 Good morning. Welcome to the 7th Sunnergy International Broadcast. We're coming to you from California and we're reaching North and South America and Europe. The delayed broadcasts will be in Asia and of course all this will be digitized and available on the internet. Let me describe just a bit how we're reaching you so that you'll have an idea of the technology involved in all of this. We have two C-band satellites and three KU bands. We're on G-star 2 for US, KU, domestic. We're on Galaxy C7 for US and Mexico C-band. We're on IntelSat KF1 for Europe for KU and for Latin America and South America for KU. We're on IntelSat KF1 on a different transponder. The reason I tell you all this is that those of you that watch this broadcast are technical people and you all have the capability of using a variety of technologies to receive information. That's the point of this show. The next show we do, and I want to get you prepared for this, the next show we do will bring you in the television signal all the documents and information that we'll be providing you today by referenced FTP sites or other internet transport mechanisms. So the way this will work, and this is what it's time to be prepared for, the way this will work we can encode almost a megabit in the vertical blanking interval of any video signal, which means you need about a $20 in parts device to decode that and that brings you all the text we'll talk about. So that's enough about technology. We'll talk a bit later about how this works. What are we going to do today on Synergy? We talked about cyber jockeying. We don't like these strange words, but the point of it is to say we're going to talk about how people are linked today and talk about the internet and its remarkable growth. We're going to talk about how you find things. Once you're linked, how do you read? What's a library like on the internet? If you can make the network go faster, what technology, ATM in this case, will revolutionize the interconnect of the world, and then how do you keep things secure? How do you have privacy when everyone's linked at very high speeds? Our guests today to talk about these questions, Carl Malamoud, the founder of Internet Talk Radio, Brewster Kale, Wide Area Information Service, Larry Irving will join us at the end to bring all this together as the director, the head of the National Telecommunications Information Administration weaving together in a national policy perspective what Carl, Brewster, and Dr. Whitfield Diffie, the inventor, the creator of the Diffie-Hellman public key encryption algorithm, how these different elements all can be combined into a national and global information infrastructure. So with that said, I think it's time to begin the program. We have a mechanism for you to be able to ask us questions. We have a telephone number to call, there'll be an internet address, and there's a method for you to be able to send faxes across the internet. Carl Malamoud and I will discuss that in a bit more detail. In fact, why don't we join Carl now over at the table where the program will take place. So let me go over there. Well, here we are, and here is Carl. Good morning, Carl. Here we are. Well, we've talked a good deal in the past about what the internet means when you bring people that actually didn't know before that they would be connected suddenly into connection. They want to read things, they want to hear things, they want to see things. In the past, the internet has been too slow for audio and for video. And today, the internet, excuse me just a second, today the internet has become quite fast, faster than anyone anticipated, and we see in the next two years it will become much faster again. You created something that would take advantage of this. You became a new international network. You became a radio network. You were almost a television network. Can you describe a bit what you've been doing and how this works? Yeah, we publish radio programs on the internet, and by that I don't mean transcripts, I mean actual audio files that are sent out over the network. Sometimes they're sent out in real time, so you're sitting there on your computer and you listen to the radio. You might listen to the Dalai Lama speaking at the National Press Club, which we broadcast, or you might listen to my show Geek of the Week. You know, when you say you broadcast, describe what you do. I mean, here we have 20 million people today. At least that's the most recent number from three weeks ago at the internet conference. 20 million people and some of them are just on phone lines. Some of them are on fast lines. How do you broadcast to them? What's important about the internet is that there's a variety of ways of interacting. You can talk to the internet by electronic mail, or you can use a real-time video connection. And we try to span all those different speeds, all those different media that are available. So in the simplest thing, we just take a file. And the file has audio data in it. Just like your word processor, word processor is a document. You play a file. You know, when you have your Macintosh and you open it up and it quacks like a duck? My program is a 30-minute version of that duck quacking. And so it's just a file. They're big. I'm not sure you're giving the Dalai Lama. Is it different? Well, his holiness does not quack. There's no doubt about that. On Wednesday, the National Press Club, you mentioned the National Press Club. So here's the prestigious site in Washington where the world press has their offices. If you read something or hear something or see something anywhere in the world about Washington, a good chance is that it's emanating from the National Press Building. You have an office there, and you have fiber, and you have lines to microphones, and you have devices. Describe how you built this. We are right next to the Kansas City Star in the Arkansas Gazette. And on this door down a long hallway, you'll see Internet multicasting service. And if you walk in, you'll see a room that's kind of half radio station, half TV studio, and a whole bunch of computers. We have the fastest link in Washington, D.C., to the Internet. Our Internet connection runs at 10 million bits per second, which is the speed of your Ethernet. And I actually get better performance talking to the outside world than I do inside my network. If I'm talking to a fast spark station, for example, I get much better performance going across the country than I do going across the room. So I've kind of inverted the fire hose, if you will. And what we're ready to do there is to pump large amounts of data into the network. And one of the pieces of data we pump out are the National Press Club Luncheons. Now, how often do these occur? Oh, two or three times a week, there'll be a National Press Club Luncheon. And if you walk into this grand ballroom, which has got 15-foot-high oak ceilings and oak walls and panels, and you'll see C-span there and all the media, and you'll look up and you'll see two broadcast booths. And one of them has National Public Radio. And the other has us. Do you have a banner? Does it say Internet Talk Radio? No, but when the president introduces his guest, he introduces, yeah, sorry, Arafat last week, for example, who came and spoke at the Press Club right after his signing. And he welcomes the members and the guests that are there, and those of you watching us on C-span, listening on National Public Radio, and listening on the global Internet computer network. And whenever that happens, everyone's sitting at the head table, all these distinguished journalists, kind of looking. What was that? We think if we repeat it enough times, they'll start remembering. Well, I think you're demonstrating by doing this that the oldest conceptions of what a radio station are, you have how many listeners do you have? We're estimating about 100,000 listeners in 30 countries. So you're an international radio network, and all you're doing in some sense is taking the audio from the podium, digitizing it and putting that file out on the net. It's a little more than that. What we're trying to do is dispel the myth that you have real media with real producers, and then you have the Internet, which is just a bunch of colored students sending mail to each other or academics, and the Internet is not. It's 20 million people, 140 countries. It is the global village, and we need people producing real information. I don't want to hear every graduate student in the country tell me his views on politics. I want the New York Times to tell me about politics. You're adding the... We want to see professional production on the Internet. We want to see real media. We're a nonprofit, and what we're trying to do is show national public radio and CNN and these other groups. Here is how you, who produce information after all, they're not in the satellite business. They're in the production of information business. And we're just saying, look, here's a new medium that you can send your information out onto. Well, I think people are beginning to understand that when you do this, it changes the definition of how you make money with television stations broadcast. It essentially destroys... Maybe that's too strong a word, but it destroys the monopoly that a small set of people have had on broadcast, in some sense, on information. The entry cost is much less, although I should say that the hardest part of running a TV station ultimately is not the fancy studio, it's the people. And ultimately, that entry cost is still going to be there. And if you can assemble a New York Times, or an Atlantic magazine, or at all things considered, you will still have an advantage over the kid who is doing desktop broadcasting. But the kid can do it. So what we're doing, we're giving you a new pathway, a new career path for the kid who becomes a media murdoch of... A mini murdoch. A mini murdoch. There's a scary thought. Well, I think that since it's global, and Murdoch is attempting to be global, and certainly has become global, if his new satellite, Asia sat with 53% of the world population, covers and brings down in the vertical blanking interval the South China, any of the newspapers of Asia, it's going to alter how publishing works. And you're essentially the bellwether of this by bringing this today to all of us on the Internet. So how do you get this? How do I...is the software free to bring the audio... Oh yeah, it's data. It's data. We make our money off advertisers. In fact, Sun Microsystems is one of the advertisers on Geek of the Week. When you think Geek, think Sun. I'm not sure your advertising budget's going to stay after that. Yeah, well, we'll see. But it's basically just files. If you want information, you just send mail to info at radio.com, and that sends you back a frequently asked questions list, and it says, here's where you can find the files, and here's how you can get on the announcements list, and here's how you play the data. Will we roll this later on the FTP sites? Oh yeah, that's out there. It'll be listed so everyone will be able to find it. Now, this notion that we're suddenly allowing anyone to become a producer, at least start training, anyone can become a reporter, it gives a new career path for those people currently doing this. I would very much like to have the moral authority, the truth telling capability, of reporters whose material I read, I trust them, and when they go someplace for me and report on it, bringing that on to the worldwide Internet gives them a new pathway, and lets us read more things. Now, there's something new you started, and this I think is, again, a warning shot in the world of telecommunications policy. You're allowing, the Internet is allowing, the Internet Engineering Task Force has arranged for fax bypass of the telephone. Well, the IETF really didn't, that's the Internet Engineering Task Force, didn't really arrange this, although we did document what we did and explain it at the IETF, which is what one does in the Internet when you have a new set of technical specifications, but what we did essentially is, about six months ago we started a radio station to demonstrate the amount of leverage you have with the general purpose infrastructure, and that started working, so we started a telephone company. Now, how do you mean you started a telephone company? Well, here's what we did. We looked at all these special purpose devices out there in telephone land, fax machines. When you send a fax, it's just a silly way to communicate, but it's an important way to communicate, and there's millions of people doing it, but think about this, you're taking a document, you scan it in at full resolution, you set up a dedicated link to the other end of the world where the other fax machine must be waiting for you, and be ready to receive, and you print it out. And we look at that and say, you know, that's expensive, it's slow, it's silly, but there's a lot of people that use this. And so our goal is to integrate special purpose devices into the Internet so that you can send e-mail to a fax machine. Now, what kind of an e-mail address would this look like? Well, all you need is the phone number. Can we see this? Is there a graphic there? All right, now here is a picture. Let me explain what this funny-looking address is. Here's what you do. You're sending to the tpc.int domain. You know, there's sun.com, so if you look at the third line on your graphic, it says tpc.int. So there's sun.com, com is commercial, sun is the next level down. It means sun, and then underneath there there may be ink for engineering. What's tpc stand for? Well, int is international, and the international domain actually has two major groups in it. It's got NATO, and it's got us, and we're the phone company. tpc.int. And so if you want to reach a device in telephone land, you send mail to tpc.int. Now, how do you tell it which phone number you're going