TechTalks Transcript
Wireless Technologies: Where Are We?
September 12, 2002
Audio
• Streaming
MP3
• Download
MP3 (Download
Tips)
 Judith Boettcher [JB] |
 Howard Strauss [HS] |
 David Kotz [DK] |
 Joel Hartman [JH] |
JB: Welcome to the CREN Tech Talk series for Fall of 2002 and to this session on Wireless Technologies�Where Are We? You are here because it�s time to discuss the core technologies for your future campus. This is Judith Boettcher, your CREN co-host and our session today is coming to you with the support of the CREN member institutions. If you�re not yet a CREN member, be sure and join and support this sixth season of Tech Talks. It is now time for me to welcome back Howard Strauss from all of his summer travels and he�s now back at his Tech Talk Technology Anchor desk. Welcome back, Howard.
HS: Thank you, Judith. I didn�t know I had a special desk here, but I will certainly ask the people here to buy me one!
JB: I keep envisioning this, so yes!
HS: Well, thank you, Judith. I�m Howard Strauss, 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 about wireless technology with our guest experts, David Coates and Joel Hartman that will answer the questions you�d like answered and ask those very important follow-up questions. You can join in this dialogue�and we hope you will�by sending your questions via e-mail to expert@cren.net anytime during this webcast. If we don�t get to your questions during the webcast, we�ll provide an answer in the webcast archive. Many young people now have a great deal of trouble with the concepts of �clockwise� and �counter-clockwise� because most of them have almost no experience with anything except digital clocks. Try to explain �clockwise� to someone without referring to a clock with a round dial as its face and you�ll need to engage in great feats of pantomime. Even such phrases as �quarter to ten� or �half-past-seven� now sound as meaningful to our students as �And palmeres for to seken straunge strondes/ To ferne halwes, kowthe in sondry londes� which are two lyrical lines from Chaucer�s Canterbury Tales. The day is coming very quickly when our students will not understand why anyone would ever have waited by a phone for an important call or gone somewhere to make a phone call or even use the pay phone. Well, why not just use your cell phone? The California Highway Department is starting to reduce the number of roadside emergency call boxes for lack of use and cities have cut way back on those red fire alarm boxes that used to be on telephone poles on every street corner. Several years ago, when I got one of the first wireless cards for my laptop and Princeton had but a single experimental access point, I tried to show off this new amazing wireless technology. During an open house for faculty and staff, I took my laptop, unplugged it from the wall and disconnected it physically from the network. I then walked out into the hallway with it and asked random visitors to give me any URL. I immediately entered their URL on my laptop and displayed the web page they requested. �Wow!� I said. �Isn�t that amazing?� But my audience was never amazed at anything except the fact that I thought there was something to be amazed at. �But I�m fetching web pages! And I�m not connected to anything,� I pled, trying to spark some acknowledgement of this great technological feat. �But of course you�re not connected to anything!� was the most common reply. You�re on a portable laptop and it doesn�t have to be connected to anything!� Well, I still remain amazed at technologies that allow people to do mobile communications. Wireless laptop computers and PDA�s are rapidly being accepted on campuses, a few of which�including Dartmouth�have made their entire campuses wireless-accessible. For a time, people kept searching for the killer application that would justify the use of wireless technology on campus. Was it remote synchronization, mobile messaging, content rendering, mobile commerce or perhaps wireless flirting? While we searched for the Holy Grail of wireless applications, wireless mushroomed with no killer application in sight. It seems clear now that wireless is just a better way to connect to the network. No killer application is needed to justify it. Wireless has become affordable, fast enough for most purposes and widely available. Using a hard-wired connection for your phone, laptop, PDA or any other intelligent electronic device is only done today when there is some special reason not to have it wireless. Those people baffled by the meaning of �clockwise� today will soon wonder why we old folks ever went to our offices to do computing. While wireless technology is being readily assimilated into the everyday life of students, faculty and staff, there are challenges, especially with respect to security standards and who is leading the move to wireless. If you�ve already started down the road to wireless, this Tech Talk will give you lots of tips and ideas to make your journey more comfortable. If you�ve been sitting on the sidelines, you need to know that within two years, over 70% of laptops will have wireless capability built in and that Intel�s new Beneas [?] chips, due the first quarter of 2003, will come with integrated dual band [inaudible]�that�s 802.11 A and B wireless networking. Whatever the percentage of laptops that can do wireless, 100% of your students will be demanding it tomorrow. David and Joel, who know wireless as well as they know clockwise, and Judith who can help us with Chaucer, will join me in helping you catch up with those elusive people who are no longer tied to their desks on today�s webcast of Tech Talk. Judith?
JB: Well, thank you, Howard, and within the communications bubble that we have available to us now with wireless I�m going to be carrying around�not carrying around, but downloading my Chaucer for our session next week. Let me go ahead��
HS: That�s going to take some preparation for that!
JB: Right! Our experts, as Howard has mentioned, are really both from very fully wired campuses and yet they�re doing more and more wireless computing, as are all of the members of their community. And so we�ll just be talking�one of the questions we�ll be exploring today will be how people�s behavior actually does change between the wireless and wired. So with that, let�s welcome our first guest, David Kotz from Dartmouth College in Hanover, New Hampshire. David is an associate professor of computer science and his research interests include context-aware mobile computing, intrusion detection and mobile agents. He is active in many organizations, including being the chair of ACM SIGOPS and a member of the Computing Professionals for Social Responsibility. Welcome to Tech Talks, David.
DK: Thanks, Judith, it�s good to be here.
