized. APNIC or NIRs check, you know, the efficient utilization, so there's not much available space in the first place and ISPs are not likely to be willing to give out space, so those organizations, maybe universities or companies that have additional space, might be willing to give out space if ISPs pay them. So that's the kind of thing that we expect.

So the target of address range to be transferred is probably quite limited.

Last lastly, transfer will just be one of the options. Not all ISPs will do the transfer. They will do cost analysis and pal with other possible technologies such as NAT or deployment of IPv6 and if they think transfer is a better and more cost effective method of providing service to the customer, they will transfer address space but if there are better cheaper ways to make this happen, they don't necessarily do the transfer.

So as a result we got the impression that, you know, not everyone will do the transfer and it will happen just a few years after APNIC pool runs out. The target will be limited to historical range and so maybe the scale is not very big. That's what the feeling we get at the moment. So we don't really speculate like large confusion as a result of, um, you know, um, APNIC pool running out.

So this is just an analysis at this stage, so doing a little bit more check to confirm if our - what do you call it? - our speculation is correct.

And then our other interim review result is that, yes, as mentioned, it's likely to be - transfer transactions are likely to be limited in time frame, scale and method of transactions. And we confirm there's no legal framework required for this because it's not different. IP address is not different in nature from other commercial transactions and it's very close in case of Japan to credit obligations, such as, you know, providing - renting land or, like, membership for a tennis club or something like that. It's very similar in nature. So it can be handled within the law framework for transferring these kinds of things. I'm sure it's different in each country with this second point.

And we also haven't found a major issue in handling disputes for transfer transactions and registration. We think the biggest concern we should take care of is duplicate transfer, i.e. transferring the same address space to multiple organization so you'd give a particular space to ISP A and give the same space to ISP B. That's the strongest concern that registries should have to be involved in dispute. And this can be prevented by APNIC's defined procedure - making sure you get the document from both the source of the transfer and the recipient. So this could be avoided and if disputes happen, even with documentation and afterwards, we can say that we actually requested proper documentation and if you, you know, resubmit the decent documentation, we'll simply update the database, so the problem can be handled this way, at least in the case of Japan.

And if we think rationally, it probably should be more or less the same in other parts of the world too. And we actually have the two remaining issues. One is how do we handle leasing? It's actually likely that rather than selling address space, some ISPs are likely to lease it for a limited amount of time, for example, maybe for three years and, in that case, do we actually register this information about leasing? Or do we simply not. And if we don't register at all, it doesn't really meet the original intention of the transfer proposal maintaining the actual use of that space. But then, if we actually simply update the leased user as the address-holder, it will cause condition fusion about who the authentic holder is. So I think we should clearly define a policy about this.

And I will explain what we have in my next slide.

And another issue that we have is for inter-registry transfer, we should clearly define responsibility on which registry is responsible to which part, so it's registry that handles the source of the address transfer should take full responsibility on the allocation part and then the registry that - the resilient - sorry, it's the other way round - the return part and then the registry that handles the recipient part should take full responsibility on allocation procedure and it's very important that we share what our responsibilities are and agree about it to make sure we don't reach disagreement after we have problems with the transfer.

So the summary of the current status is that, yes, we need to fix the policy for leasing and we also need to have a clearly defined procedure for Internet registry transfer for smooth transfer between registry.

And regarding leasing, at this stage, JPNIC think maybe we want to create specific information specialized for leasing that's separate from allocation but then, it's, I think we want to make it consistent with APNIC as much as possible so we want to keep, um, you know information exchange on what APNIC has in mind about the policy for leasing. And same for Internet registry transfer. We want to keep close contact about what we do on this.

And in future we'll try to also check the situation on domain name and then, I think, our review will more or less be completed.

So a lot of the areas we think that we share common issues, we hope to exchange information with APNIC as well as other, um, NIRs, especially in the two points I mentioned - leasing and Internet registry transfer.

We'd also love to hear any idea you have about how to handle taxation issues in your economies. In our case, we tried to hear opinion from experts but they don't really want to say. They think it's really difficult to say and speculate. So if there's any useful way of doing analysis, that would be very welcome. And JPNIC is also happy to share any information we have about our review. So if you're interested, feel free to contact me at this e-mail address I've written.

Thank you very much.

***APPLAUSE***

CHING-HENG KU: Thank you, Izumi. Izumi's presentation gives us many very important discussion topics and a very important point of view of Japan in transfer proposal. Do you have any questions based on some topics or issues in this presentation?

