TechTalks Transcript
Video Over the Net: Where, When, How and Why
October 18, 2001
Audio
• Streaming
MP3
• Download
MP3 (Download
Tips)
 Judith Boettcher [JB] |
 Howard Strauss [HS] |
 Bob Dixon [BD] |
 Jill Gemmill [JG] |
JB: Welcome to the CREN Tech Talk series for fall of 2001 and to this session on Video Over the Net: When, Where, How and Why.� You are here because it�s time to discuss the core technologies for your future campus. This is Judith Boettcher, your CREN host for today, and our session is coming to you with the support of the CREN member institutions and our sponsor for today, Tandberg, who provides distance education videoconferencing solutions over both IP and ISDN networks. I�d like to welcome Howard Strauss of Princeton today who is our technology anchor for Tech Talk, and as you all know, Howard is a well-known web technology expert and portal expert. Hi, Howard, how are you today?
HS: I�m fine, thank you, Judith. I�m Howard Strauss. I�m the technology anchor for the Tech Talk series of technology webcasts. In this webcast, I invite you to join Judith and me in a lively technical dialogue with our guest experts, Bob Dixon and Jill Gemmill, that will answer the questions you�d like answered and ask those very important follow-up questions. You can join in this dialogue by sending your questions via e-mail to expert@cren.net anytime during this webcast. If we don�t get to your question during the webcast, we�ll provide an answer in the webcast archive. We have always tried to use the very latest technologies to bring people together, but the large distances�even on our small planet�have made that very difficult. If only we could move people and objects at the speed of light, then our meetings would not be constrained by long hours of travel or compromises in communication such as e-mail. Many science fiction authors found that teleportation, the feat of making an object or person disintegrate in one place while a perfect replica appears somewhere else, to be the ideal solution to this problem. In 1993, an international group of six scientists including IBM fellow Charles H. Bennett confirmed the intuitions of those storytellers by showing that perfect teleportation is indeed possible in principle, but only if the original is destroyed. While we wait for these distinguished scholars to work out the kinks in teleportation, video technologies�while not quite as good as teleportation�have solved many of the problems that up until now have limited their use. Equipment has become much cheaper. For a price of $600, you can buy a complete business-quality videoconferencing device, including camera and multimedia processor, that will attach to any computer, even a laptop, and fit in a shirt pocket. Video quality has become much better. Gone are two-inch-high jerky images with scratchy sound. Now you can have full-screen, full-motion, high resolution video with stereo sound and better compression techniques make better use of whatever bandwidth you have so you don�t have to rewire your campus to take advantage of all this. None of this comes at no cost. With better quality video, the way you produce them becomes much more important. In a grainy, jerky, two-inch-high video window, the lighting, camera quality and control of the action doesn�t matter that much. But it does with the video quality you can get now. Because the video we transmit now uses the Internet protocol, or IP, all of the features and facilities of the Internet�good and bad�are available. This has the potential to put video on every computer, PDA, cell phone and all other web-enabled devices. Many people are already doing great things with video over IP. Virtual conferences are being held, video job interviews are being done and complex instruments are being controlled. Your college or university can build on the experience of these early adopters. As Yogi Berra said, �You can observe a lot just by watchin�!� While we are not quite ready to support you saying, �Beam me up to the CREN Technology Conference� and have you teleported there in a fraction of a second, Jill and Bob will tell you about fairly amazing video technology that will bring people together at the speed of light on today�s audio webcast of Tech Talk. Judith?
JB: Oh, thank you, Howard, and it�s good to be here with audio being �almost perhaps as good as video with just a few little changes.� We�ve got our pictures up of Bob and Jill, up on the website, so you can just imagine them as they talk and you will almost be there! Let me welcome our experts for today, Bob Dixon and Jill Gemmill, and they are both members of the VIDE, the Video Development Initiative. And I would like to first introduce Bob, and then we�ll introduce Jill. Bob is a Senior Systems Developer and Engineer at OARNet and a Chief Research Engineer for the CIO at Ohio State where he has been for many years, since 1972. Bob, have you been working in video all this time?
BD: No, I started in video maybe three or four years ago when it began to be possible to do this over the Internet.
JB: Okay, well, it will be really great hearing your experiences about that. Let me also, then, introduce our second expert, Jill. Jill Gemmill is an Assistant Director of Academic Computing at the University of Alabama at Birmingham. Jill also has spent time in Ohio at Antioch College before migrating south where she is now completing an MSEE in Electrical Engineering. Jill, do you need that degree for VIDE?
JG: No, I actually don�t need the degree at all. I�m doing it for fun!
JB: Well, good!
HS: Sixth grade would have been enough, right, Jill?
JG: Yes, that�s right. For the technology, it�s off-the-shelf, ready for the average consumer to use.
JB: When I saw that on your bio, I thought, �Oh, my goodness! If that�s how you�re preparing for this VIDE, this stuff has got to be pretty hard!��
JB: I actually got into the videoconferencing from the point of view of applications that work well over Internet 2 or high speed networks. My last job description was Network Applications Specialist and my engineering degrees related more to the Internet technology than videoconferencing per se.
JB: But, great�it gives you a nice sense of actually, wow! We�re doing a lot with video right now. It�s still kind of a tough problem out there.
HS: It seems like�we�ve all seen video. I mean, everybody has watched TV and all that kind of stuff so Jill, I wonder if you could tell us what�s special about video over IP?