JB: Great! And let�s welcome our second guest, Joel Hartman of the University of Central Florida in Orlando, Florida. Joel is Vice Provost for Information Technologies and Resources. Joel came to UCF after decades of work in IT management at Bradley University. Joel also is active in many of the IT organizations including EDUCAUSE SAC and has also held a number of leadership positions within regional networking and information infrastructure organizations in Illinois and Florida. Joel originally was a guest on CREN Tech Talks just about exactly two years ago, talking about the mix of wired and wireless that he was instituting at UCF so we�ll have a chance to get an update on how his campus network/wireless network has been evolving. Welcome back, Joel.
JH: Thanks, Judith, good to be here again.
JB: Great!
HS: Judith, since you brought it up, David, how do people�s behaviors on campus change now that we have wireless out there?
DK: My sense of just anecdotally is that the students are taking advantage of the ability to work anywhere. In fact, some of them have said that they really like to be with their friends and so rather than holing up in their room with their desktop workstation and their wired port, they�ll grab their laptop and go down to the common room or the cafeteria or some other public space where they can hang out with their friends and do their work. Sort of touching a tradeoff between spending time with their friends and getting their work done! And so other than that, I think they pretty much do the same kinds of activities, but they can do it anywhere they want.
HS: So they�ve just changed the venue, basically.
DK: That�s the general sense that I get, yeah.
HS: Joel, are you seeing the same kind of changes?
JH: Yeah, I think it�s more a matter of convenience than it is so much a killer application and what we�re seeing is that the flexibility of place�including indoors and outdoors�is really one of the attractive features of wireless that�s drawing people to it.
HS: How prevalent is wireless on campus? I knew, David, that you said Dartmouth is 100% wireless. I guess I really mean 100% wireless-accessible because obviously there�s lots of wires around.
DK: Right, right. I mean, I think the intention is to have coverage in every campus building and most if not all of the outside spaces.
HS: You said even near a ski slope you have it.
DK: Right. Dartmouth has its own��
HS: You can ski with your laptop!
DK: I suppose you could, yeah, but Dartmouth has its own ski way and the base lodge is covered by 802.11. So if you have your little IPAQ or something, I suppose you could check it while waiting in the lift lines.
HS: And Joel, UCF, you said, does not have coverage everywhere. Is that because it�s just too early?
JH: Yeah, we�re in the process. We take a little different approach than Dartmouth and some others have and are implementing wireless in stages, basically a building or two at a time, based on needs and demands of those in the buildings. To date, we�ve put in about 120 access points and are serving a couple of thousand registered users so it�s still in its early stages, I�d say, at UCF. However, at the same time by comparison, we have about 15,000 wired locations.
HS: Do you expect that when everything is wireless�actually, David, you�re in that situation right now. What�s going to happen to the hard wired connections or what has happened?
DK: Well, I mean, I still use a wired connection when I can and that�s because it is faster, maybe ten times faster than the wireless connection. And for some things that I do, that�s helpful. I do know some people who just never bother with any wired connection and so it�s not quite as important. But there�s certainly still need to be there for older machines, desktop machines, servers and things like that where it�s just not worth using a wireless connection.
HS: Do you expect the number to drop, or is it dropping on campus? Are you having to support
DK: The number of ports?
HS: The number of hard wired connections.
DK: I�m not sure. I think Dartmouth is fairly committed to having a wired port in every room at least, every office and dorm room and so forth. And I�m not sure that the number of those ports would go away. But they might not find the need, for example, to put wired ports next to every table in the cafeteria whereas they might have thought that important a few years ago. So I would guess there would be some decrease, but probably not a dramatic decrease.
JH: Yeah, I�m seeing that wireless, in most instances, is overlaying wired and people�s expectations for wired at the desktop where things don�t move as much still can be met quite well with a high speed wired connection. And where I�m seeing wireless really take root is where the devices and the users themselves are mobile. And so I think you�ve got a continuous need for high bandwidth wired, as David said, at servers and desktops but wireless is becoming the network of choice for mobile devices, laptops, PDA�s and whatever.
HS: There was a study that we�re going to talk about a bit, the EDUCAUSE Center for Applied Research did a study�let�s see. This was around December, November-December of 2001 and we�ll be referring to that as the ECAR study�again, for EDUCAUSE Center for Applied Research. But in the ECAR study, when they went out and looked at universities and they looked at how prevalent wireless was, what did they discover with universities beyond your two universities?
DK: What type of universities or how many of them?
HS: Well, how prevalent was wireless? Actually, you�re bringing up another good point and that is where�what kind of universities, what were the differences of the type of universities?
JH: Just to give you an idea, the total number of responses in the ECAR study was just under 400 to a survey and there were follow-up telephone research with 17 of those and then site visits to six of those. So there was quite an in-depth study of the state of wireless, at least here at the EDUCAUSE membership. And what they found is that currently at the time of the study, a full 60% of those who responded were moving forward with wireless and the remaining 40%, virtually of those were planning to do so in the near future so wireless is either well-established at more than half and coming in virtually all the remainder, so I think you could consider it on its way to becoming ubiquitous.
HS: And was there a difference among different kinds of universities? I mean, you can divide universities in lots of ways, but universities, let�s say, that offer PhD�s or community colleges or��
JH: Well, the study is masters, doctor�s, baccalaureate, associate, that kind of a classification and by percentage of applicants within each category, the highest percentage of institutions of a type with wireless were master�s level institutions. Baccalaureate institutions were second and doctoral third, but the differences between those classifications was not more than two or three percentage points. Roughly 25% in each case. And so I think it�s primarily a matter of particularities of the campus itself. The other thing to say is that in general, wireless seems to be more prevalent at large institutions than small institutions.
