Paul Wilson's speach Good morning. This is the opening plenary. Break time is at 10:30. I'd like to mention some important announcements before we get started. Firstly, thanks to TWNIC for sponsoring the meeting today. That's very generous and important support so thanks very much to TWNIC. We have meeting breaks - morning tea, afternoon tea - which will be outside the meeting rooms. Lunch will be in the crystal ball room number 3. You may note that there's a social event tonight so we will have a river cruise on the Han River. Please meet in the hotel lobby at 7.00 pm to be ready for the social event. APNIC is providing a hot desk with APNIC secretariat staff which is opening during the break. Members or others who would like to have a discussion with APNIC staff or any inquiries at all, please feel free to approach the APNIC staff at the hot desk and they'll do everything they can to assist you. We have a notice board during this meeting. You can visit the APNIC meeting website and there is a link there to the onsite notice board which gives you details of the venue and maps of the different parts of the venue, program information and various local information. Please visit that for up-to-date details and, finally, we've had a bit of trouble yesterday with a new Windows worm called the Welchi worm which, apparently, I understand means 'welcome to China' but that worm is quite disTRUBLGT/ ive and has been spreading around systems, including some of the commuters here at the meeting site. If you're running Windows, it's highly recommended that you install the latest Windows security patches to address that problem or you may well become a victim. Again at the APNIC hotdesk, there are staff which can assist with patches. That's the announcements for this morning. What I'd like to do is look at the agenda for the opening plenary. We have three speeches to come. Firstly is Mr Jong Rok Yoon from KISPA. He will address us about from network provider into solution provider. Paul Twomey, the CEO of ICANN, is joining us to give an important address. Unfortunately, Philip Harris from Cisco, due to health reasons, has not been able to attend and so the third and final Speaker will be Geoff Huston of Telstra. Without further ado if I could please hand over to Jong Rok Yoon to make the first address. Thank you very much. JONG ROK YOON: Mr Paul Wilson and distinguished guests, it's my honour to be here today as a Speaker for this prestigious APNIC open meeting. Let me introduce myself first. I'm an executive Vice-President and in charge of the marketing group. Today will be threefold. First, allow me to give you a brief introduction and let me offer you the Internet access network in Korea. Lastly, I would like to address four major accomplishments of that networking for a successful transformation from a network provider into a solution provider. As you may know, the FIFA 2002 World Cup was co-hosted by Korea and Japan last year. Contrary to popular predictions, Korea proudly advanced to the semi-finals after three years of careful preparation and hard training. The world media were claiming unanimously that the reason for the Korean success was the versatility of players and their endurance. In other words, the ability to cover multiple positions beyond their own. Like this - the successful network of the future will be dependent on the ability to provide a powerful IP network as well as various services based on them. As of September 28 last year, the broad band Internet access in Korea has more than 10,000,000 subscribers since its full-scale launch in 2000. The number will reach 12,000,000 by the end of this year. For remote households and apartment complexes where ADSL cannot be small we use a particular cable that is already in widespread service. Considering that the number of Korean households is 13,000,000, broad band Internet access market seems to have been situated beyond the stage of maturity. KT enjoys a 50% share in this market. Also, while telecom business has experienced a big decline, KT has managed to show steady growth, because it could create a Newmarket - that of broad band Internet service business, with 2.5000000000 users in three short years. To survive in this increasingly competitive DSL market, our corporate strategy is to maximise the network valual by introducing a variety of new services. In short, KT is now trying to transform from a common carrier into a solution provider. I believe that. -- I believe that this transformation is a key to our future success. In the 21st century, people need network service just as they need water and air. It's already an accepted part of our daily lives. Last year, KT announced a special pricing policy for our existent customers. We chose to introduce a flat-rate system in which residential customers can enjoy unlimited useage of local and long-distance cost if they decide to pay an additional 10% to their monthly charge. What caused this was an explosive increase in the number of mobile phones and Internet replacing the PSTN. I believe that this drastic pricing policy was a last resort to prevent our revenues from decreasing. As witnessed by the drop in share prices, a large of carriers in North America, North-East Asia and Europe have already experienced a slowdown in profitability due to the emergeence of replacement services such as wireless communications. Looking back, KT has maintained its tradition in telecom-based network service for almost a century. I would like to reemphasise that a decisive shift into a solution provider from a common carrier is the only viable choice. Based on Brad band access infrastructure, KT has begun its 5-year plan to emerge as a solution provider. To make this transformation feasible, we need to be reborn as a networking company with combinations of elements such as broad band access capability, IP net infrastructure, fixed-net mobile convergeence and an e platform. These four ingredients represent a shift into IP infrastructure from traditional network infrastructure. Based on that, communication network has been based on four As - anywhere, any time, any device and any service in the future. IP network at home will serve as a powerful pipeline for connection. It will enable us to be connecting to PDA, laptop, TV, refrigerator, washing machine, air conditioner, telephone and PC. To put it simply, the networks for media and household electronic goods will converge on to a single IP network. It will be able to enjoy all the convenience home services provided in the so-called 'smart house'. As a new media platform begins, HDS will allow us to watch TV and movies and surf the Internet on a single IP network. Refrigerators and washing machines will also be accessible via the Internet. By producing creative convergeence, HDS will boost alliance and incorporation on relevant industries, such as telecommunications, home electronics and content business and it will offer ambitious opportunities for traditional profits. KT is now beginning to diverseify into home network business that will be launched late this year. We will be able to offer a variety of multimedia services in cooperation with broadcasting companies, mobile telecommunication companies, household electronic appliance companies and social provideres. Let me now make some comments on four elements of the networking. First - broad band access capability. In many countries, telecom networks still serve as an access network. I suggest that the establishment of a prod band network like ESL, over the Internet with a particular cable will be the right choice for the future. This way, devices like a telephone and PC will be evolved into various multimedia high-tech devices and all the networks - different devices can be integrated into a single network. The second element is IP network infrastructure. The Internet with broad band access offers us an unprecedented amount of information. We need a solid IP infrastructure for transmission of information throughout this country. As is the case with KT, Internet traffic is now heavier than voice traffic. 48,000,000 people in Korea en-- 47,000,000 people in Korea enjoy 2000000000 minutes of telephone conversation. Internet traffic is forecast to be heavier than voice traffic by six times in 2006. Voice traffic will be losing ground to IP networks. I anticipate that the telephone centre will be replaced by the Internet data centre. The data centre will play a role as the Central computing power station in local areas. Last year, KT launched a video on demand service. This service will be popular. KT have established contents distribution networks to deliver such high traffic efficiency. Thus, subscribers will enjoy services with band disc capacity of 1 mega pixels on their TV. In an attempt to strengthen IP infrastructure, KT is now in its first stage of establishing next-generation network. revising its original plan to replace it's system by the end of this year, KT decide today adopt an access gateway which will eventually be a next-generation network. From 2005, IP networks will be employed for long-distance calls and normal calls. All networks including class five will be replaced by next generation networks by the year 2010. Thirdly, as I talk about the fixed and mobile convergeence - since the introduction of a DSL service in 1999, Korea has more than 11,000,000 subscribers, it seems to me that the DSL market has already been saturated. In a bit to develop new areas of business, KT will integrate the ADSL and wireless land service. In the past, most households had one PC each. However, with the recent introduction of laptop and PDA, the household finds it necessary to connect all of them together in a phone network. Now, a tome of 270,000 households and public places are beneficiaries of NESPOT service which was launched in August of last year. By the end of this year, NESPOT service will be offered for over 20,000 hotspots such as schools, bus terminals, train stations, airports and banks. With this service, KT is now beginning to extend beyond other companies by mixing wired and wireless services and adding value on the wireless services. 2002 FIFA World Cup was a wonderful opportunity to prove the wonderful usefulness of NESPOT. Enjoying wireless LAN service importers from around the world were able to send e-mails, pictures, video images in their time without any problem. E-Portal is the first conference. It's three key elements I have mentioned so far form a solid foundation for IP networks. Smooth operation of IP networks relies on ISP. To sail the sea of IP, we need a port. That is the ISP. And now we have a billing system and payment gateway for users. However, a growing number of ISPs create many pitfalls. For example, there are 27,000,000 Internet users in Korea. Surprisingly, the total number of user names and pass words hits more than 100,000,000. My daughter has nine user names and nine pass words. She often forgets and mixes up the user IDs and pass words. So far, each ISP has to make heavy investment in developing a system for subscribers. In my view, this is an awful waste of money. As an alternative, KT launched gateway portal. With the function of single sign-on, it ought mates access to all affiliated core service providers through a single log-in. if connected, KT will do everything from transactions to billing on their behalf. So far, about 50% of employees of core service providers have struggled with their users, authentication, billing systems. Thanks to KT's gateway portal, they will be able to concentrate on their own jobs. Similar to an index for an encyclopaedia, KT's gateway portal service has a directory. To enjoy a variety of affiliated services, all you need is only one ID and pass word for the gateway portal. In short, your ID and pass word will be used as your Social Security number in their world. This has been the brief explanation of the four elements of networking. As I have already described, KT now emerges as a solution provider, aided by networking. Let me close my speech to explaining Bizmeka service of KT. It is our e business solution service for small enterprises in Korea. To make Bizmeka service possible, KT