JG: Well, what�s special about it is that it is over IP. There have been in many places, including Alabama and in Ohio with the group Bob Dixon works with, since at least the early 90�s, there have been videoconferencing systems that used leased telephone lines for communication. And they�ve provided very good quality and a lot of states have used them to do classrooms where the classes are�half the class is in one city and half the class is in another and there�s very excellent video and audio quality. What�s new about using the Internet as the transport medium is that, rather than paying special fees for special telephone lines that you might only use at certain times of the day, you use your regular�if the speed is adequate�Internet connection, the same one that you use for e-mail and everything else that you use the network for for your videoconferencing.
HS: But aren�t there a lot of problems with the Internet? I mean, it�s congested and crowded and clogged and we�ve all seen these really slow kind of videos on our screens. How are we getting around those problems?
JG: Well, those universities�and there are, oh, 185 or so of them in the US plus universities around the world that are connected to their educational high speed networks�have the advantage of having a very big network infrastructure that does not have the problem of congestion. But I must say that in the commercial world, we�re also finding that because the technology is allowing bandwidth to increase substantially every year with the cost dropping, that network service providers have been able for the most part to provide more bandwidth to customers, including cable services to home, digital subscriber lines and so on. So bandwidth is there and is coming and the technology I�m talking about, which it works for sure on Internet 2 type networks, has been used on commercial networks. I have used it talking to people in Louisiana and South Carolina before their Internet 2 connections were completed and it worked quite well.
BD: And I use it with people that have DSL connections or cable modems but not if you have a 56k dialup modem. That�s just not fast enough.
HS: But what if, Bob, you�re just at a university and you don�t have Internet 2. You just have your ten megabit Ethernet connection or whatever it is you might have at a university. Is that fine? Do we need Internet 2 to do this kind of thing?
BD: Not usually, no. It really depends on how crowded your campus network is now and how big is the pipe that goes from your campus network out to the Internet. If those things are not already running at nearly full capacity, you can do video fine. If they are heavily loaded, well, then you won�t be able to.
[Jill and Howard speak simultaneously.]�
HS: Go ahead, Jill.
JG: These videoconference connections run ideally anywhere from 384 kilobits to one and a half megabits per second, one and half being closer to a T1 connection, if people are familiar with those terms. So for today�s technology, as Bob says, if the network is not congested, these are not monster truck type applications.
HS: They�re not monster applications? What keeps them from being? I mean, if we have 20 or 30 people at a conference, doesn�t that create lots and lots of traffic?
BD: Well, it depends where they are. I mean, if they�re all on the same campus, then there�s usually no problem because you�re not stressing your pipe going outside. But if they are distributed around the country, as they most frequently are, then you�ll only have one or two people at your campus and that�s not really stressing your local Internet connection either.
HS: Maybe we could talk about a real example. I understand that the recent Internet 2 Conference was actually done virtually.
BD: Yes, that�s right.
HS: Could you tell us how they managed to do that? What did people have to go through to actually put up this real conference? I guess it�s a virtual conference.
JB: That�s right!
HS: It was a real conference in that it was a conference that was planned to be live.
BD: Yes, it was planned to be a live professional conference with parallel tracks and many things going on at the same time and they managed to reschedule most of it, I would say. There was as many as four parallel tracks going on at the same time and they were being broadcast over IP video with a variety of techniques. They were using streaming video, which is a one-way technology with the relatively small windows that you talked about. It was also being sent with higher quality video like MPEG 2 so people could see that more like broadcast quality. But then it was also being sent as videoconferencing where people were actually engaged in full interaction and conversations back and forth.
JB: And it was really this interactive videoconferencing that we want to focus on today, right? Excuse me.
BD: Yes.
JB: And so in terms of where the parallel tracks were being provided as part of the conference, people were joined actually generally in special rooms with special equipment in those rooms, correct?
BD: Yes. Some of them were sitting at their desks with small equipment sitting on top of their CRT display. Others were in small rooms at various universities around the country. And we followed the motif of the meeting room by saying that, well, we had an Indiana room, a North Carolina room and we had two Ohio rooms called One and Two. And we gave out the network addresses for all of those rooms and we just told people which meeting would be in which room, just as if you were actually going to a physical conference, so people could tune in to whichever session they wanted to participate in.
HS: Okay, can we just take the pieces of this piece by piece? We had the people who were the attendees kind of things. And you�re saying they could have been in their own offices or they could have been in some special room. What kind of equipment did they need?
JG: Well, I was a person who participated in a panel and also sat in on some of the meetings so I�ll answer that part. And Bob sat in on�has been very involved in the back-end equipment setup and configuration.
HS: Good, that�ll be the next area that we want to discuss. We�d like to look at it from the viewpoint of a participant and then the viewpoint of the people who were trying to put this conference together.
JG: I have one of the small videoconferencing units in my office of about the price range that you mentioned earlier today.
HS: You�re probably talking about the one I saw on the web.
JG: Yes. And I went to a web page that Internet 2 had set up describing the conference and they actually had a sketch of�it looked like a room layout sketch and you would sort of click on the box with, okay, the Ohio One room, I would click there and up would come a page with instructions. And my videoconferencing unit is a piece of hardware that plugs into a USB port on my computer, has a camera, microphone/speaker and software that comes with it, so I opened the software. I put a short couple of dial strings in, I clicked the dial button and I was connected by magic.
HS: And you�re looking at the camera that�s kind of sitting near your computer screen?
JG: Right. Actually, part of the trick is that you have to�because on your screen, you see a picture of the conference and-�
HS: What do you mean, �the conference�? There�s lots of people attending this conference sort of virtually. What do you see?