HS: Yeah, and I guess that�s not a surprise that it would be more prevalent in large institutions. But why do you think that�s so?
JH: I think it would be hard to say, but I would guess that it would be a matter of both resources and opportunity and perhaps a demand. Also, probably more instances in which mobility would be important to students at a large campus than a small.
JB: Well, haven�t there been traditionally places on campuses that have been a little hard for a regular wired implementation and that wireless is turning out to be at least a nice alternative?
JH: Yeah, I mean, I�m sure every campus has its instances but we found a handful that are clear winners for wireless. An example would be events that require computer access in temporary locations, registration��
HS: Like Homecoming, okay.
JH: Homecoming is an example. Another is that due to growth, we�ll occasionally put in modular buildings until a full building can be built and we can serve the modular building with a wireless link until infrastructure is built. Another example I�ve heard in older campuses are historic buildings or buildings constructed in such a way that drilling and other kinds of construction are either prohibited or very, very expensive. And certainly classrooms in auditoria, labs, where you have a large number of connections in a small area, often wireless is much more flexible and less expensive than doing the wired infrastructure.
JB: So you can just go ahead and throw in�throw in! Just go ahead and put in a few access points.
JH: You could just throw it in, bring it with you when you come to class.
JB: Yeah, okay.
HS: We have a question that just came in from Ben Kearn from Humboldt State University and Ben says, �What, if anything, are other universities doing in regards to managing the proliferation of wireless access points such as Apple AirPort devices around the campus?
DK: In ones that are not
HS: Which don�t belong to the university, right. Some student comes in�and this is certainly a common thing�and says, �Hmm! I�m an access point. Connect to me.
DK: I remember that early on Dartmouth had some trouble with this because they would misconfigure them and that would sometimes cause trouble on the network. For example, they�d configure the base station as a DHCP server, which is possible at least with the Apple ones. That would cause a lot of trouble, not just for people trying to use wireless in the area but for the rest of the network as well. But I haven�t heard of any problems recently. Maybe it�s because we have wireless everywhere, that nobody needs to bother to set up their own base station.
JB: So you didn�t set up any policies, David, you think that controlled that?
DK: Right. And since our network is relatively insecure, there isn�t any benefit for someone to set up�or any concern for someone setting up an insecure base station.
HS: So we�re saying to Ben, just put wireless all over the place and your students will��
JB: Then what about
HS: Your students won�t bother to do this.
JB: Joel, what about at UCF?
JH: Well, in fact our wireless person approached me about a very similar incident this morning and so it�s fresh in my mind. We do not currently have either policies or practices that give us a clear cut answer in every case. However, if the user is in a building where we have university supported wireless service, we will invite them to use that service and take their device out because I regard home-provided access points as being a major security issue for the university, not to mention the technical problems it could cause. If the user, on the other hand, is in a building for which we have not yet provided wireless services, I think that our preferred approach is going to be to attempt to bring services up in the building so that user or users is not denied wireless access and that we provide university services, properly administered, that would allow them to remove their device without losing wireless access.
JB: You mentioned security, Joel, and when we were talking before in the prep session, you had said that you had quite rigorous security system for all of your wireless on campus.
HS: David said he had almost none.
JB: That�s right. We have two examples here.
HS: A note of contrast here.
JH: Yeah, we�re requiring every user to register the MAC address of their wireless card and that�s in a centralized authentication database which in our case means that if a user would acquire an access point different than the brand that is our campus choice, then they could not use that centralized authentication service and almost by definition, their access point would be insecure. Plus they probably turn it on the way it�s configured out of the box which means no security of any kind, no SSID whatever. So��
HS: That also means that somebody who is just outside your campus who is still within range of your access points can�t use your wireless network because you won�t recognize his card.
JH: Well, I think that, as those of you who are deeply into wireless know, there are software packages that people can get that probably can break almost any piece of security you could put in in a wireless network using today�s technology. But by and large, a novice user without technical skills or specialized software would be denied access unless their card was registered.
HS: And David, you�re not doing anything like that. You said your system is very open now.
DK: Yeah.
HS: Why aren�t you doing what the folks at UCF are doing?
DK: Well, we would very much like to authenticate access to the network and we do have some evidence that non-Dartmouth people downtown and so forth are using our network for free, but we would like to implement that kind of authentication. But so far, we�ve decided that registering every MAC address is impractical for the scale of the implementation we have right now.
JB: And how many users do you have, David, on your campus?
DK: Well, it�s hard to know exactly, but in some traces I collected last spring, this recent spring, I saw about 2,200 unique MAC addresses or cards using the network and there�s about 500 access points on campus. So it�s a fair number of people.
HS: Twenty-two hundred? And Joel, how many folks are you saying, since you know every one of them?
JH: Well, we know who�s registered certainly and there may be some people using as-yet undetected user provided access points, but we�re talking about a couple thousand.
HS: So we�re really talking about the same numbers.
JH: Roughly the same kind of numbers. But ours is a much larger number of students and so the percentage of wireless users as part of our population is still much lower than at Dartmouth.
HS: That�s an interesting question, David. How many students are there at Dartmouth?
DK: About 4,000 undergraduates and 1,500 graduate students.
HS: So you�re getting about a third or so of your student population using wireless.