to? You take the phone number, the whole phone number, okay? So if you're in the U.S., you've got a country code of one, an area code of four, one, five, the whole phone number, and you flip it around. Oh, so what we're seeing here, the number on the screen, back to the number on the screen please so we can talk about the number, the number on the screen has, at its far right, a one. That's the country code. That's the country code. In Paris it would be three three. That's right. Okay, so you put the whole number in and you flip it around, and I know it looks cumbersome, but there's actually a real good reason for doing that. We don't need to hear the reason. But you don't want to hear the details. Okay, so yeah, flip the number around and you add tpc.int at the end and you want to give a user name. Now, why would you want a user name? Well, that's because the first service we're deploying is remote printing, reaching fax machines, but we'd like to see other services. We want to see all the special purpose devices like TDD, telecommunication devices for the death. Telephones to me are a special purpose device and if somebody leaves me voicemail, I don't want to have to call up a mailbox on the telephone. I want to see it in my electronic mail folder. It says you got a fax, you got some voicemail, you got some email, you can respond, you can print it, you can't really print voicemail, but you can simply look at it. Well, so at the moment now, I would say I'm in Singapore or I'm in Mexico City and I want to send a fax to someone in Washington. I use that formula and I put my document in post script or in some form. That's right. And then I just mail that mail, that address, the document comes across the internet, it arrives at some machine in Washington that knows that it's to receive it and that machine makes the local phone call, delivers the fax. So you see there's two costs involved. There's your cost of sending electronic mail, which I should point out is probably a lot cheaper than your cost of sending a fax. It's not free because the internet's not free and if you're MCI mail or CompuServe or one of those groups, it may be 10 or 15 cents a message. But email is distance insensitive and so when you're faxing to Australia where we have service already or the Netherlands where we have service. Japan, how many countries now have service? Oh, there's three or four that have service. There's another half a dozen that are going to have service. In fact, after I'm done with this broadcast, again, I'm an airplane for Japan where we're seeing what we can do to get service up and running in Tokyo, for example. Oh, good. All right. I thought Tokyo was covered already. It's not yet covered. We're working on it. We're going to do an experiment of Professor Jun Marai. So this is in the next month? We hope to see it very soon. Good. I like these timescales. I don't like waiting. We're not talking years. We're talking years. All right. Well, and voice will be something. Well, this is a remark to me. It just makes an astounding case for the ubiquity of the internet causing new services and the point you made about advertising, inserting advertising on the fax, a new source of revenue, new sources of business, new kinds of jobs are derived by making this kind of interconnection. Yeah, because remember I said there's two costs. Your cost was a local cost. And you might be willing to send faxes within your research lab. That makes sense for you to pick up that five cents or zero cents. But if you're sending a fax to some stranger, how do you pay for that local call? And the way you can pay for it, if you want to be a commercial operator, is you can send up to one-third of the cover page can be an ad. It's not junk fax, but it's actually putting advertisers to useful work. They're sending your fax for you. Well, they're paying that person that has a computer that will make that local call and send that local fax. This is a brand new mechanism to reach people. And the volume of fax traffic is increasing exponentially. Well, it's nine billion dollars in the United States as the fax markets. And if you do store and forward and messaging and all these techniques that we're using on the internet, it's about one-thousandth of the cost of doing a regular fax. That's a wonderful drop in price. A thousandth of the cost. We want to see that nine billion dollars that American consumers are spending and more that are being spent internationally be reduced significantly, and the rest of that money be diverted instead of into point-to-point expendables, phone calls. We want to see that money invested in the internet infrastructure. That's a very important point. When we have the volume of information that we have, we have the next problem. How do you find what it is you put away? How do you actually retrieve? This is a bibliographic problem. This is the librarian problem. How do you put five or ten million volumes on the shelves and someone wants to come in and find something? How do they find it? Our next guest discussing these questions, Rooster Kale, is going to take what you're building, which is creating an enormous volume, an enormous flow of information as the internet continues to grow and use it to be able to retrieve the, in fact, I think before we show a bit of what Brewster does, put the number up one more time if you could, the graphic of that phone number, because that is actually the internet address if you'd like to send a fax to us here. So it's one, reading it backwards, it's 1415-988-7003. That's the phone number for the fax machine here. But you're not going to call the fax because that would cost 1,000 times more than the way you're going to do it. So you're going to send your document to remote-printer.sunnergy at that number dot, and this is all one line, this whole thing. Am I right? That's right, spaces, that's right. Make sure it's remote-printer not underscore. Great, let's try that. All right, fine. So let's now see what happens when you're attempting to retrieve information. You need to ask questions. And Brewster has created something at Thinking Machines and now in the WACE company that he started here in California, something that allows you to answer questions. What library has the coordinates of satellites in space? How did we get to the healthcare situation we have today? What library has information to compare traditional Japanese values to Japan's modern society? Brewster Keo may not know the answers to these questions, but he has WACE, that's W-A-I-S, to find them. WACE, Incorporated, is Keo's company. They produce wide-area information servers, a software product that allows users to search dozens of databases and libraries around the world in plain English. WACE is basically a system to try to actually do what computer science has been promising for 20 years, which is offer the Library of Congress on everybody's desk, allow people to find what they want out of lots of different information sources. The user community that we're trying to have be able to use WACE resources is really very broad from K-12, kids in schools, universities, government people, librarians, people who use libraries. We want those people to be able to use the sorts of tools that the whole WACE community is making. At MIT, Keo studied computer science and artificial intelligence. Then, at Thinking Machines Corporation of Cambridge, Massachusetts, he put his skills to work designing and building parallel supercomputers. My personal background comes from very large, fast computers where we were able to search through gigabytes interactively. Now you can do that with a sun machine and the old type of UNIX machine to be able to look through those types of quantities of information. The question is, then, how do you make that kind of power available to lots and lots of people? And that's where the Internet comes in. You can put a database up once and it's accessible to millions of people. This potential for easy global information exchange via electronic publishing and online libraries became the focus of Keo's work at TMC. He developed software for searching databases and for a nationwide computerized library system. From the beginning, Keo recognized, as he says, people want to be able to pose a question to the net and not care where the answer comes from. Keo designed WACE as a standard for use on the Internet. To be able to have WACE expand quickly, we need cooperation between international groups, multiple vendors that often are competing with each other. The best way to do that is to base it on international standards. We picked Z39.50 as a major piece of the WACE protocol suite, as well as some of the IETF protocols, such as URLs, URNs, and starting to move towards Kerberos and public key encryption. There are now 475 WACE databases registered with the directory of servers. They're served from 12 different countries. So you can go and search resources in Australia, Japan, Finland, Singapore, often in their native languages, making it so that people can interact, not just using English language, to find what they want across continents. This equipment here is basically what it takes to become a modern publisher or an electronic library. It's not very much equipment to be able to reach worldwide audiences. We have a 56-kilobit Internet link, which is a few thousand dollars a year, some CPUs that allow basically people to search and retrieve through gigabytes, and about 5 gigabytes, 6 gigabytes of disk space and tape drives to be able to store the information. Basically what this is, is a collection that is the equivalent of the printing press from 5, 10, 50, 100 years ago. This is what it takes to distribute information worldwide. In this case, it doesn't cost almost anything to make an additional copy, or in the printing press it did, and also the printing press at the disadvantage of producing physical copies that had to be shipped around. Now, what does a computer wizard who builds libraries in cyberspace count as his most prized position? A printing press. This antique press occupies a place of honor in Brewster Kale's home, as well as in his heart. I've been learning more about printing history, basically as a way of figuring out how should we do the electronic version of all of this stuff. We've got printing history. It's got all sorts of bits and pieces of setting and piece of lead for every letter that you wanted to put, the separate pictures. And it's a lot like where the Internet is now, where we have pretty much separate text and images, and we're trying to figure out how to get them together. The business model in printing took about 50 to 100 years to really figure out how to get the copyright structures together, how to get that all to work, and we're all in the middle of that in the Internet now. Perhaps only a man with a love of words, a vision of libraries that are always open right on your desktop, and the respect for text that all printers feel, could have invented ways. I'm just amazed by that because you putting ink on paper, the Gutenberg touch there, is I think very appropriate to talk about what does Internet, this sort of ugly electronic communication mechanism do to the beauty of a book, and what are books like? One point that really struck me in this, the announcer that was going to put this piece together, didn't hear you say, but you said very clearly that this is not just English. This is every language. Tell me how this works and what does Waze do? Waze is basically a mechanism of making information available, whether it's in Japanese, whether it's in Cyrillic, whatever the original form of the document is. It's a mechanism for moving it around. We're not really saying that we should burn all books or we should read all our books on screens. Screens are too ugly for that, but it's a mechanism for making people be able to put the stuff out, so there are people in Singapore, Guatemala, that are making their works available like they never were able to before because of the Internet. That's really, I've heard pundits. There are people running around the computer business. Who was it? The guys, McCracken from Silicon Graphics, saying books are dead, our graphics are going to kill books, and I thought this is one of the silliest things I've heard. It just, the centuries, you mentioned how long it took to put the business model of printing together. Just the technology of the book, the index, the table of contents, what's a footnote, numbering the pages. There are businesses in the United States that take federal data, the law reports, and add page numbers and some indexing, and that's a very serious and valuable business. In fact, I want to talk to Larry Irving later about this because this is a very relevant topic. How do we add things to the pile of information that you're going to help search through? Network publishing isn't about getting rid of books or surpassing books or saving trees. It's about the new types of literature that it's going to come out of the new mechanisms of distributing information. We're not trying to make books obsolete and anybody who says so probably hasn't read a book lately. It's the new types of information that we're getting over networks. We're just finding that it's not that much time to be dealing with the types of paper information that we don't care that much about. We're talking about pages, classifies. Those sorts of things do much better over communication networks like the Internet already. And the key is to make it so that there's critical mass so that there's enough out there to be able to find. When you started this at Thinking Machines, you had a 16,000 processor, a highly parallel of 32,000, 64,000 processors, ripped through years of the Wall Street Journal and gave you instantaneous text recall. Now, how do you do this on the Internet with slow communication? You have more than 64,000 processors on