JG: What you actually see can vary a little bit according to how the conference is set up, but you will typically either see the last person who spoke�because these systems are voice switched, which I�m sure we�ll be talking about more later�so you see the person who is speaking, or in a moment of silence, the person who last spoke. Although it�s possible, like for the panel that I was on, that one can use a video mixer at the back end where Bob was sitting so that it�s kind of like a Hollywood Squares panel so you might see three or four pictures all inside of a single what looks like a TV frame on the computer. And since that�s what you�re looking at, it�s important to sort of position. And the camera I have is one of these little eyeball cameras that are often found at stores, so you have to kind of put the camera where, when I�m looking at the picture on the screen, it looks like I�m looking at the camera because if the camera were off to the side and I�m looking at the screen, it would look like you were looking at the side of my head when I�m talking.
HS: Right, so you�d be looking away. It�d be impolite.
JG: Right. It would appear impolite.
HS: And so, Jill, do you have any ability�I mean, can you control what you�re looking at or somebody else is controlling this? I mean, if you wanted to look at other people who are also attending this conference virtually, do you have any control over that?
JG: The only thing that the end user has control over is the all-powerful MUTE button. Your microphone is either on or off, and remember that I mentioned that�and these sessions that we�re talking about had dozens or sometimes as many�I don�t know, Bob can tell you what the maximum number tuned in on any single conference way�but we�re not talking about three or four people, we�re talking about large numbers of people, all of whom can have live microphones in a voice switched environment. So��
HS: If your microphone was on and you sneezed, for example��
JG: The camera would snap to me. Yes, I could make people look at me by sneezing or rustling papers.
JB: But if you had your mic on MUTE, it would not do anything?
JG: Well, not if the software is behaving itself! I did get caught making a noise and my MUTE button said it was mute, but apparently it wasn�t.
HS: Now, can Bob control this? I mean, Bob, could you prevent�even if Jill sneezed, could you prevent the focus from going to her?
BD: Yes, at the central location, we can lock the conference on to some designated site and disallow other people from butting in, so to speak, inadvertently.
HS: Does that include not even hearing them if they speak?
BD: We could go to that extreme. Ordinarily, you don�t, though, so if they said something we would hear that. But yes, we can selectively mute any or all of the participants.
HS: I mean, so if somebody had a crying child in their office or something, you could just turn that off.
BD: Well, in fact, that�s a very important diagnostic tool because it often happens when you have 20 or 30 or 40 people at a conference, somebody is inadvertently causing trouble. They�re carrying on a side conversation or they�re clacking away on their keyboard or something and disrupting everything and not knowing it. So then we just mute them out completely.
HS: Okay, you have equipment at the place where people are making presentations. What kind of equipment do you have there?
BD: Well, the people are making presentations from all over the country in their offices and really, the originating point is no different from the receiving point in that sense. Although some of the speakers actually came to Ann Arbor because Internet 2 wanted to fly them there and there they had a professional studio with very good lighting and sound, where they originated many of the actual plenary presentations. But the track sessions were originated from anywhere.
JB: That�s kind of an interesting point, I think, Bob. So you�re saying that really the origin and the receiving equipment, really it�s the �same stuff�?
BD: That�s right.
JB: Okay.
JG: For instance, the panel that I was on, I was sitting in Birmingham, Alabama. There was Egon Verhaaren who was in the Netherlands at the time, Tyler Miller-Johnson from North Carolina, Mairead Martin from the University of Tennessee-Knoxville and we were all at our home institutions, at our desks.
HS: And you all had basically the personal equipment on your machines?
JG: Yes.
JB: Okay.
HS: But Bob, you had some special equipment that enabled you to decide who was locked out and what would be on the screen. What is that?
BD: If you just want to have two people in a videoconference, you just call each other like you�re on the telephone. If you want to have more than that, then you need some central device just like you do with a telephone.
HS: Like a telephone bridge? Actually, we�re on a telephone bridge right now.
BD: Yes, we are. You can call this a video bridge, but it�s more commonly called a Multipoint Control Unit which essentially everybody calls into, just like in the telephone system. And then the default action is that everybody hears everyone else all the time but the video is chosen to be seen by everyone depending on who is talking the loudest at that moment. But that can be overridden by the person who is actually controlling the Multipoint Control Unit.
HS: Okay, so that MCU is located in some spot.
BD: Yes
HS: And in this case, you had control of the thing.
BD: We have two large MCUs here at Ohio State University which we used. Those were called the Ohio One and Two rooms. There was one in Indiana and one in North Carolina.
JB: So those were the four rooms, then, that took care of the four parallel tracks that you mentioned.
BD: Yes, that�s right.
JB: Okay.
HS: What do you mean by �call into this thing�? Do you literally mean pick up a telephone, dial some number, what�how do you call in?
BD: See, this is just the analogy that the industry uses. They think that people are comfortable using the telephone analogy.
HS: Okay, but what�s really happening?
BD: You�re really typing an address into a field on your screen using your keyboard.
JG: And that address might look like an IP number or it might look like a numeric string that references a conference number at Bob�s MCU.
HS: But is it an IP address? Is that what it is?
BD: If you�re just going point-to-point, it typically is, but if you�re calling an MCU, it typically is not and it typically is a series of five or six numbers which have been preassigned and publicized, in this case, in the I2 web pages.
HS: And I mean, you type this number into your browser, into some special software, what is
BD: No, there�s no browser involved. You type it into the videoconferencing software.
HS: Okay, and how does one get that videoconferencing software? Does that come with one of these little camera-like devices like Jill had?
BD: Yes, the whole thing comes together.
HS: Okay. So we�re not using web browsers at all to do this kind of thing. The video window that opens is opening in this specialized software.