DK: Yeah, those 2,000 or so wireless users are not just students. I mean, they include me and other faculty and staff and I don�t know the breakdown because I don�t know who they are. But I would guess probably two thirds or so of the cards we see are students, just a guess. Then of course, in addition to the 5,000 or 5,500 students, there may be another 3,000 faculty and staff who might be users.
HS: So you�re nowhere near getting the majority of students or the majority of folks using wireless yet.
DK: On the other hand, just this month, the new students arrived and they�re all required to own a computer or bring a computer with them and of those that bought computers from us as they came, 88% bought laptops and all of the laptops we sell are wireless configured from the start. So if the freshmen, 88% of them are wireless and incoming business school MBA students are wireless. They all have the same laptop.
JB: Well, with that, you should see a--[inaudible] should predict quite a growth, then.
DK: Right, right. So in the last year, I think it was maybe 70% laptops, all wireless. Year before that, it was 50 or something like that. And so we have at least half the student body is wireless today and probably in another couple years, it�ll be near 100%.
HS: Plus what about PDA�s? I mean, we�re talking about laptops. Now, do you have lots of users with PDA�s using the wireless access points?
DK: I don�t think so. My sense is that there are relatively few. I don�t see many around, either with or without wireless. I think that will change now that the wireless options for PDA�s are small physically and relatively cheap. They used to be quite bulky and expensive, but I would guess in the next few years we�ll start to see a lot more of that now that students can afford them and fit them in their pocket.
HS: Joel, are you seeing the same sort of thing?
JH: I�m seeing the same thing. I�m seeing relatively few PDA�s but what I do see is a variety of devices. I�m seeing the Windows CE or Pocket PC type devices. I see a lot of Blackberries. Of course, there�s the ubiquitous cell phone issue and as far as the Windows or Pocket PC things is that only recently have they come with integrated 802.11 and so consequently before, you had to buy a sled and a card, other things to add to them, making them somewhat more bulky and power hungry than the standard unit would be.
HS: When a campus is planning for wireless, who has to get involved and what are the things you really have to do to start making this work?
JH: Well, I would strongly suggest that the central IT unit or whatever campus unit handles campus wide networking has to be there from the beginning because of the security access and network configuration potential conflict issues that arise. And plus you have to have a certain amount of wired infrastructure or use the wired infrastructure to connect the access points to to begin with to have wireless. And then you have helpdesk and support issues that typically the computing organization supports. But on the flip side, I think you also have to actively involve users in both the academic and the administrative community because they�re the ones that are going to develop the applications to use the wireless service. And after all, this is about supporting applications and users� needs.
HS: What kind of applications are we seeing once we get one of these things up? What are the big applications? I know there�s not a killer app, but what are people mostly doing with this thing/�
JH: Well, by percentage, it�s basically the same things that we�re doing over the wired connection except more flexibly and the two primary applications are the very ones you suspect, e-mail and the web. Library access turns out to be, on a percentage basis, one of the most popular because libraries tend to be among those buildings that are first provided with wireless and then classrooms and other teaching locations are next, followed by a handful of administrative applications. At the institutions studied by ECAR who are furthest into wireless, including Dartmouth, you found a lot of experimenting with curriculum or administratively specific software, that is, specific applications using wireless as part of the implementation. Class-specific software, collaboration software, that kind of thing.
HS: And what about when we look at departments? I mean, I�d expect that Computer Science would be the big user, but are there other departments
DK: Actually, it�s not. I�m sorry to interrupt.
HS: That�s okay, interrupt all you want!
DK: I would have expected Computer Science to be the big user on campus as well, but actually it�s not. The business school, it turns out to be by far and large the sort of single place where wireless is used. I think that�s because all the students there have wireless laptops whereas in�at least in our Computer Science building, most people work at a desk with a traditional desktop connection. The other biggest place where usage happens on our campus is in the dorms. Of all the sort of categories, academic vs. social vs. dormitory, the dorms just completely blow away the rest of the campus in terms of the amount of traffic that they do.
JB: Now, you�re primarily a residential campus, then, too, David. Right?
DK: Yeah, the undergrads are probably 80 or 90% on campus in Dartmouth-owned housing which is all wireless covered. And so I remember early on when we were planning the network, there was some debate about whether we should cover the dorms or not because it�s obviously fairly expensive. Large number of buildings. And I�m glad we did because that�s where people use it.
JB: But at the same time, the dorms are well-wired, right?
DK: They have been for 15 years.
HS: And you said you�re leaving the wireless�or the wired connections in there.
DK: Oh, yeah.
HS: So why are they using wireless when they have a wired connection right there and it�s faster?
DK: My sense�well, most of the students are choosing to buy laptops and I think that�s because it gives them the flexibility to walk away from their noisy roommate and go to a quiet study room or to somewhere else on campus or to go down to the common room and watch TV while they�re surfing the web, whatever. So I think that they like to use the wireless for its location flexibility but the wired connections are still there. Maybe they�re not using those as much.
JB: Maybe now is a good time�we had a question from Gerald Vernor from Azusa Pacific regarding�actually kind of combining two questions. Let me take the first one. He�s asking whether or not any of you are deploying wireless connections for high security or enterprise-critical applications. David or Joel, who would like to take that one?
JH: Well, in our case, we�re not deploying wireless on an application-specific basis. It�s simply an extension of the base network. However, we do ask those users who are accessing secure applications that do not do end-to-end encryption like SSL and so on use VPN tunneling as a way of insuring the security of the session that they�re on. And the same is true of people computing from home and I was going to say that one of the things that we�re seeing is that those�particularly administrative users who sort of had access to wireless in their office and conference rooms are now beginning to request assistance or advice in establishing wireless connections at home with DSL or cable modem service. And having done that myself, I would say it�s an extremely desirable environment for a home network as well.