the Internet, but they're not linked by a high-speed network. It's hard to remember that a gigabyte used to be a lot of information, but in the early 80s, to be able to peruse through 15 gigabytes interactively required a supercomputer. Now, 386 machines, sun processors can basically make thousands of books, 10, 20, 50 years of the Wall Street Journal browsable for the capital costs of, $10,000 for a UNIX box. And you can have all the dissertations in a university made available to everyone in the world. The astonishing drop in cost is, again, sort of what drives people to this. It's not because it's grand and glorious and better. It's because there's things we can do now that we were never able to do before. And it doesn't require supercomputers for most people's collections. Library of Congress will probably need something like that. But most people, most universities, companies for distributing their own information within themselves, that's in the order of just a normal UNIX box. Yes, well, that's the e-mail problem here. How do you actually search all those pieces of literature that have been destined to you? The biggest use I make of these services is searching my own e-mail that I send and receive about 100, 150 messages a day. I save all of it and have for six, 10 years now. It's only about 400 megabytes. It's not very much information. I use it all the time. It is my memory. I don't keep an address book. I just go and try to find who the person is that I'm trying to contact is Francois from Johns Hopkins. And I can just say Francois from Johns Hopkins up comes not only his e-mail address, but what we were last talking about, that type of extension to our memory of having these facilities available to us are just phenomenal. Well, I think in a way it takes a load off your memory because you don't have to remember where things are. You have a nice way to search. I always find things. I place things in stacks. I'm one of those sorts of people that has a spatial sense of where I put the thing I should respond to or the paper I should read. So, is there in the browsing mechanism at the moment you're searching text and you're finding in a complicated, sort of multi-dimensional indexing way what the meaning of a piece of text is? Is there any spatial browsing that's coming in this interface and doing things, walking down stacks? People are doing all sorts of different metaphors on the desktop. And we'd love to see that, especially as the screens grow, most of the interfaces are designed for extremely small, poor interfaces where all you can get is a headline list. But we're starting to see apples doing layouts of newspapers so you can browse very quickly. And as we get larger screens, that makes sense. Xerox PARC has done some very interesting information wall work where you get a two-dimensional plot of time versus how interested you are in things. So you can zoom in and know that happened last year, so you can move over to that part of the wall and start browsing around. Basically, by basing this stuff on open standards, it allows lots of different companies to compete on the user interfaces without driving the publishers and the librarians nuts where they have to go and put this stuff up over and over. Okay, now CompuServe's got a new system, now Sun's got a new system. They don't have to do it. What we want to do is make it so that people put their information up once on the internet and people gateway to it with all sorts of different information. For the static databases, I can see the existing pile of materials, but the dynamic ones, the ones that change with time, there was a tape that was made of you searching through data that changes every moment. So let's see that now because I think it strikes the new source of information, which will be much larger, I think, than the existing amount we have to search through. Another interesting databases, things that are updated continuously like the agricultural market news that's maintained in Oregon. So you can find out real price quotations over the network. For instance, you can ask, what is the price of roses in San Francisco? And again, just asking the question in natural language. The servers ignore all of the junk words and try to find out what it is you're talking about by matching the words and phrases. And here we have an article on San Francisco for late last week where it has the prices of roses at the top at about $0.48 per bloom. So this is real updated information coming off of the San Francisco, from the wholesale market trends. Another sort of server that is also of the same type that changes very rapidly is weather reports, and this weather server is populated with data that's from Michigan. It's the same information that is put on the nightly news. We can ask, what is the weather in San Francisco? And this is now going from this machine over the Internet. And the top document back listed San Francisco and all many other cities that actually have a little bit to do are relevant ranked. So here we say San Francisco Bay Area Report for this morning and it's another beautiful day in San Francisco. We can also ask of the same server, we can ask for weather minutes. So Waze is not just used for text, but also for graphics. What is the map of the weather in the USA? This server is just indexed the titles of these files. And these are a little bit cryptic in terms of exactly when, therefore, but they're Greenwich Mean Time. Here we have another amateur server where it's not the most beautifully made, but it's free and people are using these sorts of resources all the time. Now it's just downloading a surface analysis map, which is 36K, and displaying it in a GIF image. And we see that Florida has some problems. But California is all clear. There are a total of 475 databases available on Waze. You can find them through gopher sources or through Waze interfaces or worldwide web interfaces like Mosaic. Any which way makes sense. Some of the different users prefer different approaches. What we're trying to do is make information available to people no matter how they want to look at it. So if we were to look through the directory of servers, we can find all sorts of different types of information. Well, that's the flood of information we're talking about. I have another example of it, since Carl and I were unwise enough to discuss the fax bypass on the Internet. Now, testing it. So I just, in the last minute, I've received two faxes. You received, when you announced this once on a show... We went to Interop about a month ago, and we had four fax machines there. And we said, gentlemen, start your faxes. And within one hour we had about 300 faxes come in from 12 different countries, including like Alice Springs Australia sent us faxes for the cost of an email message. All right, Alice Springs, get on it here. Let's see if we can beat this record. There's an absolute demonstration of the maxim that computers kill trees. We're going to have more... So here are two. One came just now from Ann Arbor, from Peter. And you'll notice that the... I don't know if you can see this, but there's the ad. And the ad shows the phone company, TPC, and it says, think globally, act locally. Which is sort of an advertisement to get advertisers. Actually, if you're not a commercial operator, we allow you to put a public service announcement. And that's our version of a public service announcement. I like this. I like this. You can say anything you want. Well, this is great. All right. So those that are involved in political campaigns, I know there might be a few of you out there. Here is a technique to pay attention to. So... All right, well, what we have now, let me see if I can get clear about this. The last set of demonstrations of the weather maps, the information that's just flowing constantly across the Internet, allowed people to see... So that was pretty. I mean, you said it wasn't so pretty, but it actually was. There were nice... You could quickly see where the weather was going. The Department of Agriculture and other people in years past have spent a lot of money trying to make sort of ugly little PCs do this thing and make a slow modem do this. And now, with connectivity, they don't need to do that anymore. This replaces it. Yep. Sun actually uses ways to distribute its own information, bug fixes and alike, and it distributes $12,000 a day, and the capital cost of that was about $50,000 bucks. So for $50,000 one-time plan, $12,000 a day, means that there are a lot fewer phones ringing or a lot more customers that they can get to. They're all free, and we're starting to get the first four-pay services. So people are charging for per access or per month for getting ad information resources. Wait, there's one little detail here. We're fixing $12,000 bucks a day. This is great news. You're telling people about $12,000 bucks a day, I'd say. I'm sure there's far fewer bucks a day. Yes, it's just software, right? That's just software. Well, one of the nice things with these $12,000 bucks to move out to people is we can do it faster. And that graphic, that weather map that's moved in a static mode, I mean, there's a static graphic. You search it and you find the most recent one, but it's still static. As we change the speed of the networks, this, I think, is going to make such a difference again. And I think in the next two years from what we've been, as the ATM revolution occurs, we're going to increase the speed of the network, the effective speed by at least one order of magnitude, probably two orders of magnitude, possibly three. That's 1,000 times faster in moving things around the net. One of the most important developments has been the use by the computer companies of the telephone companies' research money. They worked on a digital standard to link central offices. And now the packets that they have come up with the cells, the asynchronous transfer mode cells and the switches that make those packets move to the right places and the speed of the wires that this can go on, it's just a revolution. So there's a company, Make Systems, that has been working on the hard part that comes next. First we create interconnectivity, then we put information on it, then we make it go faster, and how do we manage the networks to do this? So we have a videotape that shows how you actually can control this enormous expansion of the network. So let's look at that. The ATM was engineered in the CCITT standards bodies primarily, and it was initially conceived as a solution for the BISTN, the broadband integrated services digital network, and a primary requirement of that group was to handle voice and video traffic along with data traffic in the BISTN network. An ATM, because it's based upon a synchronous physical layer infrastructure, and because of its multiplexing method, can very efficiently handle multimedia streams and deliver both connectionless and connection-oriented video streams across both a LAN and a WAN network. There are a number of interface rates defined for ATM, but basically the concept of ATM is that it's a massively parallel infrastructure, and you can create bandwidth by adding more hardware capacity to the network. Typical interface speeds for ATM range from 50 megabits per second to 600 megabits per second and beyond. If you think about it, in an ATM network, there's very little latency in the switch device, so the predominant delay is literally the speed of light in the fiber in the network. So we're talking about a network with a latency from, say, New York to Tokyo that's equal to the latency today from my keyboard to my hard disk on the order of tens of milliseconds. Okay, we've talked about the technology. Now let's go take a look at an actual network model. At this point, you might ask, how are these simulation tools evolving to support the transition to these new broadband technologies? Well, it makes systems we feel the simulation tools are very important in the transition to the new technologies even ahead of the introduction of the ATM network itself because it is a very major transition in the technology and the end user wants to model the network before it's deployed to understand how to properly apply the new technology. We're evolving the NetMaker platform into a fully distributed system, object-oriented, capable of multi-user operation but most importantly, it will be capable of multi-class simulation of TDM, router, and ATM traffic in the same network. So now let's have David Simkin, our director of industrial design demonstrate this enhanced capability. The complexity of today's networks is leading people to reject the real-time, fix-it-when-it's-broken form of network management for managing your network asset. This situation is very similar to what was going on in the computer industry, where high-scale integration is possible in part because of computer-assisted engineering. In an analogy to that, we're really putting CAD kind of processes down for network management where there's different functions that a person performs such as simulation, designing, analysis, all available to them through this one interface. What this interface needs to provide for people is a way of decluttering the mass of data that's available in much of the same way that I'm doing here. So for example, I've honed in right on the information I need about my network and in particular, maybe I'm interested in how this traffic, this circuit, if you will, routes through the network. To do that, I can just simply take it and drop it in a window over here and in the lower window down here, I can see the true route that that traffic is taking through the network. Further, I can take that and further explore it by dragging it, dropping it on here and I'm left with a picture of a demand that is the traffic through the network and the route that that traffic takes through the network. Once I'm in a situation like this, I can