BD: That�s right.
JB: But in terms of, then, the Internet people who were just watching it as streaming technology, they were using it from their browser?
BD: Probably so, yes.
JB: All right, so we�ve got a lot of distinctions here. Have we talked about standards as yet? And what kind of standards are being used in terms of using IP over video?
HS: Right, that�s a good point because I think some folks in the beginning�I forget if it was Jill or Bob�mentioned things like MP3 and that kind of stuff, and we�ve heard about those standards. But are there other standards involved here as well?
BD: Well, there�s almost more standards than anybody would ever want! In the streaming video field, you have proprietary things like Real Video and then you have Microsoft�s Media Player and you have Apple�s QuickTime. I mean, those are all things that vendors created and they�re trying to standardize on them. But then you have the entertainment industry that created things like MPEG. There�s MPEG 1 and 2 and 3 and 4, which again deal typically with one way kind of devices. And then we have the telecommunications industry, the ITU specifically, that creates standards like H.320 which explains how videoconferencing is done over telephone lines. And there�s H.323 which defines how video is done over IP.
HS: Do these standards, like H.323, does that include some compression algorithms and things like that? What�s in that thing, H.323?
BD: H.323 is a large umbrella standard which includes within it maybe a hundred or more other, more detailed standards, some of which define the actual compression algorithm used. And it can be�H.261 is the most common one and H.263 is the newer, less common one.
JG: And those are the video standards. And then are a separate set of multiple audio encoding standards that are supported and there are also data collaboration standards which fall under T.120 which reminds me that the other person on the panel with me at Internet 2 was Mary Fonner from Georgia Tech who is an expert on data collaboration and T.120.
HS: So H.323 is not using anything like any of the MPEGs?
BD: No.
HS: Like, you don�t send MPEG using a 323��
JG: That is not part of the H.323 standard at this time. There�s also an architecture for how these pieces fit together. There are, you know, encoding standards for video-audio, there are call setup standards that have numbers Bob and I haven�t mentioned yet. There�s the data collaboration piece and there�s some rules about the architecture for how these fit together so it would be conceivable in the future, if the ITU were to make higher resolution video part of the standard, that those numbers�something like MPEG�could be added. But at this time, that�s not part of what an H.323 standard device would support.
HS: But Jill, as a user of this thing, you take this little gadget out of the box and you plug it into your PC and load whatever software you want and��
JG: And that�s all I need to know for how to find an MCU or the other person I want to talk to.
HS: Right. And so it understands H.323. That�s one of the things it understands and it understands how to compress and decompress and all that kind of stuff.
JG: Right.
HS: And so most users, it�s like it�s nice to know there�s some standard out there, but for most users, it doesn�t matter at all.
BD: That�s right, and in fact, it�s even better because sometimes people will have different algorithms that their software prefers to use or is set to use and when you connect two stations together or one station to an MCU, they will negotiate back and forth to some common denominator that they both understand and then that all happens automatically and the user doesn�t have to worry about it.
HS: We know what these little gadgets like the kind that Jill had connected to her machine cost. They�re about five, six hundred dollars.
BD: Even as low as $400.
HS: Okay, even better. The price, as we�re speaking, is dropping.
JB: Getting into the range of a Christmas gift, it sounds like here.
HS: What do these MCUs cost?
BD: The least expensive MCU that works well is on the order of $18,000.
HS: So that�s the kind of thing a university might have one of those things or��
BD: Yeah, that�s right. There are some that are publicly available so that not every small university has to have one initially, but in the long run if they�re going to do this significantly, then they probably do need one.
HS: Now at Princeton University, and I think this is true at lots and lots of universities, we have these videoconferencing facilities. We have a big room with cameras and monitors and all kinds of stuff here. How does this kind of technology, how is it going to impact these rooms? Do we not need these rooms anymore? I mean, can we just say, �Hey, every office, every conference room, is really a videoconferencing facility?��
BD: Well, I would say you now have more options than before. It�s still wonderful to use those rooms if you want a very high quality, polished presentation and you really want to control everything carefully. And by the way, you can use those rooms with the new IP video also by adding some small conversion box. But now we have all these new opportunities that you really don�t have to use a special room like that. You can do it from a classroom or a laboratory or your office or your department conference room or anyplace you want.
JB: So I think you�ve made a point there that you might want to reiterate, that really those neat rooms that have been set up and have been set up for years are probably running over the telephone lines rather than over IP, right?
BD: Yes.
JG: That�s correct.
JB: Okay.
HS: So they�re using something like ISDN or something like that.
BD: Yes, that�s right.
JG: And the standard that they use is the H.320 standard that we mentioned earlier.
HS: So if one were to go into one of those rooms and say �We can still do what we�re doing, but in addition to that, if we provided an Internet connection in those rooms, then we could use this H.323 and just use the same physical facility with all the remotely steerable cameras and all that kind of stuff� to do the kind of stuff you�re talking about.
BD: Right, you could buy a box called a VIU for maybe $4,000 that would enable that room to function on the Internet.
JG: And some manufacturers, like the manufacturer whose software is in use in our big classrooms like that here, actually the software is upgradable so you can use the computer to switch between Internet or telephone connection.
JB: Oh, that�s neat!
JG: And a third option is there�s a way to bridge calls from the H.320 telephone world into the H.323 world and so when that bridge is provided, someone in one of those expensive classrooms�even if it is still on an ISDN line�can call in and people like myself sitting at my desk with H.323 could participate in the call if it [inaudible] available.