established ASP platform on IP networks. With the solution loaded on the platform, all the subscribers can enjoy the bizmeka services. KT offers two types of solution services. The first solution is commonly designed for all small and medium enterprises. On the other hand, the other solution is expected to benefit specific businesses, such as beauty shops, optician shops, taxation offices and so on. With an ASP, small and medium-sized enterprises can enjoy many bizmeka services as subscribers. In spite of its short history of over 16 months, bizmeka now has 150,000 subscribers. Even an incredibly widespread distribution that can handle short business services in Korea, bizmeka is expected to enable its subscribers to perform customer relationship management more efficiently. I am sure that it will get more popular. Our goal is to help the bizmeka subscribers to save time, reduce cost, increase revenues and efficiently manage their customers. If we reach our goal within five years, all the small and medium-sized enterprise will lose in the competition if they are out of bizmeka. Bizmeka is only one example of solution services. Networking will create a variety of new solutions in the future. Privatiseed last year, KT is now planning to diverseify into innovative construction business. Unlike traditional construction companies, we would like to build residential places of a new concept. We plan to take advantage of our expertise in home service in building residential places. One of the future business opportunities will be a distribution of invisible solutions offered by many venture companies in the 21st century. Just as traditional trading companies have traded products and commodities, in the worldwide networks, IT companies like KT can distribute a large number of invisible solutions across national orders. To this end, we need to have network in countries and regions respectively in order to transport the solution in a matter of seconds. Now, I would like to draw your attention to one product KT has in mind, tentatively, may we call is 'solution 1' project. For this project, each region in the world will have representative IT companies. For example, KT will represent Korea, AT&T represent the US. We can have company A representing Asia and company B representing Europe. Equipped with extensive worldwide networks, these IT companies can work on the distribution of solutions. If any of you here are wishing to share my view, please do not hesitate to contact me and I will be more than happy to see you for further discussion. Finally, I wish to thank APNIC for allowing me to share my view on these issues, which have kept me up many long nights thinking about. I would also like to thank APNIC for its efforts in arranging this forum. Thank you for your attention and enjoy the rest of this forum in Seoul. Thank you. APPLAUSE PAUL WILSON: Thank you very much, Jong Rok Yoon, for that interesting presentation. I think, in case there are any questions, we might leave Question Time until the end of the session and move straight on to the next Speaker. Paul Twomey from ICANN. PAUL TWOMEY: Thank you, Paul. Can I say how pleased I am to be both here at the APNIC meeting and also how pleased I am again to be back in Korea and in Seoul. Presentation is more a matter of a discussion and so there is no materials on the screen. I wanted to basically, I suppose, have a 3-part conversation. One was to share with you - if you like, some of you who are not fully aware of ICANN, its originaliness and philosophy -- origins and philosophy. Secondly, to talk through priorities that are the ICANN community is pursuing and especially how they relate to the Asia Pacific. And then, thirdly, talking a little bit about some of the priorities of the ICANN staff and the sort of things we're doing to assist you and assist the IRR. Many of you will know that ICANN was founded in 1998 and was founded in response to a need to coordinate any issues that would be coming into -- intermixed as the Internet became very commercial. And the origins of ICANN in terms of organisation, I think, is very interesting. Traditionally, when areas of economic coordination have been needed in the international community in the 19th century and into the 20th century, the traditional model was to have governance to -- government to government interaction take place, often through an international treaty. This model for coordination and cooperation worked well when the economy was mostly a national character, that, essentially, businesses operated within a national framework and that the government would represent those businesses to other governments. But we now live in a globaliseed world and, in a globaliseed world, the new technologies have meant that the power of the nation state has diminished somewhat. Importantly, the flexiblity of companies and the flexiblity of skilled individuals like yourselves has meant that the ability of those people to interact with each other directly has much increased and is more important. So the ICANN Model which was developed for coordinating - helping to coordinate many of these technical aspects in various parts of the Internet took a new approach. And that approach was to bring together directly the people who actually made the networks happen and who dealt with the business on the network to help identify and solve the problems they saw as applying to this internationaliseing Internet. This model has encountered battles. Importantly, it sees a very important role for governments. I don't want to give you the wrong impression. ICANN is a public-private partnership which has a very important role for the governments. But it also has an important role for the CCTLDs, an important role for the people who have experience in addressing like yourselves, people with IETN and other backgrounds and the forum of ITN is base -- ICANN is basically a place where people can come together and identify those issues which affect people across all those ranges of technology and problems, if you like, and prioritise those and come up with common solutions. It