start doing what if scenarios, what if this link broke, what if this traffic increased. And in this way, I can really come to the optimal price performance for my network. Well, this is, I was going to see, looking at that graphic, that's part of a new waste interface. So you can see when you're querying where everything is. This would be a geographic location, I think. Already, the USGS has made it so you can mark on a map what maps you want and then it will go out to databases to find the right maps for you. You can imagine it finding the right text for you as well, but right now, they're doing it based on maps. So they mark a map and it retrieves documents relating to events within the region you mark. Is that the... It's really, it's bringing back maps that cover those particular areas. So using one standard, Z39.50, we're able to get at maps the same way we can get at text. I think that's so powerful. I don't know how many people noticed on the wall behind Stephen in that last video. Steve Bell had behind him what's been called the world's best graphic conveyance of information. It's a map showing Napoleon's armies marching from Paris to Moscow. And the width of the line shows how many people were alive at any given moment. And the line gets smaller and smaller and smaller, it reaches Moscow and it gets very narrow. And coming back, there's just a tiny trickle of people all the way back to Paris. It's the most powerful, single, one-page demonstration of graphics. And now I think with these new tools, they're in the green. I love your graphic. I think we're going to have those sorts of graphics in the front of these kinds of searching. Well, Carl, I mean, you heard about ATM. What do you think that's going to do to your business? Oh, I think ATM is crucial. I think it's important to understand that we already have an operational internet work that uses a lot of different technologies. So we already have a global village and ATM will be an important addition. It's a kind of addition that I think has a national interest. I think when we hear of a national information infrastructure, we shouldn't be looking for telephone companies and large corporations to provide information because we're finding that the individual or the university can do that themselves. But there is a federal role to provide a national, seamless, high-speed network. Now, there might be many providers, but it's going to require government leadership in order for that to happen. We'll unleash you on Larry Irving, who I think echoes your sentiments, but we'll have to have him speak for himself in a moment. We're going to link to Washington and bring him in before we do that. Some people have made videotapes of questions that they have sent in. They're from Guatemala. Could we see the video of the question from Guatemala? Hi, I'm Freddy Rosal, challenge manager for the Sun Distributor in Central America. From Temple Fort, the tallest temple in the ancient Mayan city of Tikal, Miss Greta de Pache, one of our customers in Guatemala University is going to make the question for us. Tikal was the most powerful Mayan city in this area, thanks to its commercial and cultural links to other cities in Mexico and Central America. Today, this cultural and commercial synergy remains just as critical. However, our region lacks the modern communications links and we have few resources to invest in them. Therefore, we have to be very careful in choosing the right technology at the right time. Our question to you is, if you were starting from zero right now, what would you invest in? Would you invest in an internet infrastructure to recreate a commercial system based on commercial standards? And how would you do it? That's a fundamental question. So, Tikal, the city we saw, had roads and that was a fundamental element of the Mayan empire. You moved. Today, Guatemala needs more and new kinds of roads. So, what do you think? How would you answer this question about commercial versus this thing possible today in Guatemala? I think the question of commercial versus internet, the answer is you do both. The internet is not federal standards. It's not OSI. It's not some other thing that the federal government did. The internet is a set of very practical things. Waze, for example, was deployed by Brewster and his colleagues and they just did it. And I think the internet is an environment, a national infrastructure on which the commercial operators flourish. And so, I don't think you do one or the other and it would be a grave mistake to say our internet will be provided by this corporation or that corporation. It has to be a competitive information economy. So, the answer is both. Without knowing more about Guatemala it would be difficult to say just who can do this. But I think as the pattern in the United States, it began with universities, research centers. Those people, in some sense, have all this time to spend to fix the 12,000 bugs that appear in this. And as that happens, then people that have commercial interests can just use it. But it doesn't have to start with the researchers. We've leapfrogged over that now. It's at the point where if you want to put a national internet in place, it is a trivial task. Jeff Houston in Australia sat down and put their national network in in 30 days with a staff of two people. Australia is now one of the largest countries on the internet. So an answer here would be jeff.hays at let's find his address, because his experience in doing that in Australia could be useful to Guatemala. He is jeff-g-e-o-f-f dot Houston h-o-u-s-t-u-n Australia at a-a-r-net that's the Australian Academic Network dot edu. So that's jeff-dot-huston at a-a-r-net dot edu. And he'll shoot me the next time I'm there for putting his email address up. Do you know his fax number? They have the entire continent is available. If you need to send a fax anywhere in Australia, you should do it via the internet. We have another question from Guatemala a follow-up. The Maya recorded important events and fax on storm monuments, which have survived to our days hundreds and thousands of years later. Today we seem to lack this long-term perspective. Aren't we relying too much on fragile computer media to store important information? And also on the internet, how are we reserving the information flowing through it for the use of future generations? Now this is she's standing in front of preservation of information in stone. How do you search stone? There's an interesting question. But the deeper question about how does this affect what a library is. Brewster. All librarians have the yin and the yang. There's the access side and there's the archiving side. And I'd say right now the internet librarians we call ourselves librarians are really heavily on the access side. The archiving side is just starting and there are initiatives going on most of the people that are serious about censorship on the internet are really working at this particular problem and what's interesting about it is it's cheap. And we're working with the Library of Congress to figure out what their role should be. Because as their role diminishes in the copyright securing one copy of every book because there's too many, right now you could actually get a copy of all electronic journal archives of the scholarly peer-reviewed electronic journals. Is that going on now? No, but yes Greta is completely concretely correct and that it's not being done it needs to be done and those that are interested in it are starting to assemble now. And when they assemble is there any organization that they form into? How do you find out more about this? Right now at the ad hoc there's the group that's coming together it's called the Library of Alexandria Foundation that I'm a part of. So if people are interested in the archiving aspects they can contact me to do a scary thing. My email address is Brewster B-R-E-W-S-T-E-R at Waze.com. So those that are interested in the archiving aspects because basically all librarians hold both tasks fairly importantly. Well this is good. There's a question I'm hearing from Argentina. Could we see the Argentinian question? Good morning my name is Marcelo from Open Systems Argentina. My ride is UWLB. It's the National Director of Information Systems in Argentina. I have a couple of questions for you. I know there is a new approach to encryption technology, the Clipper chip developed by US government engineers. Would it be available to foreign countries? And also I would like to know which other trends in compatibility between Gigabit per second wide area networks and local area networks. Thank you very much John. Mr. Del Pino let me take the second question first. The Gigabit and local area network conjunction with ATM as we heard before from Stephen. The telephone switching fabric will become the fabric for network communication. That makes the network and the wide area network the same. The atoms the cells are the same the switches are the same the protocols are the same and the goal which we hope we will achieve soon. We are now at 155 megabits with boards that fit into the S bus on the sun. We will soon have 622. I think it becomes really interesting when it's a Gigabit. The volume of that information flowing to a desktop changes how we perceive things. It changes all user interface. It is based on quite a slow link. So let me show you the name of the president of the ATM forum. This is the organization that began about 2 years ago which now has over 400 companies every telephone every major computer company every PTT, every ministry of telecommunications belongs to the ATM forum. This name Fred San Martino Fred is the president of this. He works at Sun. We were instrumental in the founding of this organization when we saw the opportunity to merge the Gigabit networks and the local area networks wide area and local area networks in one. Now the second question you pose about the clipper chip is a this is a major issue in national and international information policy and so to talk about this let me introduce Dr. Woodfield Diffie. Dr. Diffie created in 1975 and 76 Diffie Helman at Stanford the Diffie Helman public key encryption algorithms. Those of you that today use public key encryption use Witt's code well at least your algorithm. So the question about clipper availability for those not familiar with clipper this is an initiative put forward through the White House of the United States to place a chip just named the clipper into every telephone that would carry communications inside the United States government. So what's happened with this well in partial answer to that question when this was first started it was promised to various companies that they would be able to export these products unlike current cryptographic products using a present standard that's called the U.S. data encryption standard the DES when the announcements of this program finally emerged they contain the statement that there would be no change in the export control policy that is to say that cryptographic products would continue to be classified as weapons and that their export would be approved on a case by case basis. So Mr. Delpino who would like to would be interested possibly in having a clipper chip in something in Argentina for Argentinian phones would for each phone have to get a license or I would... Well I think the more plausible question is why he would want this because the clipper chip is a proposal for making communications sure against everybody except the U.S. government. Now that may or may not be a good idea within the United States but it's very hard to imagine some country outside the United States I think the government of Argentina might very well want the ability to spy on its own citizens but I think it would hardly want the United States to hold that ability on its behalf. So if we put it in a sun we'd have a label we could put on the sun that would say J. Edgar Hoover inside. Actually maybe they're being printed this morning. They say big brother inside. But now we need security. I mean we've talked about the energy and the speed and here we have enormous retrieval. How do we know when we use WACE or WEB or go for any of the retrieval mechanisms that we've retrieved something that is the document it should be? Well I think that's a very important question because at the moment we are depending on a lot of informal security mechanisms. This is very much like depending on unspecified properties of say integrated circuits. And if they change these security properties change from time to time. So that for example the introduction of optical fiber made it a lot harder to get at to put wire taps on many circuits. The introduction of cellular telephones made it a lot easier to put wire taps on other circuits. We are depending on security properties that are not promises to us in the design of the network. In designing the network nobody ever said explicitly okay these properties guarantee that the customers will have security of this kind but not of that kind. And in order to have security in the in the future and I think what underlies this is the fact that the position of telecommunications today no matter how important it is to be dwarfed by the significance of telecommunications in the future. That's just plain going to be the fabric of everyday life the fabric of government the fabric of commerce and it is going to be essential that that fabric be protected by what I call honest security mechanisms. That is security mechanisms that everybody around the world can trust the way we were able to trust the mechanisms that in fact have protected mechanisms of guarding and escorting and so forth that have protected roads and ships and so forth for centuries. This is going to take the place of the network of roads and ports and rails and we have got to have an international standard security system so that when I am engaging in a