HS: We have a couple questions that have come in. Here�s a question specifically for Jill. Most folks will just send a question in saying, �Who cares who answers it?��
JG: Uh-oh!
HS: Not this one! Okay, this is Timothy Wynkowski and I�m trying to figure out where he�s from.
JB: I don�t know, yeah.
HS: From wherever LUC.edu is. At any rate, Timothy says, �Jill, you said that you connected your videoconferencing appliance to your computer via the USB port. What operating system are you running? We have NT and NT doesn�t work well with USB ports.��
JG: Well, he�s an astute person and he�s absolutely right. I have Windows 2000 on my desk. But for those of you who do have NT, there are devices in the same price range that are PCI card based so the little more determined person with a screwdriver or technical support person could get one of these PCI card systems. And when I first got into videoconferencing I did have an NT system and I did use my screwdriver and I installed the PCI card and merrily videoconferenced away.
JB: But then you almost have that engineering degree, Jill.
HS: Yeah.
JG: Well, that�s what tech support is for. If you don�t have an engineering degree, get one of them!
JB: Actually, LUC is Loyola University in Chicago, Howard.
HS: Okay. And Tim has another question, a follow-up question which we said we would do follow-up questions and Tim managed to get one in here. Tim wants to know which desktop videoconferencing appliance you�re using. We sort of have danced around this, but why don�t we talk about the thing? Jill?
JG: Well, I�ve used more than one so I�ll talk about the two that I�ve used.
HS: Okay, that�s fine.
JG: I have used�VCon is a company whose PCI card and USB port systems I have used, both, and been happy with. And there�s PolyCom makes a product called Via Video that is a USB port based one that I used. And I like all of them.
HS: Okay, we have another question here from John Snead at Whitman Community College and John�s bringing up the question of quality of service when we�re doing all this. How do we deal with quality of service issues on videoconferencing and general video over IP?
JB: And I think, and let me just add this, he also was asking particularly how do you deal with across networks rather than being on the same one?
HS: Yeah.
BD: There is no guaranteed quality of service today and everything still works fairly well even without that. I mean, some people are working on implementing quality of service particularly for video, but that isn�t here today, and so what we rely on is using networks that are not fully loaded. And especially this is true of Internet 2. I mean, they have plenty of headroom, but even old networks, if they�re not physically heavily loaded, you really don�t need quality of service.
JG: Yeah, when the Internet 2 whole project started there were grand expectations that developing some quality of service techniques for networks would be a big focus of it, and that is still a research interest. But the fact of the matter is that the industry�s ability to provide excess bandwidth through things like Wave Division Multiplexing and lots of changes in the technology that just makes bandwidth increasingly cheaply available, the industry has just taken the�we�ve developed networks that know how to work with congestion. So translation, just like people I know�almost everyone listening, I�m sure, has watched some RealPlayer stream and if some network congestion happened, you�ll see that it drops the rate down a little bit to try to give you�adjust the quality by lowering the bandwidth. And these videoconferencing devices will do the same thing.
JB: Jill, do you think we�re kind of adjusting our expectations in this area a little bit, too, as a way of rather than saying we have to duplicate what we were able to achieve with the high quality and predictability of the telephone networks, that we�re kind of finding application areas where it�s good enough?
JG: Well, that�s certainly not the objective of the systems that Bob and I are putting together.
JB: Okay.
JG: We want to be able to duplicate that rock solid television quality telephone line based kind of service and we quite often achieve that with our systems.
HS: We have sort of a comment and a question by Randy Cross and I think he�s at Washington State University and it fits in perfectly. His comment is that when video fails, we talk, and when audio fails, we walk. His question is, �What�s the actual impact on a network with 384k connections? Is it a steady 384k drain on the network? Is there any way to tell what the impact is going to be?��
BD: Well, actually the traffic speed varies, depending on how much the picture is changing. The audio goes pretty much all the time but the video varies. But still, you should plan on having 384 kilobits of bandwidth all the time. Otherwise you�ll have some drought frames or blotches on the screen or something like that. And it�s like any other application, like web browsing or e-mail or whatever. You just have to have enough capacity for the number of users that you actually have. And in most cases, the number of simultaneous users even on a huge campus like Ohio State University is very small. So it really has no impact on the total amount of traffic generated by the university.
HS: Okay.
JG: A place that I would pay attention to those kind of questions is where they MCU is located, however. That�s a little bit different.
HS: Okay, so there�s going to be more impact there?
JG: If I�m sitting at my desk, I�m generating one 384 kb stream but if Bob is sitting at an MCU and he has 50 people connected to it, each of whom is sending in a 384 kb stream, then take the 384, multiply by the 40 people connected and he has a lot more bandwidth to worry about at his end but he knows where it�s going to be headed so he can design the network ports that that system has available to prepare for that.
HS: Randy Cross just sent us in a note here that addresses the NT-USB problem. Randy points out that PolyCom has just come out with a software version for Windows NT that activates the USB ports on a machine running Windows NT 4 with Service Pack 5 or higher. So��
BD: Yeah, I believe I�ve heard of that.
HS: You may not need the screwdriver here.
BD: And I�m using Windows 98 and it works fine there, too.
HS: Okay.
JB: While you�re catching your breath, Howard, I�d just like to remind folks that it is a great time to send in your questions and you can send in your questions for Jill and Bob to expert@cren.net.
HS: And I also appreciate the fact that we�re getting lots of them right now. This is great! Keep them coming! Tom Head at VT.edu.�that�s Vermont-something, I guess�says, �Are there inherent limits on the quality of the H.323 protocol?� Bob?