is a model based around introduced self-regulation. In that respect, I think it's somewhat similar to APNIC and ARIN and RIPE NCC and LACNIC, in that it's a model based around practitioners themselves coming together to identify problems and then to help devise solutions. At the heart of the ICANN policy forum is the concept of balance. There is a balance between the private sector and the governments. There are some 76 governments that participate regularly and five international treaty organisations who participate regularly in the ICANN forum. There is a balance between the supply side and the demand side. Part of the ICANN framework has a large process, a process for representation of consumers and I'll come back to that sortly. There is a balance between registries and registrars, particularly in the domain name space. There's a balance between the naming part of the Internet and the addressing part. So I want to emphasise that, at the heart of the ICANN policy forum is this concept of balance. And it is that balance which I think makes ICANN unique in being able to address in some timely manner issues that affect the broader Internet community. ICANN was founded and if first addressed the issues it was confronted with, I would say it was mostly a North American organisation. It was founded in the mid-1990ings after problems that were -- 1990s after problems that were encountered in North America with the Internet. It was founded partly because the United States Government and other governments felt that the issues were better dealt with not by the government but more by an industry self-regulateory body. In the short time which ICANN has existed it has gone from being an entity that deals with North American problems into being one that deals with global problems. In that sense, it has reflected the globaliseation of the Internet in its broader sense. Many of you have been involved in the Internet for a long time. I'm not talking about the framework you've done over 10, 20, 30 years. I suppose our impression has emerged in a truly globaliseed Internet as it has become the basis of economic and social interaction. I think Korea is one of the premier examples in the world of how the Internet has become an important part of social and economic infrastructure. So ICANN now is very much an international entity. It meets in its policy forums around the world. Many of you will be aware of that. Our next meeting is in Tunisia. We have five regions - Europe, Africa, Asia Pacific, North America and Latin America. The board of ICANN is drawn from those regions. It's also drawn from the technical community, drawn from the naming community, from - also drawn from people with business and property backgrounds. The number of people peopling ICANN forums, I think, importantly - just to give you some indication, this is with respect to our American friends in the room but the maximum number of Americans who attend a meeting when it takes place in North America, I think we had 32%. Normally it's a figure between 18% and 20%. I share these numbers with you to give you a sense of the international or national flavour of the organisation. I know that some of the perceptions here in the Asia Pacific have been that it's very much focused in North America and Europe. The staff are also internationaliseing. I'm an English Speaker but I don't speak with an American accent. I'm an Australian. Most of our staff are increasingly not from the United States as well. I want to assume the internationaliseation of the ICANN policy forum. One of the solutions that we are used to addressing and one that is think is relevant to APNIC members - some of them are the following - we are doing a HROL/ lot of work at the moment on Whois registries, the Whois data in registries, both starting in the name registries, particularly the generic name registries, but also - these topics for discussion is having implicationings for the CCTLGs and implications for the Whois registries with RIR. This discussion is really about trying too find a common best practice approach to how do we deal with what data is kept in Whois registries, how accurate that data should be, what is the balance between privacy and discloseure and, as you can imagine, the community has a wide sense of views here. We have people from governments in law enforce T-PLT/ and people with property concerns who would like to see fully accurate data and very disclosed, through to, on the other extreme, people in privacy advocates who think data should be very much limited. The original technical reasons for the collection of this data probably sit somewhere in the middle. So the ICANN community at the moment is trying to improve a best-practice approach as to how we should address Whois and Whois privacy. Another area which you will be aware of which has come through the policy process is the introduction of domain names. I think we're in the process of stage one and several stages of how we internationalise domain names. North-East Asia really is a leading part of the world in terms of the introduction and implementation of internationaliseed domain names. This, I think is a very important issue. One of the things that ICANN is very much committed to is the support of a stable network. I think, as we look at the process of international domain names and other issues, one of the important parts of the ICANN process in this forum will be to continue to help an introduction for the discussions to take place and continue having a single Internet. Another area ICANN is looking at at the moment is new topical domains. This interacts with internationalised domain names. We're starting a process for considering how the internationalised domains should work, in other words, how do people come and apply and gape a domain, how many should there be available and over what time frame and, importantly, I think, part of that discussion is needing to talk about - I'd be looking for leadership from this part of the world - we need to talk about what character sets those domain names should have and what should the