business deal in some but with somebody in Buenos Aires with somebody in Rio with somebody in Tokyo the very first time I talk to that person I can know that our business negotiations are secure and that other people aren't listening in. How do you mean honest? You use the word honest. If you are in Argentina and you like to make the phone system honest what can you do practically? I think by honest what I meant those are two questions what do we mean by honest I mean that you don't have some mechanism built in as with the clipper chip that guarantees some spies can listen in maybe not other spies but some spies you wouldn't want to put a lock on your house that you knew very well that certain people had master keys to and how you guarantee that is in fact one of the most difficult problems in modern security engineering because it is so easy to hide functions into products in particular security products that they need to make security products secure against this kind of sneaky intrusion really flies in the face of the standard desire of vendors to have proprietary products and not to tell people how they work and to be ahead of their competition in that way so I think we face a major challenge in developing a generation of products that can be sold, that people can make a profit off of and so forth that can be built where each individual group of users can if it feels the need build its own products against an international set of specifications international standards so that it is confident that there is very little place for somebody to have installed what's called a trap door or a back door or a side door that will bypass the security in a way the power of public key encryption came when you found a class of mathematical functions that allowed you to tell someone one of the functions that encrypted something and if you knew that you still couldn't decrypt it, it was a one way function well that's right and as a matter of fact which is it from conventional DES encryption yes and in conventional cryptography in order to talk to somebody securely you have to know something in common that you have somehow arranged ahead of time and both of you have to know exactly the same thing in public key encryption we have a couple of different mechanisms but the essence of this is that we have to know corresponding things but not the same thing and from what I know I can't figure out what you know and from what you know you can't figure out what I know that's the that's the sort of second stage of it we had a first stage of it which is what you so kindly credited Marty and me with inventing which is a mechanism that allows two of us to negotiate between ourselves something that we know that nobody else in the world knows even though the negotiation was conducted in public we could talk right here in front of the TV cameras and I would give you something that's very much like giving you object code but I would keep the source code to myself and you would give me something that's very much like giving me object code but you would keep the source code to yourself and because I have one set of source code and one set of object code I could combine the two to get something and because you have one set of source code and one set of object code you could combine the two and get something we have the same thing somebody else who only has the two sets of object code will be unable to figure out the thing the two of us knows. Could we use this in conjunction with DES to provide security? Well, that's right. As a matter of fact, that's exactly what this gadget here does. This is the pre- This is a show-and-tell show. This is a show-and-tell. I hold it as high as I can. This is the pre-clipper version of something that AT&T brought out. And in some views, this precipitated the announcement of the clipper initiative. This is a device that engages first. Note that it fits between the telephone and the handset. It's a fairly small device. Could get smaller, but it's about the size of a small box of cigars. Your sister makes a secure phone call. And my jello, the fact that we don't have either the line cord or the power. We're doing pretty well. We have known people in the computer industry to hold up boxes and wave them. Well, we'll wave this box. Now, what happens when you are engaged in a phone call and you push this red button on this box? There has to be a similar box at the other end of the phone line. And what these two boxes will do is they will negotiate with each other for a few seconds. The first part of the negotiation is modem agreements. They have V32 modems in them. The second part is this Diffie-Hellman key negotiation algorithm. And at the end of that, they will know something in common that nobody else, no other box in the world knows. And then they begin to use DES to do the encryption. So this is something. Now, this box existing made all those building clipper chips desire to plug the clipper in here. But there are other ways that the world could achieve common encryption standards. Well, I think that's right. I think this is really in the area of telephones. We have all of the basic standards we need in de facto existence. Some of them are formally existence standards. V32 is the worldwide modem standard. And this particular thing is generally the 4.6 kilobit speed is what's accepted for secure voice. Then there's US federal standard, federal standard 1016, a public standard, that is a linear, predictive, coding, voice digitization standard. And then there are the public key standards. And this uses one of them. In fact, in the secure phones we hope to see come about as a worldwide standard, we hope to see two different sorts of public key technology used. One for protecting privacy as in this device, and one for identifying the person who is making the call. The mention you make of standards, let's bring Larry Irving. He's in Washington. Larry is in charge of the element of the Department of Commerce. And I see him on the television set. You're looking very distinguished today, Larry, I must say. Thank you, John. We've knocked the polish off here in California. We miss you in California, actually. You should come home once in a while. Larry is in charge of the element of the Department of Commerce, the National Telecommunications and Information Administration, that is the point, the major element of the Department of Commerce for the national information infrastructure. So the document, which has been published, the committee that Larry chaired that wrote the document entitled the National Information Infrastructure Agenda for Action. I have a paper copy here, so I'll just hold it up. This paper copy is available from Larry. Where can I turn this so you can see it? There we are. There we are. This document, you'll receive it if you just FTP it there. Carl, you have a copy that we just pulled off and printed yesterday off the internet. And there are hundreds of FTP sites around the world with this document. Let me recommend it to everyone, because it has in it exactly what it says, agendas for action among them, provision of support for standardization. And so coordination with other levels of governmental and with other bodies international on these standards, provision of seamless interactive, user-driven operation of the network. Larry, I wanted to ask you, when you prepared this document, it cites a lot of projects. There's a National Information Infrastructure Pilot Projects program. And this is supposed to allow people to get on the net. How is this going? What stage are we in in this? Right now we're in the stage where Congress is considering it. We have the Appropriations Committees are looking at it and the Authorization Committees and Congress are looking at it. We're moving forward, developing regulations, trying to reach out to state, federal, local people who can help us push this project through. We hope that Congress will approve it rapidly. We'll hope we'll be able to get those pilot projects out. But it's going to be coordinated, not just NTIA, which has some of the pilot projects, but other federal agencies, the Department of Education, National Science Foundation, NIST, and others. We'll all be working together to make sure that we're learning the right lessons from these technologies. You've been sitting in the committee meetings with all these people, and are they eager to move forward? There's a major element in this about putting the governmental information databases on the internet, or at least to make them network accessible. Is this, are people ready to do this? What about the CIA and the other agencies? Are they interested in this? I haven't had any direct conversations with the security agencies such as NSA and CIA, but I have had conversations with the folks within the Department of Commerce, people at OMB. The President, Vice President Gore and Secretary Brown are eager. They are champing at the bit to get more information out to the American people. This is an information aid. It is an information economy. The way we can help keep ourselves globally competitive, the way we can create more jobs in this culture, I mean, this economy is to get this information out to people. And the President, through an OMB directive earlier, has made it clear. The President, through his gender for action, the Vice President, in everything they're doing, national performance review, reinventing government. One of the ways we're going to reinvent government as an administration is to get this information that people want and need and can use out to those people. This is when you mentioned for those outside the United States, the OMB is the group that controls the money, the budget group inside the executive. Is that am I correct in this? They're budgeteers and managers. They really are the President's managers of the whole federal bureaucracy. And so any federal agency will deal with OMB, both in terms of our budget, but also in terms of how we manage ourselves. And the way the government is coordinated, generally goes to OMB. Quitted brought up some issues about security on the net. And there are some provisions in the standards process in this document to examine how security can be provided. Are there task forces? And would people watching this program around the world be able to send their ideas in or come to meetings? There are two principle ways to get information out about security. Let me start by saying security is important to this nation with regard to development infrastructure. People don't believe that their confidential business or personal conversations and communications are secure. They're not going to use the infrastructure. We have to do something with regard to security. But we do want to hear from people. We do want to hear from industry. Under the President's Information Infrastructure Task Force, and there is a committee that is chaired by a woman at OMB named Sally Katzen. And they're looking at information policy questions. Sally Katzen is one contact. But also, the National Institute of Standards Technology, headed by Dr. Arthi Prabhakar, is also involved in this entire security issue, particularly the clipper chip issue. And people who have questions or concerns should get in touch with Dr. Prabhakar. I can get the internet address for her. And we also have telephone numbers for her. And I'll get those to you so that our audience today can know how to contact them. That's great. So can we hear Sally Katzen at OMB? Does this work dot USG? Doesn't work. Well, let me get you the exact address. I think I have it in the back here. Sally's going to kill me for giving this address out. But let me give you what I have. I have a fax number right now. I don't have an internet number for her. So you can fax that 202-393-395-3047. And Arthi Prabhakar and this can also be contacted on the internet at arthi, a-r-a-t-i, at m-i-c-f dot n-i-s-t dot g-o-v. OK, everybody in the world's writing quickly here. So what we're going to have to do is get this on a graphic display. And we'll put it at the end of the program. We pile all this information together onto an FTP-able archive for synergy. And people can get that from around the world. So we'll get all this straight. So what kind of question would it be appropriate for Sally? She's at the top of all this. She's going to be heading a committee. And they're really just starting their work. They're looking at information policy. They want to know that they are helping industry and government decide what the right issues are, the security, privacy, copyright, all of those issues that are so key to getting this technology deployed rapidly. Letting people know that if they use technology, it's going to work for them the way they want it to work. And Sally is my colleague. She's got the entire federal government working with her. Every federal government entity that cares at all about this infrastructure initiative is part of the president's task force. And all those people care about information policy, including network reliability, security, privacy, and others, I'm sorry, security and copyright and privacy are working with Sally. So there's a great opportunity if people have their voices heard. We want to hear from them. This is great. I mean, this is a change. No one's ever known the anonymous. You're not anonymous. But the people that are inside the organizations that spend their days, long, long days working on this are generally not known to people outside. Perhaps they don't want to be because it's just our problem. How much email do you get, Brewster? No, 120 messages a day. 120 messages a day. Well, that's getting up there. So this is good, though. This is letting people know. Now, for those that are outside the US, what are you doing in linking the national information infrastructure, which is important for the United States, with what seems to be growing even more rapidly the global information infrastructure? Singapore, Germany, France. People are moving very rapidly to establish and have established interconnectivity among citizens of those countries. Do you talk to them, or what's the relationship with them? We are working closely with them. Dr. Prabhakar and NIST work with NTIA and others in the government on domestic and international standards. We also have an international policy committee or subcommittee under the telecom policy committee. We are going to work with anybody domestically or internationally who wants to work with us. There are no borders anymore. Information gets back and forth so rapidly. The only way the system really will work is if we are in cooperation with our friends in Britain, our friends in Germany, our friends in Japan, Singapore, Guatemala, Argentina, wherever people want to connect to people in the United States, we want them to be able to do it. We want this to be a seamless, interactive, mobile network. We're going to start our principal mission to make sure it's working here in the United States. But it only really works well in the United States if it also works across the globe. Let me give you Sally's number one more time. I understand that you may not have gotten it the first time. It's 202-395-3047. It's our fax number. And I guess you should be prepared. I'm going to call her as soon as I get off camera to let her know to expect some calls. All right, we'll give her a breathing spell before the flood starts up here. It's 202-395-3047. 