BD: Well, there is a defined resolution which I�m not sure I remember. On the order of 350 by 250 or something like that is the actual resolution and you can�t ever exceed that using that standard. And so once you get to a totally noise-free, stable situation, that is the ultimate limitation. But you can send that at whatever rate you want so that you could be having very rapid motion taking place without loss of picture quality.
HS: Okay, we have another question from John Farquhar at wwu.edu and he says, �Bob mentions a conversion model for ISDN to IP.� And John wants to know where he can find one of those things.
BD: The one I mentioned is made by RadVision. You can go to www.radvision.com and look at the Video Interface Unit or VIU which they sell.
JB: Okay, and that was from�John sent in a note saying that he�s at Western Washington University and the question before, I would just like to note that Tom Head was from Virginia Tech. Not Vermont, Howard. Just wanted to kind of get you further south here.
HS: That�s fine. We actually have a question from Serge Goldstein at Princeton University. That�s probably two doors from here right now! And he said that he�d really like an offline response, but I think it�s an interesting enough question really�and not just because he�s here at Princeton�that perhaps we should address it. He says, �Why was Net Meeting excluded from the I2 virtual conference?� Can we talk about Net Meeting as a player in this videoconferencing game?
BD: Yes, and it�s actually very timely. There was just a discussion going on on the Megaconference mailing list today and the last few days about this. The problem with any software-based system, exclusively software-based such as Microsoft Net Meeting or White Pine�s UCMe is that they are inherently too slow because there is no PC today that is fast enough to do the real time video compression and decompression needed for full-screen, full-motion video. So those systems, while they work, they provide typically small windows and relatively slow video. And even worse, they sometimes are incompatible with other systems and so they cause other systems to fail or have difficulties. The discussion that was going on recently with a number of messages posted pointed out all sorts of different areas of difficulty and nobody seems to really understand the underlying cause of some of these problems.
JG: And as it happens, VIDE, the Video Development Initiative group that Bob and I are members of, has recently started a Microsoft conferencing working group that wants to study a number of issues with the Microsoft conferencing products and the question of these incompatibility issues with all the other�see, the interesting thing is that you can take end point systems from at least a half a dozen other vendors�and there are many other vendors that we didn�t mention, but people who are using that equipment participated in this virtual meeting that we talked about. There�s several manufacturers of the Multipoint Control Unit Systems that all this equipment works well together and that�s what a standard is for, so that manufacturers can manufacture to the standard and the equipment will interoperate. Well, Net Meeting is supposed to be an H.323 standards based system but it doesn�t seem to interoperate well based on our experience so far as these other vendors� equipment seems to.
HS: Bob, just on a different topic here, are there any security issues with all these people having videoconferences and things here? Does this give you any kind of security exposures?
BD: No more than with anything else. I mean, your e-mail could conceivably be ready by somebody else, depending on how it�s routed to them and where the exposed points occur. With video, it�s more difficult because the person who would intercept it would have to have knowledge of video and have more equipment and greater processing capacity to intercept it. But I mean, if they actually had physical access to the network at some point where your video passed through, yes, they could intercept it.
HS: What keeps another person from getting into the conference? I mean, I know the IP address or I know the number of the MCU and things like that. What kind of thing keeps me out?
BD: Well, typically, you don�t know that because for most conferences, you give out the password and that information only to those people that you want to be in the conference. And so there�s that security, but secondly then, if I�m the monitor of this conference and I�m running the MCU, I have a listing on my screen of all the people who are there. And if I see someone who shouldn�t be there, that�s a red flag to me that something has gone wrong.
HS: Okay, and you can just turn them off as soon as you spot them.
BD: Yeah, I could just throw them out immediately.
JG: There�s also software that runs on something called a Gatekeeper and the Gatekeeper, if there is one, is part of the H.323 standard. And it gives you some management so that you can, for instance, restrict�like I could, for instance, say, �All right, only people from my university are permitted to register with this Multipoint Control Unit.� Now, I might still have to see who from my university was using it, but if I wanted to, I could prevent people not from my university from connecting.
HS: But are there things, are there equivalent things in this videoconferencing stuff like certificates to make sure you know I�m really me?
JG: Well, you know, you�re asking�these are important questions and for the most part, the actual answer is don�t do anything particularly private or secure using videoconferencing.
HS: So we shouldn�t have videoconferencing where we�re setting faculty salaries?
JG: Right. And I�ll tell you what�s actually a serious and yet practical application where this is important has to do with telemedicine. There are certain branches of medicine where using H.323 quality video is clinically tested and accepted in the practice and those are in dermatology and certain types of�you can do ear and nose exams and so on, and that�s considered sensible. But then you have patient privacy issues. So again, some of the vendors offer�if you�re going from one client end station to another, they�ll allow you to encrypt that signal, but it�s kind of semi-proprietary. I mean, they always try to say, �Well, we�re using Kerberos� or we�re using whatever public key certificate you may have. But the way that they do it may not be�they won�t guarantee that you could connect to anyone but their brand at the other end and have this decryption work. So it�s a very early stage of that important question.
HS: Is that something that the ITU is working on now? Is that going to become part of H.323?
BD: Actually, I think there is some standard for that but I just don�t think it�s been widely implemented and most people are off doing more proprietary kinds of encryption.
HS: Is that because up until now, it�s just kind of amazing that you can do this at all and so��
BD: Yes, that�s right, and there is no publicized case yet of where someone actually intercepted a video conference and did something serious with it.
HS: Yeah, but I think like in Jill�s case where people are exchanging medical information and things like that, that�s certainly something you will not�I mean, if it was medical information about some public figure or something like that
BD: That�s right.