DK: Yeah, I have a home wireless network and it�s really very nice. I don�t think that we are�we don�t have easy connection to our wireless users because we don�t know who they are and so we haven�t specifically instructed them on avoiding sensitive applications on the wireless network. But in general, our take on security and encryption is that people should use some sort of end-to-end encryption scheme like was mentioned rather than depending on the wireless network itself or for that matter, the wired network itself, to be secure.
HS: Are there any pitfalls or things one should watch out for, any surprises out there when you go out and start implementing a wireless network?
JH: I think that there are several issues, but one of them is certainly the question of campus wide control. Is it something that�s going to be a phenomenon owned by the end users and controlled by them or something that�s under the control, or at least within the standards created by the central IT organization? And I tend to follow the camp that suggests that the IT organization has to be closely involved all along because of the potential for wireless to basically take control of the network and all that goes with it, security access and manageability, away from the IT organization if the users are permitted, allowed or encouraged to do whatever they want with the network. So that�s one of the things and the other is the support that you have to provide to wireless users that comes along with wireless.
HS: Could either of you talk about the support that wireless users need that�s different from the support that people need when they�re just hard wired into the network?
JH: There�s just a little more setup to a wireless net than there is to a wired and some of it involves things that are embedded in campus standards, the SSID and so on. And as users request wireless installations for their buildings, there needs to be standard planning that should be the same throughout campus in terms of channels used, channel assignments, how to determine the placement of access points, how to remember where you put them so you can maintain them and so on. And then the authentication as well.
DK: I think Dartmouth found as a pitfall when they were planning the network that they significantly underestimated the cost of the wiring for the access points. I think they ended spending almost twice as much as was originally planned, not because the access points were expensive, but because the electricians were expensive. You typically put access points in places where you didn�t already have a port, up high on the ceiling, and often don�t have power. And in many case, we used the access points that can be powered over Ethernet but you still had to run a wire from some wiring closet or port up to the access point in the ceiling and that cost turned out to be significant.
JH: And you have to have hubs or switches that provide power to the access points to do that to begin with so it�s another networking issue.
JB: What is a good planning number, David, after your experience, for an access point?
DK: The number I saw, which doesn�t include the time spent by some of the staff of Dartmouth, we did hire outside electricians to do the installation and that number is included. The number I saw is about $2,000 per access point for the campus wide average.
JH: That�s about what I would say, too, half of which is the access point and the other half is what it takes to get it installed.
JB: Interesting!
HS: Running the wires and all that type of thing.
JH: That�s right.
JB: And assuring you�ve got good power and backup.
JH: And the installation part of that is highly variable, depending on how the campus does it and what kinds of infrastructure they have and those kinds of issues.
JB: You�re right. I think it was kind of a surprise for people to realize that those access points aren�t near the floor but rather toward the ceiling, right?
DK: Right.
JB: Let�s see. We had another question that came in regarding what kind of enterprise wide solutions you might suggest and I think one of the questions had to do with the actual product, but I wanted to talk about actually where we�re going with standards right now. Right now, most of the products include the 802.11 B.
DK: Right.
JB: And there�s two new standards coming? What should people be planning or figuring for these standards here?
HS: Well, what are all these standards about anyway?
JB: Thanks, Howard!
JH: The standards have to do with the nominal rated speed and the frequency band in which the access points and wireless network cards operate. And the first standardized of those is what we now call 802.11 B.
HS: And that�s the slowest, right?
JH: That has a rated speed of 11 megabits and a practical throughput somewhere half of that, and like all wireless, the further you are from the access point or the more material things like walls are between you and the access point, the slower it gets, down to as low as two megabits. The 802.11 A standard�check me on this, David�but I believe it�s a 54 megabit rated speed��
DK: Yes, right.
JH: �which delivers practically about half of that, 27 megabits.
HS: Why the ratio of two here? Why is it�I mean, why do they advertise it at 54 and you�re saying you don�t get that? You get half that. Where�s the other half going?
JH: Where�s the other half going?
HS: Yeah.
DK: There�s a certain amount of loss in the protocol itself, and this is true even for wired connections. You might say you have a 10 megabit Ethernet, but in practice you might get five or 10 megabits out of it depending on how many other users there are and how much noise there is on the line and things like that. So some of it�s ways to do the protocol design itself and some of it depends on how busy the network is. The busier the network, the more waste there is. And so that�s the rule of thumb is you lose about half.
JH: Another part of the answer is that like hubs, these are shared devices and the more users on them, the more places the bandwidth available has to be divided.
JB: Now, David, you had said that the students, 88%, I believe, of the freshmen coming in were buying laptops from you, all with�or 88% of those bought had the wireless now. What cards and what standards are in those laptops now?
DK: We offer for sale either Dell or Apple computers and about 78% chose Dell and 21% chose Macs this year. HS� But they�re all 802.11 B.
DK: They�re all 802.11. The Dell happens to be, I think, a Lucent based�I can�t remember the name.
JH: The Dell is a Cisco based and the Apple�s Lucent or [inaudible] based.
HS: But the point is they�re all 802.11 B.
DK: But they�re made by different manufacturers using potentially even different chip sets.
JH: Yeah.
HS: Aren�t all the 802.11 B�s compatible with an access point that�s an 802.11 B?
DK: Yes.
HS: We don�t care who the manufacturer�s are.
JH: Correct.