407, exactly. That's a... I do recall when the Clinton administration took over, trying to call them, it was busy all the time. So I think we're going to have a busy fax machine now. So bypass it. No, it's... She'll be on the next one, right? Now, actually, what is the provision? Does Bill Clinton have an internet address on his card? You have on your card, but does every... How is this working with the existing government that's been there before the new administration came in? We are rapidly moving to an internet-friendly administration. The president does have an internet address. I have an internet address. The Department of Commerce is moving rapidly to get all of us connected to the internet. It's... We have found the utility of it. The reality is when I was over on the hill, the internet... And I left the hill in March of this year, Capitol Hill, I was a congressional staffer. We didn't have internet access. We were not connected. Since I've gotten over to the Department of Commerce, I found out how important it is. What we... Last, in September 15th, when we released the NII agenda for action, we were able to release eight times, eight times the number of this document using internet that we were just through hard copies. So we had about 1,000 the first week that we went out and distributed by hard copy. 8,000 people got it off the internet. I think that's very, very important. And it's significant, and it's an indication of just how useful these new technologies, new tools will be in getting information out to people. That's the kind of thing the president wants to say. This is great. I think you're underestimating. 8,000... I presume what you know is that 8,000 primary requesters took it off the internet by FTP. Exactly. Those copies were probably multiplied many times since they were in digital form and could be reproduced trivially. Yes, what it brings up is this very interesting phenomenon. Everything moves and moves and moves and moves. One side of this world will depend upon advertising revenue. Advertisers never believe the numbers you give them about the millions of people that you're reaching. And so in the internet, the provision of some accountability for this is going to turn upside down. The current Nielsen and ways of people estimating the reach of a piece of information. It's going to alter considerably. I think, well, that's... I think, Larry, can you hold on with us for a bit because we want to show a little bit of interconnectivity and then we have the world wanting to ask you questions and ask everyone here questions. What I wanted to show people is on the net now how we can do the teleconferencing. We're doing it right now with Larry from Washington by coming down and C-band and going up and KU-band and bouncing all over the country. But why do that if we have lots of fiber running around the country and we just put the video on the net? So let me... All right, let me go over here to where we have machines which are going to show this. Don, are you here? There you are. Good. My microphone. Do you have a microphone? Okay. Good. You have a microphone. I have a microphone. Okay, basically what we're doing is I'm going to show you some of the tools that are in common use on the internet for doing remote teleconferencing. Hold on just a second, Don. This is Don Hoffman. There's going to be on the screen a telephone number for you to call here and you already know how to do the fax business. Can you send questions? We're going to have about 20 minutes to ask each of the members of the panel what's going on and what do they mean by this and how could we reach this piece of information? So the phone number you're seeing on the screen, call that up, give us questions. All right, that's the advertiser for this. So what are we seeing here? I see here's a screen and I can see, oh I can see some faces moving here. Basically what we're doing is I've put together a small community of people for the demo here. Typically on the internet, you may find conferences may involve large numbers of people, particularly if they're audio only conferences or conferences where so you're having a lecture style of interaction. For instance, Palo Alto Research Center, in fact I'll maybe switch over to this window over here. Let me just point out something I've been told because we're speaking quickly and because this conversation is going in Spanish to all South America, we've been overwhelming everyone that has to expand these short phrases into longer Spanish phrases. So how would we expand the multicast backbone audio, in bone audio in Spanish? I don't know, I'll hear later, I'm sure. So what do we have here? We have a long list of conferences going on. So these are sessions, this is directory of people who are broadcasting on the internet. Correct. The tool was done by Van Jacobsen at Lawrence Berkeley Laboratories. It's called SD for session directory. And so basically what we have is a list of currently active conferences. The way it works is if somebody wants to advertise a conference, they bring up a version of SD and then hit the new button down here to create a new conference. So this tool is the television guide, the reader's guide to what's going on in the air, all right? So what we have here is I'm gonna do very quickly point out a sample of what's on SD here. I have the cursor already pointed to something called M-Bone Audio. This is almost like a chat channel, a large number of people involved in research and multimedia over the network. And in fact, if we real quickly open up the window here, you'll see a list of people, this is a list of people that are currently there. Oh my, now here, so everybody here is sitting in front of a computer someplace and they talk and it's like a party line. Everyone else can hear them. That's correct. So we have on one conference, I see Ron Frederick, I see Van Jacobsen who wrote the code for the video on the internet. Who else? Yeah, Van we did the part for the audio and as he, Ron did the network video. Ron did the video, sorry, Ron, I'm giving credit to Van here for this. Did I just see? You notice it's also fairly- John Postel is on the net? Well, the god of the internet here is sitting talking, chatting. He doesn't have time to chat. Get back to work, John. You notice that the participation is fairly international. If you look for instance, we have a person from the UK, you can tell by the domain it names. Somebody from Japan joined in. Somebody just dropped out from University of Queensland. Person at Edinburgh, UCL London. So fairly international. Oh, Lord. So basically what we're doing here is sending this, it's an overlay internet over the internet. So basically pieces of the internet that have sufficient bandwidth that you audio and video support something called IP multicast and then we use those protocols to communicate among them. But how does someone sitting in front of a machine get the IP multicast code that put in there? Okay, at the end here, I'll be able to bring up a window showing where the FTP servers are before we sign off. Great. So this could happen at commerce anywhere. So suddenly we're going to be able to give you a fax path to someone and a chat path. So every, let's say Friday afternoon, around three, we'll just open up the channels and the world will chat with you. Oh, I love it. Okay, so now let's, I see video here. I see myself actually in a little window. That's the small video camera. So why don't we close this over here and move over to the other quadrant of the screen here. What you'll see then is another version of the audio tool again of that. This time with a smaller group of people. We have us here at the studio. John Austin who'll be joining us from the son of North Carolina and Michael Spear who's down the street at Sun and Mountain View. All of these people are all remote from here. We're reaching them using T1 links. The local T1 links provided by Peck Bell, the remote T1 link. I don't really know who we get that from. So John Austin is in North Carolina. And to get from here, I wish we had that network routing tool. We could see exactly how the packets are getting, all they'd be split up and going in different pathways around the country. But anyways, all our communications are just packets, just bits headed across the net. And they're gonna go, they could as easily be going to Japan or to not the line is up to Singapore Yes, okay, so let's try this. What are we gonna do? On the video site here, what we're gonna do is we'll briefly, let's start talking to John a bit. John was, in fact, let's quit it. Gonna make your microphone live here at this point. My microphone. Why don't you say hello to John there? Hello, John. John's thinking. Oh, here we go. Yeah, really. You have to speak up a little bit. I can't hardly hear you. Ah, okay, all right. Let's see, we have a volume control, all right. Can you hear me now? It's fine. How are you doing? Fine, your mouse not moving. There it is, all right. Great. Now John was responsible for, his group was responsible for a lot of the hardware, particularly the video card we're using. And maybe he'd like to say a few words on some of the architecture. Yes, how complicated is the hardware that's making this picture, putting this picture together from the internet? Is it a big board? Is it, well, how soon will this be a $5 ASIC? Because we're always from a $5 ASIC, but the board is just a little S-Bus board right here. I'm holding it up, maybe a dozen or so chips. And what's the integration going to cost? I mean, is it a long time? To answer the question again, sorry. Will it be a long time to integrate this into something on the motherboard? It's gonna be a matter of when the cost comes down and the usage goes up enough that it's prevalent enough that enough of our users are interested in using video. I can see as I talk to you, there's a little meter that's just showing the packet traffic. Is that showing? Basically that's a cool one of a VU meter. And let's bring that window up to front here. So as you go ahead, in fact, my voice is showing up on the VU meter as I talk. You can see the meter going up and down. Also, you notice as I speak, our entry becomes highlighted in the entry there. Now, the tool uses various different mechanisms for doing what's called floor control. In this case, I'm using something called push to talk because of the open speakers in the room. But in some cases, you might say use headset microphones and that supports certain forms of echo suppression you can use in some cases. Is there code, assuming that it was hardware support, is there code that would allow a Macintosh or a PC to participate in a conference like this? Now, most of the tools I've been showing here work under X-based workstations. Oh, so just it's an X protocol. So pretty much most of the workstations that support some sort of audio support X-Windows can support the tools you see here. Now, there are versions of those tools available for, say, the Macintosh platform. There's something called CUCME done at CMU and Ron has provided gateways between the NV video protocol and the protocol used for the Macintosh. When you say Ron, you mean? Ron Fredericks. Ron Fredericks at Palo Alto Xerox. The Palo Alto Research Center at Xerox. So the NV tool we see here can talk not just to other NV applications but also to the Macintoshes. Okay, great. Currently I'm not aware of any PC applications but I know there are fair number of people working on that sort of thing. Oh, there must be at some point. Well, this is great and you actually look quite clear, John, is there, now, if we have higher bandwidth, this board will go, how fast will this board go? If we have an ATM board, we have the ATM 155 megabit boards now that plug in the S-Bus. Is there any interconnectivity yet between that speedy data flow and the board that you just held up? I'm sorry, I didn't quite hear the end of the question. The end of the question was, if I have 155 megabit ATM coming in on an S-Bus, that's the board that Jeff Baer made and we have in thousands of suns now around the world. If we have that data speed coming in, what will it require to get video across that up into this set of tools? I think it's flexible enough that it's just a matter of matching the video bandwidths that we have. We'll just compress the data less, have higher quality, have higher frame rate to match the