SPEAKER FROM THE FLOOR: Actually, it's not comments or questions. It's clarification. As Izumi has a point in our presentation talking about historical transfers. And CNNIC has a proposal to supplement current historical transfer policy so if anyone in this room - it doesn't matter if you're from a NIR or doesn't - if anyone has a concern or comment, we would like to hear from you and we might go through with the review of it or get it better done of these proposals. Thanks.

SPEAKER FROM THE FLOOR: I'm from TWNIC. You said that you expect the transfer would happen to historical address mainly. Is that before the final /8 or after? Also do you encourage transfer or not expect it to be happened or how it will happen very frequently? Or you will try to encourage those policies?

IZUMI OKUTANI: So, regarding your first question about whether historical address transfer is likely to happen before or after, I think it's pretty much like right after the exhaustion of APNIC pool runs out, maybe start to have gradually a little bit before, you know, when people start to expect - OK, maybe we'd better be prepared so have a little bit of transfer before APNIC pool runs out but I think the majority will be right after APNIC pool runs out.

And then beyond that, like a few years later, it's very difficult to speculate, so we don't really have an idea about it.

And whether JPNIC would try to encourage transfer or not, we are neutral, we don't really, you know - it's just up to the community what they want to do about it and our role is simply to register what's going to happen and keep people aware of, you know, what policies are available. That's all we plan to do.

Does that clarify your question?

SPEAKER FROM THE FLOOR: I'm just thinking that JPNIC has spent a lot of effort on pushing IPv6 and if IPv6 is to be successful, maybe there's no need for a transfer policy. So that's why I'm asking why you're not discouraging a transfer policy.

IZUMI OKUTANI: We actually think that it's - it can be pretty much independent because it doesn't really provide long-term solution. It's just for a few years. And it's - it would just - to help ISPs provide one of the options on how to, you know, address issue a few years after APNIC or RIR pool runs out and then complete transfer to IPv6. So we don't think this will disturb the deployment of IPv6. Although some people have a similar opinion as you in Japan as well.

JOHN CURRAN: John Curran, ARIN. I just had a question about this summary of how ISPs see transfer transactions going on. What version of APNIC policy 50 was under way when this was being asked? Because the companies likely to see addresses through integrators and consulting firms presumes that the transfer can go that way and the current policy doesn't allow that because - unless the consulting firm has a need for it. So can you comment?

IZUMI OKUTANI: It was based on policy proposal in Manila so it's assuming that transfer will be allowed and we want - they share the situation.

SPEAKER FROM THE FLOOR: Um, this is not a question directing to Izumi-san, but I'd like to know if other NIRs having similar discussion within your community or not? Obviously JPNIC's community is planning to talk about this proposal within JPNIC in November this year. Um, how about other NIRs? I'm just curious. Is there any specific, um, plan you have to talk within your own NIR's community?

WENDY ZHAO: Hi, Wendy from CNNIC again. I might put forward information on that. As early time, I'm talking about the largest operators from local. Their comments are totally different with these smaller ISPs. Um, they have their own troubles. They don't see any benefit from getting addresses being transferred. They will bring lots of hassles and a workload for identifying which addresses belong to which and if they are able to be routed or should they not route it and who is going to hand over the logistic things. They have to build up a whole system to support this kind of policy proposal. But considering the smaller ISPs, they are welcome this kind of proposal, this kind of thing happened because they see the transfer - the address transfer as inevitable, so we cannot forbid it or not avoid it, as we cannot get any addresses, IPv4 addresses, from RIRs from that makes either the historical or the - sorry, not only the historical, also the current address, which not be used or not be routed become available.

So by that time, people will start selling the address, leasing address, to make the profit and from the possible point, it's really nice because we actually generate a system to let the address not be used come back to a system, but the other way is we should be ready for how to manage those kinds of transfers. So it's kind of, you know, half-half way. Some of them are - sorry, some are against, have their concerns and worries about that. And the others do in favour of it but they haven't think about the problem from the, um, the big operator's perspective because they are the organizations having more trouble than the small ones. So that's, um, updated information from home.

CHING-HENG KU: Does anyone have some situation or comment about the transfer proposal or about this presentation? No?

OK. Then thanks, Izumi. So give her a... a...

The next topic is presented by CNNIC.

Please welcome Dr Wang Xin.

WANG XIN: Good afternoon. This is Wang Xin from CNNIC. I'm very happy to be here talking about IPv6 study at CNNIC. We did this study from the angle of DNS and I would like to share some interesting results and pictures.

As we know, the DNS is the fundamental Internet service and it allows users to access resources and applications by typing domain names. Therefore, we believe that Internet behaviour could be reflected and perceived by domain name querying behaviour to some extent.