HS: With respect to having them, but we�d certainly care if they were not 802.11 B.
DK: Right.
JH: That�s correct.
HS: You also talked about a standard called 802.11 G. Could you talk about that?
JH: Yeah, G is a standard that�s emerging that will provide access to both 802.11 A and B network cards or embedded NIC�s and will provide whatever bandwidth each of those devices would expect to receive. So an 802.11 A network card or built-in NIC would get the 54 megabit access whereas the 802.11 B card or embedded NIC would get 11 megabit access and that is not yet shipping. At least I don�t know of any manufacturers that are yet shipping the G standard but it�s what we at UCF are looking toward as the way of migrating from where we are to higher speeds without having to replace all of the existing cards and disenfranchise those users who are still buying PC�s with the B standard.
HS: I had just read about this Intel Beneas chip, if I�m pronouncing it right, and they said it was going to be dual band, A and B.
JH: Yes.
HS: What�s that all about?
JH: That�s their solution for places that have A and B access points that are not compatible, and so rather than to go to a G axis point that makes cards compatible, this is a card that makes both access points compatible and both are roughly and equivalently effective solution. And of course, a machine with A or B capability with a G access point could have either of the speeds they chose and probably gravitate toward the A level speed because it�s obviously faster.
HS: If you were a university today, it sounds like you have no choice but to put in 802.11 B.
JH: No, you could do a university today based on the A standard if you chose.
DK: But I think you�d have a hard time��
JB: But then most people�s cards are B�s.
DK: Yeah.
HS: Oh, not in cards, but my understanding is that virtually all the built-in stuff, if not all of it, is B.
JH: Yeah. The other choice you have is that some manufacturers of access points allow two radios in the same chassis which would allow you to have both A and B operating simultaneously in every access point at, of course, additional cost.
DK: And I think some of the access points that are planned will also be dual band, just like the Neg Fitter [?] recently announced. So you could�you really have sort of two choices for the future, or three. One is to get one of these chassis with two cards and the other is to get a dual mode A/B access point and the other is to get a G access point. And I think Dartmouth is leaning towards G at the moment.
JH: Yeah, that�s what we�re leading as well. But I think it�s something where campuses who are sorting this out for themselves need to get good technical advice and I�m sure most campuses would have their network staff involved in this and would make whatever choice they thought was best and most economical for the institution.
HS: Just one other question about these access points. Next week or next month or next year, when G is replaced by Z or something like that, if you have an access point, are they built so that you can just pull a card out of them and just replace the card with whatever the new standard happens to be?
JH: I think the answer is some are and some are not. And I�ve seen examples of both. And I can�t really answer from a technical point of view whether the new products that will be coming out even later this year, the G standard, are [inaudible] upgradeable or forklift upgradeable. I just don�t know.
JB: Um hum. We�ve gotten quite a few questions in. Howard, have you been looking at them?
HS: Yeah. We have a question from Garrett Yoshimi from Hawaii and Garrett, if you�d like to spend your time up here and I could go out there and check things out where you are, we�d be glad to do that! Garret says�a question at the appropriate part of the discussion, which I think is now��Can you discuss scalability with respect to authentication, both in general and specifically in the solution at UCF?��
JH: Yeah, it�s my expectation that by using Central Radio Server as the authentication device, that we�ll be able to support pretty much as many wireless users as necessary. Previous to the central authentication, each access point stored its own control list, MAC addresses, and the two problems were that list was finite�in fact, in some of our older devices, we�ve exceeded it already. And the second is that you have to upload that list and maintain it in every device. There are some automated ways to do that but it still is blasting an access list to every access point. And so for any institution looking at a medium to large scale installation, once you have chosen a vendor or vendors, I recommend strongly that you look at their centralized authentication technology and at least plan to adopt it in the future, if not immediately. I don�t know the exact upper limits, but my expectations, what I�ve been told by our networking folks is even at a campus of our size, it�s not going to be an issue.
HS: Okay, we have a question from Mary Grusch at [inaudible] Magazine and Mary wants to know about the potential of Blackberry. �Will these be used very much on campus?� she asks, �and why or why not?��
JH: Blackberry, I happen to be a Blackberry user and there are quite a few of them around. And of course, they are wireless devices, but they are not 802.11. And the attractiveness of the Blackberry is, speaking from a personal perspective, is three things. It has a very high resolution screen so it�s easy to use. It has a keyboard which allows people who don�t know how to do the graffiti thing to enter information quickly and it has a very pleasant push wireless e-mail delivery system where the mail follows you rather than you having to go fetch the mail.
HS: Oh, that�s nice!
JH: Yeah. So that�s the attractiveness of the Blackberry. And the later models, like the Trio and other devices, now have integrated cell phones along with them, although that combination of devices and functions is still a little bit unrefined in most cases. So Blackberry, the Blackberry�s mainly for people who want to carry their calendar and address book with them and are e-mail addicts. But it doesn�t address many of the things that people use 802.11 for, which are campus based services and so on, so forth. And you can do web over a Blackberry at slow speed but it�s not really replacement for a laptop with an 802.11 connection, in my opinion.
JB: David, did you have some response?
DK: Blackberry, of course, is a service that you buy as an individual from some provider, not from the campus. And so for all I know they offer some sort of a university package, I don�t know. We don�t get Blackberry service here so we have no Blackberry users at all.
JB: Now, that�s an easy answer, then, right?
DK: Yeah.
JB: There was one related question, let me see, from Galen Work at Rutgers that was linking back to about managing wireless authentication and he was asking if either of you had looked at any solution that provided authentication without registering MAC addresses. He mentions two, I think they must be vendors. Blue Socket or Reef Edge.