network capabilities that ATM is going to bring us. Good, I can see now, we've just brought up, I didn't realize this, but I can change your colors here, John, I can, oh and okay, so the bandwidth of this tool is very interesting. It shows that we have a current bandwidth of 400 and there it went, 400, about 400 kilobits per second is carrying this quality image. That's really quite remarkable. Why don't we ask John to increase the frame rate a little bit for a moment? Can you increase the frame rate? John, can you increase the frame rate? I can adjust the bandwidth on my side. I'll go ahead and increase it somewhat. Oh, so it went down, we're 288 kilobits per second. I don't know, John, in this implementation, when I increase the bandwidth, are we increasing frame rate? Yes. Yes. This is interesting. I'm taking it down to about 100 kilobits right now. Now can you take it up as high as you can? At slower bandwidth, you have a tendency to divide your head in several pieces. That's not bandwidth, that's just typical. Oh, great, there it goes. Well, I think we're getting the idea here with people participating in teleconferencing like this, which can be done across the internet for nothing, essentially. The more people that do this, the better we'll understand how the tools should interact with people, how we can make the normal cooperative work environment transmit itself through this mechanism onto the screen. We should show you one more time, the telephone number, so that you can call for questions for the panel. And where to get the software? Oh, yes, we have on the list the FTP sites for all the software, the X-Windows software, all the software you see. I don't know if we have the site for the Macintosh, but we do, okay. No, but not up here, we can. They're not up here, but we'll add it to the FTP. When you come into Sun Energy, the FTP site will have all that information there. The phone number for the question and answer period, the Q&A, is 415 in the United States, 415-988-1171. Thank you very much, John. We're gonna go back and ask questions now. You know how to send faxes on the internet, so it's not as nice as a video, but it will be handed to us. Thanks a lot, John, talk to you later. Bye-bye, Michael. Thanks a lot, Don. Good, we'll go over here. The first question we have, while the faxes are coming in, is a question from June Murai. Those of you that were at Interop saw June, and those of you that have been on the internet a long time know that June created the academic network in Japan, linking the national academic libraries together, allowing people multilingual access to library information years and years ago. Let's see June's question. As an international language on the internet, in future, okay? So there are a lot of people in the world who's got a very good idea, but have a little bit of problems representing their ideas in English. So is there any way we can solve this problem with the technology and the internet? Yes, June is being uncharacteristically polite in this. What he's saying is English is not the only language smart people use. And what is the technological change in the internet to allow all languages, everyone, all literatures to participate? And I think this is for you, Brewster. This is... Getting multiple languages is completely essential towards getting worldwide acceptance of this. What we've got is standards based on ASCII, which are just ancient standards based on U.S. character sets only, so you can't even represent a tilde, much less kanji. We're struggling with trying to move forward into Unicode, there's patchy standards, all around people are starting to pass around images of pages just because we computer people can't get our acts together enough to get the standards in place. Even Unicode is complained about by the Japanese as not being able to represent Japanese well. So maybe it will be images of pages and we computer people won't be able to to be able to get the character sets together. One thing that would help is to be able to search that material even though the documents are in page images, be able to search it in this native language. That's now coming about. So maybe, just maybe, instead of getting a lot of the next level of ASCII, we'll get page images that are searchable so you can find the right ones and you'll get all of the doodles that anybody wants on it. Well, this is good. So be able to search English and Japanese and even legal, American legal language. Could we search that? I mean, that's- That's impossible. Is that impossible? That's technically unfeasible. Larry, can you help us on this one? Is Larry still on the wire? Still here. Great. There was an interesting flap recently about the audio records of the Supreme Court, the oral arguments. Some, a professor in San Diego apparently taped these things in the archives, the National Archives, and is selling them. It brings up the broad question of access to records inside the federal government that used to be constrained to a few people in a few rooms, which now seem to be available. If Carl Malamud can go to the National Press Club and put the voices of General Colin Powell, I guess, Wednesday, tomorrow, goes out across the world network, is there a chance we'd have any access to the federal depositories of this? We're trying hard to figure out exactly what those rules will be. And again, that's what the Information Policy Committee will be looking at. We want to make available to the American people those pieces of information that don't have a security component to them that are set in repositories around this country to the extent that we can get them out to people we will. That is a priority of the president. We have to come up with some new rules. We have to come up with new thoughts and concepts. But this is something we're working on and I think we'll have some good things to say about it later this year and definitely over the next several years. A fax just came in from Peru, let me read it. There's a new Peruvian academic network that's connecting all universities and research centers across the country. I'll translate it. It's the Peruvian Network of Science and Technology known as RPCYT, the Science and Technology Net. Question, which is the most economical and reliable media, or I would say mechanism, to connect the RPCYT to connect Peru to the internet. Carl, what do you think? Well, there's a variety of programs that in fact the US government has provided some very good leadership not out of our information agencies like NIST and NTIA, but actually out of our mission oriented agencies like the National Science Foundation working on behalf of US education and ARPA and NASA. And those groups have programs like the International Connections Manager Program and they've been very active in bringing other countries into the internet web. Now let me say that you don't have to go into the US government backbone. It is not the center of the internet and there are a variety of ways that Peru can connect itself into this worldwide web. How much of the internet's paid for by the government? The estimate that Dr. Bob Kahn, who was the ARPA program manager who founded the internet many years ago and he's now a distinguished researcher. He's president of the Corporation for National Research Initiatives. His estimate are that maybe 5% of the internet is federal money in the United States. And that's always been true. There was a myth that the internet is in fact a federal government project and some of the first internet users were DEC and IBM in the early days. They were putting their mini computers on and the vast majority of the internet is now paid for by its users. They're directly on a per fee basis at the end user or by their organization such as their campus. Larry, is there money in the billions you have to spend? Well, I'm exaggerating this a little bit, but you have a significant amount of money to spend. It's not, as Carl's saying, it's not the foundation, the internet's grown so large, it's not the United States government. Some links and very important ones are funded by you. How much money do you think will be available to reach the schools and libraries, to reach the elementary schools in the country? It really is a modest amount of money, actually. We're looking at a project this year, a program this year of about 30 to 50 million dollars. The president asked for 50, the Senate and House right now are looking at monies in the 20 to 30 million dollar range this year. A lot of that money will go for planning grants and then next year it'll grow a little bit more. Again, the president's asked for 150 million. We're hoping to start growing up to that level. That's not a lot of money given the number of schools, libraries, social service providers, hospitals, museums that we hope to work with. We hope, one, that we can leverage some of that money. That there'll be some non-profit money, private sector money, state and local grant money that will be combined with the federal money. We also hope, however, that what we can do is demonstrate to states and localities how important this technology is. And they'll undertake projects on their own once they see how they can benefit the citizens of their communities by these technologies. So part of the pilot program and the demonstration program is to really get the message out. It's an educational component as well as a connection component to these pilot projects. Brewster, you had a comment? Yes, and some of the government information, getting the information out, the networks do cost a lot of money, but getting the information out is often even cheaper than what they're doing currently. So I find that with the federal agencies we're working with, the most expeditious way of moving is to cut their budgets on information dissemination and that will force making better use of the information technologies that we have. A lot of these technologies are in the tens of thousands of dollars rather than the tens of millions. And so dealing with the executive branch, the judicial branch, a library of Congress, the patent office, often they're gearing up these massive programs, government printing office, massive programs when they could get by with very inexpensive technology and do it today as opposed to planning on spending tens of millions and doing it in a decade. I just heard Larry, I just heard a volunteer here to come sit on all over the long, you know how thick the budget is to sit there line by line and see what you think. Well, I think we have historical precedent. It seems to me we could argue that the reason we got to the moon in 1969 is because of the cutback of the budget in the mid-60s that forced them to focus on the moon and stop thinking about monstrous nova rockets that would. And this isn't rocket science. Let me say that there are some easy things. The government can provide the best leadership not by issuing reports but by healing itself. And for example, in commerce the securities and exchange commission is now spending $13 million to buy back its own information from meat data. And there are examples where that would be a very simple ways database. That's 20 gigabytes of information. It's currently a multi-million dollar project and the end result is government information is being sold at high prices. And these are examples of areas where the government can really heal itself and save money for the American people. I think in fact I read a segment in the infrastructure that mentions just this form of analysis. But I have a question coming in and then we'll let Larry jump in on this. The questions from Sylvia, I wrote it down here someplace. It's from Sylvia. Sylvia is in Englewood, Colorado. She's on the phone at EDI. Now is this electronic data interchange, Sylvia? What is EDI? Yes, it's actually EDI Systems Integration Company and I'm based in Englewood, California. California, sorry. Two short questions. The first one is I would like to know Sally's last name. Larry, Sally's last name is? Katzen. Katzen, K-A-T-Z-E-N. Okay. The second question is how will the various networks that government agencies such as IRS and HICFA are establishing fit into the total NII architecture? A great question. Anybody up to tackle this one? Hopefully with some of what's... Let me repeat the question. How would the IRS net and in fact the many separate nets in the federal government fit into the overall NII architecture? There ought to be a principle which is if there's information for the public it should be available on our public network and that doesn't mean that the internal IRS computers should be in any way connected to the internet but certainly a piece of every government agency's mission is to give out information and that piece should definitely be on the internet. Good, I'm hearing in my ear that Peter is on the phone line from CITI and where is Peter, is Peter are you on the line? Where are you Peter? I'm in Ann Arbor. And what is CITI? It's the Center for Information Technology Integration at the University of Michigan. And ask away. I'd like to know what's being done to integrate security and authentication mechanism into information location protocols sort of integrating what Brewster and Witt have said. All right, the two of you seem up on this. You locate it, you do security. How do you integrate this? Actually the ball is really very much in the court of the tool builders. The security people have generally done pretty good job of making some of the tools available. The two major ones that are available on the internet are Kerberos and public key encryption. Both of them are weird. Both of them have real problems to them. The Kerberos structure requires real time checking of authentication for each session. Requires a network of fairly high integrity. And often when you're finding you have Novell networks connected to Apple Talk networks connected to the internet and