process be. The ICANN community has a great deal of introduction of -- deal of interest in the introduction of IPv6. I know that's a topic for discussion here at APNIC. I would have to say that part of the ICANN broad family brings sometimes not great fields of understanding of all the issues and understanding how much floor space is still available and not a great understanding of the IPv6 details but there is a great deal of interest, particularly from governments, as to how it will be introduced, what are the costs associated for the networks and what the issues will be. I think the work you're doing is of great interest to the broad community. I think, finally, I'd say in terms of the broad areas that ICANN is bur suing at the moment, one of the big -- pursuing at the moment, one of the big ones is truly establishing outreach throughout the world. We have established at-large advise ire committees which is tasked with bringing together consumer views throughout the world to be brought into the ICANN process. We're establishing regional organisations there is one being established in the Asia Pacific. So far, the Advisory Committee has approached over 350 consumer and technical organisations throughout the world looking to Internet consumers' issues that need to be addressed in the ICANN process. That's another area where I think you can expect to see more interaction here in the Asia Pacific with ICANN as the Asia Pacific Internet outreach organisation is established. In terms of the third part of my talk - the things we do in ICANN - and one of the important ones is the operation of the IANA. Just to give you some update on that, we are in the process of appointing a new general manager to the IANA and increasing the resources we have allocated to the IANA function. Secondly, we are finaliseing a new process mapping and improving to processes in the way in which we operate the IANA function. And I think as part of that, we have been in discussions with the four RIRs on a number of issues. One is the introduction of an allocation address box and also a discussion about allocation, which is something the RIRs are involved in in discussion. The interaction with the IRRs is very important to us in the operation of the IANA but you're also aware of the IANA interaction and so far there is a lot of support for the IETF. It is actually a multipurpose function. Finally, if I can make some points about security. One of the things that the ICANN community is very concerned about is obviously promoting the security of the Internet. No one group, obviously, can determine or ensure the security of the Internet but I can report to you that I receive a lot of questions and a lot of interest from governments around the world and from other players about the security of the Internet. ICANN has a security instability -- security and stability committee. It involves many people that you would know. We have a very narrow function in terms of Internet security, particularly around the root servers and the DNS. But, more broadly, there is a lot of experience about network security and that's something that we think governments should be involved with in that community. Part of the benefits, I think, of having an ICANN is at least having one place where people can come and direct questions and traditionally where they go to find out more information for security about the Internet. Again, I think it shows us the advantage of having a not-for-profit international private entity which can engage with governments on issues around security. So that was really the main point I thought I would share with you today. Thank you very much for your hospitality and I very much look forward to taking questions either here or talking to people outside during the breaks. If you have any questions, please speak to Paul. We'll move on now to Geoff Huston who will give us an interesting presentation entitled Boom and Bust. GEOFF HUSTON: I must admit what I'd like to do now is set a slightly different tone and take perhaps a light-hearted look. On the immediate future, what's happening over the next few years. Maybe it will become obvious. I'm probably old enough to remember that rockband called The Who. There was one fantastic song in 1971 called 'Won't Get Fooled Again'. I think it was appropriate here. Boom and bust is nothing special. The Internet boom is very mild by many standards If you ever go to Amsterdam, and you said, there's a great museum. Way back on the first floor there it has a picture about as large as that thing there, a lot more beautiful, Rembrants. It was sold in 1537 for 2,000 gilder. 2,000 gilder was pretty spot on. He was pleased about himself when he looked out the window. by 1637, the price of just three bulbs, went quite expense That's a boom. It crashed immediately afterwards and the Dutch economy took a while to establish itself. By about 1719 the French court were getting pretty broke. Along came a Scottish gamblear with the great idea that can't we persuade the French to give us our gold and we'll give them bits of paper. It was a great scam and the French court get themselves out of trouble for a while and then they demanded the gold back. Even railways had their boom. In 1947 there was a great boom in Britain. They said the rail will never work, no-one will ever use it. Boom and bust is nothing special. Perhaps the best way for the Internet is to look at T-shirts. This is one inspired by an Australian, psychoen Hackett who manage -- Simon Hackett who managed to hook a hi-fi into the Internet. He mistakes a campact disc player. This is technology of putting a toaster on the Internet. Of course they ran out of bread by about launchtime and started recycling it. By about the afternoon, the same blackened piece of carbon came down and went up again. A boom is not without its ability to criticise the old ways of being uniquely old-fashioned and ridiculous. The way that always struck me as the way of doing networking by sneezing and standing at the other end of the room and trying to catch all the droplets and re-assemble it. That