HS: You certainly wouldn�t want that intercepted.
JB: Talking about the quality of transmission, though, we have a question that we might want to move to from John Fritz who is at the University of Maryland Baltimore County and he�s commenting that they�re exploring how to provide IT training over the web but have been running into problems with poor screen resolution of computer applications on video playback. And so he�s asking if we have any recommendations on products or procedures to enhance the �user�s viewing experience.� Any thoughts on that? I know it�s not real specific, but perhaps you�ve got something that could guide him here.
JG: Well, if he�s trying to use the videoconferencing window, the TV screen to show written material or something, that�s quite limiting because of the resolution limit that Bob mentioned. But there�s the whole area of data collaboration. A part of Net Meeting that works well is the data collaboration tool and many of the videoconferencing vendor systems allow you to use Net Meeting as the data sharing tool with their applications, so with Net Meeting you can bring up as detailed a picture as you can bring up on your computer screen. You can share a web browser window or a PowerPoint slide or a word document, whatever, so that�if that isn�t the answer to his question, I hope he�ll ask a follow-up.
BD: We have to understand, there�s two ways of using Net Meeting. One is using it as a videoconferencing tool, which is one that we don�t recommend. The other is using it as this T.120 application sharing in which it works really well. And so typically if you do want this high resolution graphical image of some kind, you run that as a parallel operation on a separate PC.
JB: Okay.
HS: Okay, we have a question from Carl Berger�actually, our friend Carl Berger at Michigan. I think lots of us know Carl and I�m just going to paraphrase Carl�s question here. Carl basically says that he�s enjoying listening to Tech Talk and things even without seeing a bunch of talking heads with lips synched to the audio. And he wonders, is all this videoconferencing necessary just to see a bunch of talking heads? Isn�t this kind of overkill? Who cares if you see the heads with their lips moving?
BD: Well, I think it�s to each his own. I mean, many people find it very useful to be able to see body language and facial expressions or to be able to look at some object the person is showing to you or to see slides as long as the detail is not too great.
HS: So is it really that this becomes more useful and more important when you have something to see other than somebody�s head?
JG: My personal opinion on this is that if you don�t know the people, looking at them may or may not be particularly useful. I mean, other than, okay, now I know what they look like. But once you know someone, there is a�you have a personal basis that you use in visual communication and I just find actually looking at them, even if it�s not the same as being in the room with them, that I get something extra out of the fact that I know them and I can sort of read things about our conversation. Although that�s very hard to do with a stranger and a person standing at a podium speaking may or may not�their voice is either dynamic enough or not and what they look like. But I�m all for, you know, trying to achieve high resolution folk dancing for all to make visually stimulating images.
JB: Okay, Jill, you�ve had quite a bit of experience with real applications on campuses. What is your favorite application where there�s really a nice match between the current state of technology and the application?
JG: Do I just get to pick one?
HS: No, you get to pick two.
JB: You get to pick�well, we have another question!
JG: I will mention��
HS: Whatever Judith mentions, I will offer you one more!
JG: Okay. I�ll take two, then, both of which are very different. One is there is a research collaboration between UAB, Purdue and Rutgers that focuses on pharmacology and it�s a��
HS: Who�s UAB? I�m sorry.
JG: University of Alabama at Birmingham, my institution.
JB: That�s her host campus, Howard!
JG: My host campus.
HS: Okay.
JG: Everyone�s heard of UAB, of course!
HS: Certainly. Just an involuntary slip.
JG: That�s right. And it�s very common in scientific research today that there are multidisciplinary teams that cross a lot of disciplines in that you won�t have all the experts you need on a single campus, and this pharmacology team is an example. So as part of their research collaboration, they�ve each gotten small room conferencing systems�which we haven�t talked about in detail yet. These are $3,000 to $10,000 boxes that are portable that give you excellent quality and they�re appliances you just plug in the network. And so every week, they have a regular research seminar where the faculty, post-docs and graduate students take turns presenting their information to each other, which is what research groups will do on a campus. And in this case, they have five to 15 people at each of the three sites and this is part of how they�re building their research team. Another example is between our English department here and some of our local public schools. There�s a professor of British literature that�s been working in a kind of mentoring project with high school teachers who teach high school British lit. And the name of this program is Eureka! There�s a link to it on the website for this, the CREN website. And they�ve been using a website to facilitate communication among the high school teachers with the college professors who can help supplementary research background and so on and also help bring the students into the universities. So we�re going to be expanding this program using that H.323 technology which will work in schools that have a T1 connection where they�re on an educational network that�s hooked to our university network. And so we�re not talking about any high-class, untypical public school network, and that will allow students in that school to sit in on a college class on occasion and to have college professor experts come talk to the classes without actually visiting the school to discuss Shakespeare and other things in British literature.
JB: Those are really exciting, and you did get two of them in, by the way, Jill, so that was good. Can we go back? We�ve talked about, we�ve had another question that would, I think, be answered by your telling�either of you�talking about the cost for the room level units.
BD: Well, in addition to the small, more personal kind of units we talked about that cost approximately $400, they are not suitable for use with more than two people, say, sitting in front of them. Then there are the room style systems which cost on the order of $4,000 for a reasonable quality one and up from there if you want more features. They are not PC-based, that is, they don�t plug into a PC. They are an appliance by themselves. You connect them to typically a high quality television monitor and the Internet and then that�s all you need and you�re on the air. And you can, if you wish, provide a nice room with sound-deadening qualities and good lighting and all that, but most typically, people just use them wherever they normally have meetings.