DK: Yeah, we�ve looked at Reef Edge and personally, my personal opinion of it is that it looked quite promising. The possible drawback is that it looked like we had to buy and install a physical box in every subnet on campus and the architecture of our wireless network is such that we just stick the access points on the existing subnets and by and large, every building has its own subnet. And so we would need to buy 160 or so of these Reef Edge boxes or any other vendor that works that way and that currently is not attractive. But we�re very interested in those kinds of techniques and we�re looking at that and the technique that UCF is using for authentication.
HS: Okay, we have another question. We�re getting lots of mail, which is really great, from Doug Grammer at Ohio University. And Doug says, �What type of services are the staff and faculty using the wireless for? Are staff and faculty finding the use of PDA�s with the wireless network useful, and how? And as a trainer here at Ohio University, I wonder if you have been offering training to students, staff or faculty and if they�re well-attended?��
JH: Training on wireless?
HS: I�m not sure what his question��
JH: Because the question involved both training and PDA�s and I guess I�ll just try to answer what I think the question is. We do not see very many PDA�s in total, at least not visibly in use. And what I do see indicates a wide spectrum of devices, so we haven�t yet had any kind of a critical mass or standard around which we�re building applications like some other institutions are. And as far as training is concerned, we provide no campus based training on anything having to do with PDA�s or wireless, either one.
DK: That�s true here, too, we have quite a bit of training��
HS: [inaudible]. And both of you said that the end user support, additional end user support has been really pretty minimal.
DK: Yeah.
HS: When we see people coming into classrooms that have wireless, are some of them doing things that are disruptive? Are faculty finding that there�s any problems, or is this all positive stuff?
JH: The indication from the ECAR study�I�m trying to find the percent of cases�but there were many of the institutions that had wireless that reported they were having or seeing some cases where students were surfing in class and conducting activities unrelated to the class content, which required the faculty to have some protocol with how to deal with that. That�s not necessarily true on every campus but I�ve seen it reported in the Chronicle and other places and I think it�s something to consider.
JB: Does that mean that the faculty member will confiscate their student�s
JH: No, I think typically what they will do is say, �Close the lid, pay attention and we�ll go from here.��
HS: We just put tinfoil over the access point!
JH: Yeah!
HS: That makes you work. One of the problems that people have had with access points, as I understand, is the load sharing kind of problem. That is, when you transmit and you�re within range of several access points, which one gets the thing? Do you wind up overloading one when another one is relatively free? How does that work out?
JH: Well, I mean, the computer itself, the NIC, seeks to an access point sort of like a cell phone will roam and as a good example, you could take a wireless access point, tune in an Internet radio station, put on headphones and walk around the campus and see what happens. And what should happen is that the music would be an uninterrupted stream and you could roam with a single uninterrupted session just as you would be able to talk on a cell phone from one side of the campus to the other, even if you switched cell sites. Now, the other part of is loading on the access point and in a well designed wireless network, the access points will load balance.
HS: And is that�I mean, how do they do that? Do they have special software?
JH: I don�t know�well, yeah, I think it�s built into the technology in the access points themselves and I think it�s true of almost all vendors. David, you may have more technical detail than I do on that.
DK: It does depend a lot on the particular brand and model of the access point, to what extent the access points talk to each other and do any sort of load balancing. But by and large, it�s up to the client card which access point it associates with and we�ve actually seen some problems with that. I�ve noticed in my tracing data that sometimes a client will roam from one access point to another, sometimes 20,000 times in a session, which means several times a second, and that�s at the very least inefficient.
HS: Now, that certainly doesn�t mean he�s moving around.
DK: Right. The client card can�t make up its mind and ends up waffling between several access points. And this is a problem on our campus because, as I mentioned, we have several subnets, a hundred or so subnets and if you�re sitting in a place on the edge of one building and you can see an access point in this building and another in the next building, switching from one access point to another means your IP address becomes invalid because you switched to a new subnet. And so the vision of walking across campus while listening to streaming audio is, unfortunately, not possible at Dartmouth.
JB: Are you going to��
JH: There is a point here, if I might just take off on this, which is some campuses put all the access points on a single subnet or VLAN and there are some issues that go along with that. One of the benefits is you have seamless transparent access anywhere in the institution. The downside is that you can run out of IP numbers and there are some institutions, Carnegie Mellon one reporting it, that they are now having to ration wireless IP numbers in order to maintain a seamless wireless presence by containing the number of devices.
JB: David, you had mentioned as a result of the research that you had done that you were going to design two fewer subnets, right?
DK: Well, the next round of network infrastructure at Dartmouth, as I understand it, will be structured so that they can set up a single virtual LAN for all of the access points, but I don�t think that they�re planning a single physical network for all of the access points, which is what they did at Carnegie Mellon, actually, many years ago. Because that�s quite a bit more expensive. You actually have to pull wires for the entire wireless network, all across campus. And that�s too much! So I think the plan is to use a VLAN for the next generation.
HS: When a campus decides to go wireless or when campuses decide to do this, why are they doing it? Are they doing it to save money? Are they doing it to just somehow make students more mobile? Are there some reasons that people give for wanting to do this?
DK: The value proposition, as I understand it at Dartmouth, is one of convenience. Basically, you know, you want to make it possible for people to do their network activity everywhere and not just at places where there were wires before. So it�s not to save money. In fact, it probably costs more money in our case because we keep the wired network. But I do know some smaller institutions that perhaps had not deployed a comprehensive wired network who saw wireless as an opportunity to provide networking everywhere for less money than it would cost to install a new wired network. In Dartmouth�s case, we already had the wired network and so it was a different decision, a convenience.