it's sort of stringing around you don't necessarily have that. But it's a pretty good code base and it's freely available from MIT. Public key encryption is more mature and it's actually more the right thing. The problem with it is it's patented. The group that's doing it is very flexible about how they license things, but they have a license. And most of the people on the internet are dead set against licensing any kind of protocol. We're working to do both the Kerberos and the public key for the ways and that would extend to both some of the functionality of web and gopher and trying to figure out how to work those in to those different interfaces. Now, so I'd say in a year we will have at least one of Kerberos and public key built into that particular, the Z39.50 based systems. So the work is going on. It's not necessarily getting funded very heavily, but it's certainly in demand mostly from the commercial sector. What do you have anything to add? The, well Larry this, the merger of these two in the committee meetings, which as you say the work is beginning now because people have to come together have not come together before. I think this question does touch, it's the merger of these two concerns, finding it and finding what you want, trusting what you find that should be on the agenda in these meetings. Well, that's my editorial opinion here. I'm just throwing things in. I have a question that's come by facts from Richard Keane at NCAR, the National Center for Atmospheric Research in Boulder and the question, and I like this, rather than the usual date and time of facts, the date is now, the time is now. So Richard's wasting no moments here. How does one send an image or post-crypt facts over the internet? Well, there's a whole variety of standards that deal with multimedia data types over the internet. There's a set of standards called MIME, which is multimedia messaging for the internet and MIME includes different body parts. It's the thing that X400 tried to do, but X400 forgot one important piece, which is to write some code. And the folks at the IETF actually sat down and developed some working code. There are several interoperable implementations, several commercial version shipping. Waze is a multimedia environment. The World Wide Web is a multimedia environment. Our new fact service accepts post-script code and TIF images. And so we now have a multimedia remote printing service. Well, that answers that question. I have another question that's come in from the University of Michigan at the Medical Center. Once a document is searched and located, do you foresee users being able to edit the documents using the Waze interface? And it goes on, will there be an ability to tag documents with electronic signature, which is this merger of the document and authentication mechanisms? Generally the way people do it is they download the document and they can do anything they want with it. If they go and republish it, then copyright law can come in and help stomp that out. But the idea is not to lock a document so that it requires encryption hardware on every PC to be able to read documents, but to do it more like the way books were done. Do you control the distribution of the work at the publisher or at the library and offer easy access to it? With the new URL, URN structures, you can pass references around. So instead of actually having to pass the whole document around or zeroxing it and passing it to somebody else, you can just pass a pointer, then you double click on that and you go back up and get the original resource. If it costs money, then you might have to pay the 10 cents, the dollar that it costs to be able to get it. Electronic signatures are further away, but knowing who stands behind a document is important and that's often where it came from and being able to ask what the copyright disposition of it is. Larry, this does sound like there's enormous turmoil in the law surrounding copyright and intellectual property. What element of the NII, the National Information Infrastructure, is dealing with this? Again, it sounds like I'm throwing everything to Sally, but this is again the information policy group. They're gonna look at those issues. One of the things that was raised, if somebody can go in and edit a document, there are some questions about altering works and what are the copyright implications of that? It's one thing to appropriate and we know there's copyright liability there. It's another thing when you take it, alter it. What's the copyright status of a document that's been altered? We have to look at those issues and the Information Policy Committee of the IETF will. Sally's got a very, very full plate. We'll be looking at some of the other issues, including applications and Arthur's committee. We'll be looking at telecom policy in terms of what we have to do and to build out this network, what structures we have to look at. From a regulatory perspective, we'll look at those issues, but Arthur, Sally, and I all have a lot of work. Fortunately, we have the backing of the president, vice president, and virtually every cabinet secretary, including Secretary Gore, I mean, Secretary Brown is gonna be very, very closely involved. Vice president, president, and secretaries of every department want this to work. We're gonna have a full plate, but we think we're up to the job over the next three, four, six, eight, maybe 16 years during the Clinton and Gore administrations. Why are these numbers multiples? These are multiples of four, I hear. Multiple of the four are very important in a political process. Now, behind you in the picture, your background shows the dome of the Capitol building of the Congress. That's a multiples of two building over there. And six in the Senate. And six in the Senate. How are your relations with those people who actually have to pass the legislation to make these changes? Well, I spent 10 years of my life on Capitol Hill. That was my residence before I moved over to the Commerce Department. We think that the Senate and the House are ready to work with the administration to do what needs to be done to get this infrastructure deployed. We will be later this month or late in October or early November, there'll probably be an opportunity to testify in Congress as to what the president wants in legislation. And we're working very proactively and very cooperatively with the Congress on doing the things that need to be done. There's a question that's just come in from Juan Carlos at EMSCA, EMSCA in Caracas, Venezuela. Juan Carlos, go ahead. Hi, John. Glad to talk to you again. I have a question really for the whole group. If I were the telecom manager for a telephone company in one of the developing countries in Latin America, what would you recommend for me to use as a technology, given that I probably have a pretty clean slate and I can start with digital, cellular, ATM, ISDN? Could you give me some recommendations about which way I should go? Now, what problem are you solving? You're establishing in a building or for a company or a government, a telecommunication system? This will be for the telecommunication system for a wide area to cover the whole country to offer this service to companies. Well, do you want to tackle this one? You go to the supermarket and if you're in the Bay Area, you go to Fries or one of these other places and you get out your shopping carts and you buy a couple routers off the shelf and there are several dozen vendors and you go to your telecommunications supplier and you see what's most cost-effective. If they let you lease lines, you go for those. If they don't, maybe you have to use X25 or something else. If they have ATM or ISDN available, go for it. I mean, but really, this is an off-the-shelf problem. It's like, how do I put a network of PCs in today? You buy the PCs, you buy the Ethernet cards, you buy some cable and you install it. And wide area networks have reached a point where they now can be done that way. All right, now this is, in a way, Larry's been doing this. He does not have an open slate. I mean, you don't have everything's there already and there are huge contracts in place for providing telecommunications for the federal government and there are immense changes in the regulatory environment with the telephone companies becoming cable companies by sending video over the phone wires and the cable companies becoming phone companies by getting dial tone across the cable. How is this all evolving? Is it total chaos or do you see a pattern emerging that would allow a national policy to establish standards or some level that the elements of the infrastructure should reach? Larry, are you there? Oh, I think that we have a great opportunity here. I don't believe it's chaotic. Legislation is always somewhat of controlled chaos, but this is probably the one great opportunity. Industry wants to build this infrastructure. The administration wants to get out of their way and permit them to build infrastructure, but we have to put the human components in place. We have to protect the consumer. We have to protect the creative artist so that his copyright or her copyright product is protected. We have to do those things that make sure that consumers and the creative process works. We have to ensure our universal service commitment is honored as we get these new technologies that will improve educational facilities, improve medical applications. We have to make sure that all Americans are part of that system, but we don't expect the administration to build it. We expect the administration and the government to help industry build it by getting out of their way in most instances and only being there where there's a real role to protect the consumer. And one other thing, we take very seriously the administration, Physician Heal Thyself, the national performance review, the Vice President's Gore initiative to reinvent government has looked very carefully at how we can make sure more information is available to the American people. They own this information. How do we get it out to them? And how do we get it out to them at a reduced cost? We want to save trees. We want to cut the cost of getting this information out to the American people. And that is going to be a high priority for every federal agency. What my colleagues and your guests today are saying on this panel is absolutely right. We are not going to forget the lessons learned through the 60s and 70s and 80s, that there are all ways to do things cheaper and more efficiently. Well, that's good. Thank you very much for that. I understand that you have to go. Thank you so much for joining us on this program. You're going to be receiving, you'll be sorry, Larry. You're going to be getting mail from everywhere. The ability of people, the question from Caracas strikes to there's a new ability for everyone to be interlinked. I think this unification of the world telephone standards on ATM has come several years earlier than anyone thought. So instantly, we have high-speed communications globally, and that brings many of these policy questions to the force. You have a tough job. I must say, in reading this document that you came out with, the agenda for action, it's very well done, and it incorporates, I think, most of the things that we've been concerned about for a long time. So thank you for participating. Thank you on behalf of all my colleagues. Thank you. The state we are now in the show, we're coming to the moment when they turn the satellites off, so we should probably give you the information we've promised you. There are a lot of FTP sites. There are a lot of places that you can reach. Of course, you can use Waste Go For The Web to find these sites. Read on the screen the place you would send mail. Now, this is a person volunteering to receive worldwide mail. Catherine Webster at corp.sun.com deals with the world of electronic libraries. If you're involved or interested in the Bibliotech de France project, the new Arabic library going in Morocco, the interchange of information globally using, for example, the Z3950 standard that Brewster mentioned, Catherine will answer your questions, mail Catherine. We have more addresses about information. May we see the next list? If you're going to reach us, that is, for all those lists of names, Larry Irving sending us the additional fax numbers and email addresses, we'll add these to this site, the Sunergy site. And you mail sunergy at sun.com. And you, in turn, will get back the book lists, the things that we may not have time to go through in any detail on this broadcast. To FTP documents from this program, there are the FTP addresses. You can have access to all this information. And probably by the time we finish this program, it will be on the net. Maybe it will take until tomorrow. So I'd like to thank everybody for participating. As I mentioned at the beginning of the broadcast, the documents you see there that are FTP-able the next time we hope to be sending in the television signal in the vertical blanking interval. Now, question for you all. How can we make this secure? How can we make this searchable? And how can we make this carry even more information than there is there right now? Are there any final comments you'd like to make on what you've heard? Silence around the audience. Well, I think Brewster raised an intriguing possibility. And I envision it's going to have a big impact on publishing. Right now, we talk about a book that's called A Page Turner. We're going to have one called A Reference Digger because it gets them to dig for the expensive reference. Sounds great. I'm hearing where we've got to go. Thank you very much. Thank you all for participating. Thank you all in the audience. It's been an interesting time for me. I hope it's been interesting for you. And please, send us in for me. Send us mail. Let us hear what you're doing. We'll see you all on the internet.