JB: So if I wanted to do that from one point to one point, then each campus just needs that one room point-to-point to talk to the other room.
BD: Right.
JB: Okay. We did have a question from Richard Danielson asking about the name of the �little videoconferencing device� that costs about $400 that you�ve mentioned a couple of times, Bob.
BD: It�s PolyCom Via Video. You can look at www.polycom.com.
HS: That�s a great site, actually. Once you look at that, you�ll want to have one. I do!
JB: And as we�re kind of getting close to the end, I�d like to go back. We had a comment, it was from Serge Goldstein at Princeton, Howard, down the hallway from you. He mentioned that you�ve got a conferencing room there at Princeton that has one of our sponsor�s units in there�Tandberg�and that supports both H.320 and the H.323, both formats there, he mentioned.
HS: Yeah. We also have a question from Dan Carlisle at Oklahoma State, another interesting question I think that a lot of campuses are faced with. He says, �Can you comment on the inevitable bandwidth collision between IP videoconferencing and the growth of wireless computing on campus?� It sounds like both of these are going to be growing quite a bit.
BD: Well, first of all, wireless should only be viewed as an adjunct to the wired network on campus and never a replacement for. And most videoconferencing is done inside, in rooms, where they do have wired connections present. So it really doesn�t affect the wireless network in that case. But of course, there will be people who have Via Video, say, attached to their laptops and they�re sitting outside in the nice weather and they want to do videoconferencing. And our experience is, that works okay as long as you are relatively close to where the access point is. It does not have the same range as other applications like web browsing because video is more critical.
HS: So we could have a videoconference that includes people sitting under trees outside.
BD: Yes.
JG: Yes.
JB: I�m ready!
JG: I would like to say, I�ve taken a personal unit into a classroom and abused it, you know, used a regular PC projector to show a speaker on a screen and I�ll pick the eyeball camera up and point it at the student who�s talking. So if you only have $400, you can still do this in a classroom which is what we�re going to be doing in the public school. You just need to do a little extra work. But I was connected to a network connection that was on a shared Ethernet and the connection was fine until somebody in the next room sent a large e-mail attachment and there was kind of a momentary freeze and that could happen to you as you�re sitting under your tree in the wireless world. But if that�s acceptable to you, the tradeoff for sitting under the tree, then it�s still doable.
JB: All right, great. Well, Howard, we�re getting�we�re actually out of time.
HS: We�re out of time.
JB: Do you have a final comment or question?
HS: Yeah, I do. I wonder, I mean, having heard all this kind of stuff, Bob, Jill, do you believe that where we�re heading is every classroom is going to have one of these things, or even more wildly, that we�re going to have lots of distributed classrooms. We�re going to use this more and more to conduct classrooms as videoconferences.
JG: I think Bob and I and most people in videoconferencing are receiving an enormous number of increased phone calls in the past few weeks about videoconferencing because of people�s concerns about traveling and also because the technology was just getting to the point of being ripe anyway, where it�s real and usable, so those two things together have really made�we�re seeing a lot more business, you might say. And if people haven�t tried this out, if you go look at the VIDE net website, which is www.cavner.org/.
HS: VIDE is V-I-D-E, right?
JG: Right, but there�s a network backbone that we and other people have worked�including Bob�to put together that offers some open MCUs so basically what I�m saying is you don�t have to go buy an MCU to try a multipoint call. We have some better-fitting open out there in the world for you to call into to learn about
HS: Are these open chatrooms, anybody can just connect to this?
JG: Yes.
JB: Is that free, Jill?
JG: Yes.
JB: Wow.
HS: Is it just full of 13-year-olds out there?
JG: Not yet! Maybe after today. Because we know how many 14-year-olds listen to this program. But if they�re out there�
JB: Oh, that�s right. And any other final comment, Bob, that you want to leave with our guests?
BD: Well, we�re here to help and also if people at universities try something out and they need some advice, they can contact us.
HS: Send all your e-mail to Bob!
JB: We may send a couple follow-up notes to you both. All right, listen, thanks so much. It�s time for our closing notes. Thanks so much for being with us here today. Be sure to block off your Thursdays this fall and particularly two weeks from today when we will have two guests, discussing PDAs on campus.
HS: I thought we were doing teleportation! No.
JB: Oh, well, that�s coming! Hold on! We�ll be talking about PDAs on campus, where they are, who has�how many people have PDA�s and how do we support their use on campus. Our two guests will be Todd Grappone from the Stanford School of Medicine and Bob Kalal from Ohio State University. Also, for those of you who will be in Indianapolis at Educause, you will have an opportunity to experience and participate in a live Tech Talk with these two experts on Monday afternoon, October 29th, about at 4:35, I think, is the exact time we start. So be sure and plan on being there. Many thanks to the CREN member institutions and to longtime members such as Georgia Tech, College of William and Mary and Arizona State. A special thanks to our corporate sponsor for today, Tandberg, who provides distance education conferencing solutions over IP and ISDN networks. And thanks to all of our folks who made this happen--our Tech Talk experts, Bob Dixon and Jill Gemmill; to technology anchor, Howard Strauss; to Terry Calhoun, Tech Talk web guru; to Jason Russell, Bonnie Boyles and the support team at Merit Network; to Susie Berneis, audio file transcriber and finally, a thanks to all of you for being here and for all of your really great questions. You were here because it�s time. Bye, Bob. Bye, Jill. Bye, Howard.
HS: Bye, Judith. Bye, Bob. Bye, Jill. This was great fun.
JB: See you all at the live session at Tech Talks. Bye now!
HS: Bye-bye.
END OF WEBCAST