T-shirt was one of the better ones of the ATM shredder. This was inspired by Bob Metcalfe. At a name there was a NANOG meeting and the T-shirt was obvious and there it is. It's a classic, and if you have it it's probably worth a lot of money. How big a boom was it? Tiny. The price vbingilated by factors of thousands, the Internet share prices went up. In terms of a boom it was relatively mild. Of course we've seen it in all other kinds of industries and they have the same kind of pattern. The initial innovation looked pretty cute. Enthusiasm comes along, this is really cool. That turns and all of a sudden we're in the popular press. Mania is always overemphasising reality. At some point the boom says, "Reality is just outside and it's not on the same floor as you are." All of a sudden people start saying, "Maybe this is not as big as boom as ever." dissolution turns into -- dissolution turns to panic. Let's go back, the Internet is a dud. The reality is different. There is a real industry out there. The Nasdaq index, the share price market does show the classic trend. It went up and crashed in March 20 00, back down again. The role issue is not the boom and bust that's important, it's what we've managed to learn from all of us. What are the lessons that have come through? This is no longer an expansion-driven industry. Growth is not going to save your business. Business models now are trying to consolidate service. It's not a case of doubling your market base every two weeks. You have to look at a more solid business model that looks at market share and how you retain it. Now there's a highly competitive market out there. This is not a monopoly model of doing business, it's a business of where you have to be efficient in what you produce and price it effectively and you need loyalty from your customers to your product. This is not the same old business model. Now, if I'm in this business, what am I looking for? Looking for the amount of traffic every two weeks? no if any manages to have 10 fingers and type for me, no. We're looking at a business model which is more conservative. We're looking at the Internet being quite a dull and ordinary business, based around integrity, being useful to people. It's not just the Internet but what can I use it for? And being price competitive. I think there's less reliance now on forcing yourself way out into vapour ware and slide ware. Folk whose are spending money want to see value for it. Even quality and service performance are less of an issue than simply making it work. So these days the whole QoS thing is treated back. And even the wilder innovations around the application area are actually slowing down a bit more. The money isn't there to take high risks. Because right now optimism alone is not enough. We've actually learned this is a very conventional business based on very conventional business practices. On a low-animated power point slide make -- although low-animated power point slide makes it cute, it's not. It's actually very old. These days the business of communication absorbed up to 15% of the GDP of most countries. Big and old generally mean conservative, generally mean slow to react, generally mean deliberate in the way they work. And the Internet to my mind hasn't changed that reality an awful lot. But although the boom and bust has happened, we now have a backable business plan which is more akin to the business it's part of. The explosive growth has gone. We're back down in shipping, railway and aeroplanes doing the same old thing. Money has an agenda. If you wish to get money to fund an idea, show the investor there is minimal risk and reasonable return. Once you make an investment, it can't be discarded. Once a couple of years we would rebuild the network, who's doing that now? Beyond the ground and up to working for 10 years, one way or another and I can't afford to recycle equipment that quickly. This is a conventional business based around trying to make your investments work. Greater emphas on service, robustness these days and the combinations of bundling are still there but it's focused on a reliable service. From being reliable I move to the next big thing we've learned. We've learned we don't have a secure network. The informal trust models don't work. And that those models are breaking down. We're now seeing a very, very visible security focus. Users are aware of the problems and want to see it fixed. so they say, "Look, it's not that it has to work, but the packet I send to my provider, my customer, should get there. And it should be a packet that I created, not a packet that 16-year-old invented on my behalf." That's more than a relationship with your customers about security, it's part of the fabric about public communications. That implies there's a relatively strong regulatory interest in making sure the Internet works. We're not in a regulatory-free zone. We're in a zone of public sector involvement. Realistic ely, the next -- realistically, we're going to see a remarkably informed trust models that exist on the Internet and replace it. That happens at all kinds of levels. It happens in the DNS, as we've seen in the past, the ability to inject nonsense into the DNS and get people to believe you is high. It happens in routing. The ability to inject false routes into the Internet is undeniably there and it is uncomfortable. It happens in hijacking address space. It happens at all levels up and down protocol stat. The data plane gets ininked as the -- interpreted as the control plane. And it happens in content in all kinds of ways. Vulnerability exist all over the place. When I send you mail, how do you know it came from me? And I mean know, not just guess. When I send you mail, how will I know I'm sending it to you and not some random other person? how do users trust the network? Can users trust the network? The answer's no. How can a user even trust web page snz we're seeing Internet banking out there now and remarkably inventive stands to try and redirect traffic to sites that look like your bank and aren't. How do I know the content I'm looking at? The