JH: Respondents to the ECAR study listed both general and specific reasons. The general reasons were fairly vague statements like �to complement and extend the wired network, meet future needs, achieve competitive advantage.� But the institutions giving specific reasons listed things such as student �anywhere� access, classroom access, flexible access for faculty, basically things associated with mobility.
JB: And from what I�ve seen, productivity seems to be coming to the fore as well. People are, in fact, suggesting that wireless does provide an edge for productivity since you are more�you�ve got it available to you more frequently. Has that factor come up at all?
JH: Well, I think the question is what is it that creates productivity here, and it seems to be convenience that allows people to take devices with them, spend time with devices more of the day and therefore have access to things they can do more time and more hours out of the day.
JB: We�re getting close to the end of our time and I don�t think that we�ve talked at all about printing, wireless printing. How is that being implemented on your campuses?
JH: Mostly the same way it is on wired networks. Selected printers you can access from the wireless and assigned IP numbers and off you go.
DK: Yes, printing uses all the same protocols, IP or AppleTalk, and all of that works through wireless.
JB: So it all just works, huh?
DK: It�s a non-issue.
HS: Another thing we didn�t talk about, when we talk about standards, we never talked about Bluetooth. And Bluetooth is being used to interconnect wireless devices. How is this playing on campus?
JH: Well, we�re seeing very, very little Bluetooth and Bluetooth is essentially an inter-device or personal area technology, not a wide area or campus technology. And of course, where I think it will play in is the PDA�s talking to cell phones talking to laptops talking to photo machines and so on. And right now, I�ve been experimenting with it and found it just a little bit troublesome to get things talking to each other. And about all you can do is exchange files or browse directories and so on. I haven�t found much yet of any compelling use in Bluetooth, but I�m still playing around with it.
HS: But could one have 802.11 whatever on your machine communicating wirelessly with your access points and have Bluetooth on your receipt machine chit-chatting with other people in your class
JH: Yes, in fact
HS: �doing both of those together.
JH: In fact, you can have an auxiliary device sharing the PC�s 802.11 connection using Bluetooth as a link to the PC. So I could have an 802.11 PC next to a Bluetooth equipped PDA, sharing the 802.11 connection to the network.
HS: Are we, as a result of all this kind of stuff, are we seeing some interesting collaboration using these wireless things that we hadn�t seen before? It sounds like we have the potential to do that.
JH: Yeah, well, I think what you�ve seen in the personal area network category, the Bluetooth category, is the standard is just now being shaken out to where you can buy devices and it�ll work. Sort of like the early days of 802.11. What will follow are more devices and more applications that make use of it and then life�s going to get interesting, and I think the idea of a short range or personal area network where all the things around you can recognize each other and carry on appropriate business is going to be an interesting environment that will let us exchange data and print and make phone calls and those kinds of things.
JB: Does it all mean that we can talk more efficiently with ourselves?
JH: I think efficiency is a matter of organization and not technology, and I hope so.
JB: Okay. It is time for us to wrap up. David and Joel, any final comment or question?
JH: No, just to thank our listeners and indicate those were excellent questions and I�m sorry I couldn�t answer more of them.
JB: Okay, well, we�ll forward a couple to you for an offline response as well. David?
DK: Yeah, same thing. I really enjoyed it a lot.
JB: Okay, Howard?
HS: We always have a final question and it�s always pretty similar. But we didn�t tell you what it was going to be until the very end here. But for people who are listening here who haven�t gotten started yet, what�s the first couple things they ought to do? If you�re a university out there and you really haven�t had much involvement in this wireless but all this sounds intriguing to you, what advice would you give them, David, to get started?
DK: I would first go out and buy a couple of access points and a few network cards or [inaudible] laptops and set it up in your office or your lab and get some experience with it. You might even buy a couple different vendors so you can have some sense of the features of each one. And for a few thousand dollars, you can get some early experience, and that, I think, would be a very important first step before you try to do anything bigger.
HS: And Joel?
JH: I agree. That�s what we did. And I would add maybe two other ideas to that, one of which is seed the technology a couple of other places to give others in the academic community an opportunity to do it. Seed a couple of conference rooms and try to provide wireless access to some senior administrators, particularly those to whom you�ll go later for money.
DK: Yeah, we did that and it worked!
JB: All right, great! Well, thank you, Joel and David, and thanks to everyone listening for being here with us today. I�d like to invite you all to join us again in just one week�a week from today on September 19th for a session on Student Technologies: What�s Hot? And our two guest experts will be Lonnie Harvell from Georgia Tech and William Griswold from the University of California at San Diego. Thanks to all of our CREN member institutions and sponsors for making these Tech Talks possible and many thanks to our Tech Talk experts today, David Kotz and Joel Hartman; to technology anchor, Howard Strauss; to Terry Calhoun, our Tech Talk web guru; to Jason Russell, Bonnie Boyles, Gayle Terkeurst and the support team at Merit Network; to Susie Berneis, audio file transcriber and finally, a thanks to all of you for being here. You were here because it�s time. Bye, David, Joel, Howard. See you all next week.
HS: See you all wirelessly!
JB: Wirelessly! All right. It�s out of sequence, so I�d like to remind you all, be sure to mark your calendars and be here next week. See you!
HS: Bye-bye.
JH: Bye.
END OF WEBCAST