At the same time, CNNIC is a national registry for Bow both TLD domain names and Internet addresses and we have always been interested in the development and deployment of IPv6 resources and applications.

Our study is data driven and we have two sources of data. First, we get queries made to. CN root nameservers in 24 hours and this date is logged automatically by server software.

Second - data collected by actively probing nameservers in IPv6 network and, of course, IPv4 network as well.

OK, now first part, let's see what we get from .CN root query logs.

This picture comparatively illustrates the query rates issued from IPv4 and IPv6 complaints. Please note that the query rates are in the log-log scale and the X axis is time in minutes. From the curves in the blue and the green curves, you can see that, the exhibit similar shapes and trends, therefore indicating that the user behaviour patterns are similar as well. Except from the IPv6 queries, far less than those of IPv4. We guess that this is because of the limited amount of IPv6 addresses allocated and the applications as well.

This is the continent distribution of IPv4 and IPv6 queries. According to the source IP addresses observed in the .CN logs, we calculate the geographic distribution of queries according to their source IP addresses. From this picture, we can see that most of the IPv4 and IPv6 queries are from Asia. Actually more specifically from China. And you can see from the red blocks, the red bars.

However, large proportions of IPv6 addresses observed are from America and from Europe - the green bars and the deep red bars.

So, compared with IPv4, a smaller proportion of IPv6 clients issued a large proportion of queries.

This is the distribution of query types from both sources.

We can see that the queries from IPv4 addresses are more evenly distributed among the types because IPv4 applications such as mail and HTML, such applications contribute much to this. However, for IPv6 queries, they seemed to concentrate on A and AAAA queries.

We guess this is due to the limited type of applications deployed in IPv6 networks. Because users have no other choices.

The distribution of client and client and target query lead. The client query load is the upper plot and the target domain query load is the lower plot. They're all plotted in log-log scale. The busiest client and the busiest domain names are first. From these two plots, we can see that the queries are not evenly distributed among both clients and domain names.

However, the curves for both IPv4 and IPv6 - the shapes look quite similar and we get this is due to the transfer from IPv4 to IPv6. This transfer doesn't cause much unusual behaviour, except from the amounts of query loads is far legs than those of IPv4.

This is the first part of our study. The second part - probing the DNS nameservers in IPv6 network actively, and to get the software installations, the recursion design of settings and other informations. These two slides show the software installation statistics, both authoritative and recursive nameservers.

For authoritative nameservers, ISC BIND 9 amounts to 96% for both IPv6 and IPv4 nameservers. However, for recursive nameservers, the ISC BIND 9 share is IPv6 is much higher.

These two pies, these two pies show the statistics of recursive design and flag status of authoritative nameservers.

IPv6 servers are better configured since we know that this flag should be disability for authoritative nameservers for security purposes. However, 61% is far from satisfactory. We guess this is because currently most of the IPv6 servers are for experimental purposes, so this flag is not carried very much.

Support supported protocols, we can see that, um, IPv6 servers show a larger percentage of supported DNSSEC than IPv4, because IPv6 servers are deployed later and are thus more likely to be prepared for DNSSEC.

Another characteristic is nearly all servers in IPv6 network support TCP. This is because AAAA records are included in DS packets, therefore it's more likely to exceed 512 bytes.

From our study, we can see that IPv6 is growing steadily. However, there's a long way to go for it to be in position.

OK. That's it. Thank you.

***APPLAUSE***

CHING-HENG KU: And thanks, Dr Wang. It's a good observation and study of DNS in IPv6.

GEORGE MICHAELSON: Hello. This is George Michaelson from APNIC. That was a wonderful and very informative presentation. I will be presenting something similar to this in APOPS but from a slightly different perspective of the 'day in the life' exercise, but it was very beautiful to see convergence in some of your slide of the v6/v4 relativity and I think it is good to have two independent measurements and research activities that show this kind of behaviour. So I am very glad to get to see this. Thank you.

CHING-HENG KU: Thanks. Any comments or suggestions?

TAIJI KIMURA: Hi, I'm Taiji from JPNIC. I think this type of study is very helpful for - not only for China but also the other countries, so do you have any plan to public your papers or graphs on your presentation anywhere?

WANG XIN: Those have been published in papers, yeah.

The conference proceedings and journals. But I can give you some electronic versions.

CHING-HENG KU: Maybe we can form a workshop in the future to include you in APNIC in the IPv6 readiness topic. So, because of time issues, we thank Dr Wang Xin for this presentation.

So I will hand over now to Wei Zhao.