content I'm interacting with? The system that's out there is the system that I intended to interact with? Unfortunately the answer is that really hard and really hard to have that set of trust. How do networks protect themselves from abuse and attack? How do networks protect themselves from abuse and attack? I've been trying to do this for 10 years and I'm not sure I have the answer. How do users protect themselves? I saw yet another worm is infecting users. How do users protect thevz from this? -- -- protect themselves from this? How do you identify of where abusers are. Whose responsibility is it to isolate them, track them down and get rid of them? The police? Law enforcement agencies? The ISP? Vigil-ante groups? Who's trying to address this problem in a structured fashion? Why and how? The answers aren't very good and to my mind the theme over the next few years is to put answers in place in this area. There's a lot to to in security but that's not everything. In the boom Internet was going to be everything. Internet everything. The issue is that's not the case. The Internet is not a resource management architecture for networks. It doesn't manage resource. End to end systems negotiate with each other dynamically on how to share a network. The network is passive. It's very good if you're running TCP - it's fantastic. When you try and run things like mobile wireless, you find it doesn't help. When you're trying to run tightly controlled realtime signalling, the network doesn't help, it gets in the way. I'm not sure that IP is the transport mechanism for every single application we could possibly dream of or every application that exists today. I'm not sure. And if anyone's making investment decisions in networking plat ferms, you'll see -- platforms, you'll see that same level of uncertainty. We might but not today trash our network platforms. So as far as I can see, IP has a role but not every role for every form of communication and that will be reflected over the next few years. The other thing we've sort of been looking at is in abundance. The bottle neck we always had was long-distance communication whether it was under the through the satellites from city to city. We've now produced a world with long-distance bandwith is among them. What happened then? The last mile becomes a headache. The last mile is capital intensive. It requires you to bury your money in the ground for 25 years and get returns which are probably less than 2%. What employment issues we face happen between the top and the user and that's not going to change over the next few years. Use wireless they say. Unfortunately wireless is a very limited spectrum. Unfortunately, wireless can't serve all of us simultaneously where we want to go. Wireless is 1,000 times more expensive per bit, probably more. It is a wide world in the last mile if you want to work fast. That's going to make a huge amount of change in the way we bill this. We looked at IPv4. We've allocated about half of the space. If you look at the routing table, we're routing about a third of the way through in terms of address space. Some reality there are about 100 million machines out there. So things aren't too bad. Or are they? There are much more than 100 million silken addresses out there. Woo did they get them from. DHCP. Various levels. All these kinds of hints and tricks, to take the address space go further. what are we really doing? We're putting pruft in the middle of the network. The network now has a whole bunch of middle-ware things. Application levels, gateways, wall gardens of function alt. The problem now is that in a network that is full of middle-ware, it's not just a case of putting out end applications on end machines and saying it will work, you've got to figure out how to negotiate through NATS, and how to identify systems. All that is part and parcel of the problem. We're freezing the network applications in so doing. If you think the Internet is web and email, pipe 4 will last forever. The middle-ware gets in the way. the IPv6 says if the problem is not enough addresses, let's have more and they did give us more. You remove the middle-ware and have a much simply base. This device I kind of thought was a great example. In the world of IPv6 it's not screens and keyboards. This is a cute Sony digital camera. I really want one. If anyone was Sony here, I'll use it and say good things about it. That screen to the side is an attach screen. The machine networks. The machine has email. It has a web browser and even Jarva on it. This is the camera on steroids. This is what IPv6 is about. Like I said, if anyone's got that camera, send it over. I'll check it out. Fundamentally, though, I suspect there is a bigger problem and we knew about this problem way before IP was thought about. Way back in the '60s it was obvious that in networking you've got three basic problems. When I want to reach you, I need know who you are. It would be good to know where you are and it would be good to know how to get there. Three very basic questions. What if we split them apart? Rather than doing everything in an IP address, we might well consider using different identity sets for each of those problems. Over the next four years, I suspect we're going to be down there in the research agenda looking very, very hard, because S silicon is cheap. When you do a production run you do a production run in volumes of millions, it's not hundreds of millions. There's a lot of silicon out there. Scaling is the big problem. One of the employment places will be wireless. In theory, IP worked over anything, even carrier pigeons. And some folk in Norway two years ago proved it. Yes, you can run IP every barbed wire. In practice, you can't. Now you're saying, "I'm useing it now." But going fast is a problem. When you try and go fast, your bit error rate has come down. Not even to 1 in a 1000000000. If you're trying to --000,000,000. You need error rates. 