WEI ZHAO: OK. So I'm chairing the last session. First of all, I want to make the point that if someone smells the food, that means you're very hungry so you can go. And if someone don't, that means you're OK to stay for the next wonderful presentation brought by KRNIC. Can you call you KRNIC? Or still KISA? Sorry? Alright, yeah, they're doing the corporation constructure.

Well, while Ji-Young is coming up, I will make my comment on her presentation. When Korea was attacked on their system, I was in Seoul for their IPv6 forum, so even though I didn't really go through the pain with them but, as a person from outside the system, from the outside the frame, I can feel the impact.

So I'm glad Ji-Young is going to share their experience and some thoughts about this attack and then we really can discover what's really happened behind the attack, then probably some work can be done to save us for the further one.

OK.

JI-YOUNG LEE: Hello, everyone. My name is Ji-Young and I'm from KRNIC.

My presentation is about July 7th DDoS attack happened in Korea. Even though I'm not a security expert, I'd like to share our experience with other countries.

OK. Um, this is the DDoS attack timeline in Korea. First attack occurred in July 5th and, at that time, targets from websites in United States, for example White House and Department of Homeland Security and other websites in United States. And the second attack occurred on July 7th. At that time, targets were some websites in United States and some sites in Korea. For example, Blue House, which is, um, Korean President's office, and Ministry of National Defense, National Assembly and some local portals of, like, NAVER and some other websites. And that is the day the first attack in Korea occurred, so that's why we call it the July 7th attack.

And third attack occurred the next day. And fourth attack occurred the next day.

This is the comparison of DDoS attack. Usually hackers make malicious code and upload those malicious codes on the website - for example, homepage or P2P sites and when user PCs downloads Michelle codes from the P2P sites or websites, they are infected and compromised and using e-mail or messenger, they can distribute the malicious code to their neighbours. And when a hacker or a C&C servers - a command and control server - sends a realtime command to zombie PCs, zombie PCs access indefinitely to the victim sites. So victim's system is flooded with traffic and the traffic overwhelms the bandwidth or the capacity of the system, so they are, um, they become - they become out-of-service status.

But this time, the attack was a little bit different. The target and the attack schedule are programmed in the malicious code, so there was no communication with hacker or C&C servers. So it made it really difficult to track down the source of attack or the location of hacker.

And some zombie PCs are scheduled to delete the partition data in the hard disk so if you are compromised or if you are targeted as a victim, some information in your system could be deleted.

As you know, KRNIC is not for security, but we have our member ISPs and we have very close relationship with them, so at the time, we participated in these activities. So we collected zombie IP addresses from the victims' sites and we sent them to each ISPs. Currently we have 127 ISPs in Korea. And we, um - and KISA uploaded vaccines in the major Korean portals and game sites and they recommended Internet users to update them. And KRNIC opened our KRNIC Whois to the victim sites to identify the zombie PCs.

And the ISPs - when we called them, we - some of them were already aware of the zombie IP addresses using their IDS, so some of them, um, blocked their access automatically, but many of them, but for many of them, we contacted the subscribers - we let them contact the subscribers and let them update their vaccines and sometimes they disconnected their access when their contact information were not reachable.

And this is the result - this table shows the confirmed number of zombie PCs from major four ISPs in Korea. But these numbers are different from investigators, from investigators. So some say there were 200,000 zombie PCs in Korea and some say 170,000, but according to our investigations, confirmed number is 77 - around 77,000.

And as of July 11th, 97% of PCs, we deleted malicious codes in their zombie PCs and around 3% still remain.

So, from this experience, we learned several things. First of all, it was really difficult to identify the zombie PCs or other command and control servers, especially when they are NATed, it's really hard to track down, because the number of zombies are increasing really fast, so it's really important to find them fast. And we think some ISPs it is better if they distribute vaccines from their websites, so there is a an ISP in Korea who freely distributes vaccines and recommends users to update it when they log in to their network. And one fast way to solve the DDoS attack is to restrict the access of the zombie PCs.

So the source of the attack is not identified and the motivation of the attack is not identified.

So, this problem is not solved, but we should be very careful about, um, other attacks because it could happen to anybody.

Thank you.

WEI ZHAO: We're only five minutes before our schedule but I think we've gone through a really long morning taking lots of informations, very important, helpful informations.

I do think all the presenters, speakers, won't have a problem if anyone has further questions - you can join them with the tables to have further discussions or exchanging of ideas. Even though I don't do eat and talk, but some people do.

OK, this is end of NIR SIG. Thank you for everyone being here with us, sharing their experience within the NIR SIG. Thanks.

***APPLAUSE***