3 G, because of its huge radius, finds it hard to give you that very low error rate. What we're finding with 3 G is esit's not working well. -- it's not working well. The error rate comes down and you're OK but 3G is having trouble. It's clear to me that the value packaged presented by 3G will be compelling. VoIP is something that's going to be fascinating. We've managed to see we can do voice over IP. The hard bit is figuring out how to take a telephone number and make it work in an IP world. And it's not even the DNS bits of that. The email part is easy. It's the 70 years of regulatory practice in a whole bunch of national regimes that have made telephone numbers a very, very strange social concept. Who owns your telephone number? You or your telephone company? Can I take my telephone number and move to another country and keep it? Can I move providers and keep my telephone number? Can I cancel my telephone number, maintain my laptop and keep my telephone number? What if I dial an emergency number? What country's emergency service will I be corrected to? What we're going to find with VoIP is all the easy bits have been solved. The hard bits are going to be the change in social means like things like telephone numbers. Of course last but not least services and middle-ware. There's a huge agenda going on there. The network is a service delivery platform and trying to understand how that works is going to be interesting. Directory technologies. Persistent identifies. What happens when the URL doesn't work sghm how do I give you an identifier that is like an ISP book number? There's a role for better authentication. PKI certificates are required and needed more than ever. What have you learned? It was always said in NASA that with another thrust anything can fly. Regardless of the thrust, some things it is bolted for the ground. Social change takes a lot longer than technology change. As you see with VoIP, the hard bits are actually social rather than technology-based. We don't move quickly. Is the international a television station or a stock exchange? It's a lousy television station but the most effective stock exchange we've ever had in human history. The Internet strength is pier to pier. The strengths are there are going to be strengths around individual negotiation. Expect much more regulatory involvement at all levels of the regulatory chain. This is high-value national ast. We're going to see the -- asset. We're going to see the fact this is no longer an anarchy. No industry booms and busts twice. We're not going to see that. We're going to see IP conservative business models now with conservative rates of return. We're going to see continued regulatory increases in what we do. Yes, the infrastructure will expand but the objective of the players are no longer bill, sell and run away in two weeks. The players are now playing for keeps. We're playing in timelines. In other words, and I'm harking back to The Who here, meet the new economy is the same old economy. Thank you. PAUL WILSON: Thank you very much. We are near morning tea time. I would like to open questions in a moment. While Geoff was speaking, I received this email which resounded quite remarkably with his issue of security and things not flying and it comes from one of the speakers of this meeting. He's saying that he's stuck on the tarmac in Canada with airlines down, all computer systems unavailable, passengers being checked. The consistent rumour is due to some computer virus. This hasn't pin verified. It's an email just arrived. It's perhaps quite unknown at the moment. Particularly to the final reminder of the slide which is the Welchi worm going around. I would like to just, in the next few minutes, open questions for any of our speakers. Please feel free to Just before you start with the questions, two of the PCs are still affected and they've generated two counts of traffic. 221.143.141. 221.143.6136. And 221.143.6.14. Looking at it has being where things interacted internationally was actually through nation states and government representation. I can't help observing Paul that the ITUT is one of those interactive models. I take to heart your characterisation of ICANN has being model that is different in terms of directly engaging in players. So my question to you is if you're being given the opportunity to give some advice as to what should pee done, what advice would you give? A. Actually we recently had a lunch and went through issues. I think I'll page 1 back toia. The answer is the answer you gave at the end of your presentation. The ITU is the world's oldest international telegraph you knowian. And in some respects I think reflects the way in which capital had to be mobilised to build with the networks. I don't want to overstate this, one of the reasons why the ITU and other treaty organisations merged was the way the government had to do with those networks. I think one of the very interesting things about the Internet was firstly the US Government... once you get beyond that, the entry cost for new capital is so much lower, but you actually have a whole series of capital. And I think even in your presentation it shows that although it's now going to limit because of the new business models, I think what you see is different sorts of capital. And I think that's why I see the traditional approach of government to government reaction in some respects is not endorseed to a market where you have a lot of private capital. The governments play a role. The interesting issue branched from the ITU is going to be that continued government role. PAUL WILSON: Would there be another question? Perhaps Jong Rok Yoon? OK, we are moving into lunch break. So again I would like to... I'm sorry, the morning tea break. I'd like to just remind us of these announcements. Thanks again I missed naming the sponsor for today. It is JPNIC. TWNIC we've thanked already. Thank you JPNIC fr sponsoring the meeting for today. KT again. Breaks are outside. I do encourage everyone to come to the river cruise tonight and meet in the hotel lobby at 7pm. I'm not sure if I got the IP addresses of the latest victims of the Welchi worm. 136. If these are your IP addresses. 221143. after morning tea, we will come back for the sessions.