Hardware • 1:07:02
Wireless technologies are becoming standard in a wide variety of devices, including computers, mobile phones, and PDAs. This session will provide you with the latest information on AirPort Extreme and Bluetooth technologies that you can leverage in creating your wireless solutions.
Speakers: Lars Rehder, Paul Rekieta, Philip Kearney
Unlisted on Apple Developer site
Transcript
This transcript was generated using Whisper, it has known transcription errors. We are working on an improved version.
Thank you very much and welcome to session 500, Wireless Directions. It's truly amazing how much wireless has permeated our lives since we introduced AirPort in 1999. In today's session, we're going to hear from product marketing going over the current state of the union of where wireless is. We're then going to look at several updates that have happened recently with AirPort 3.1 and move into futures that we'll be integrating with Panther. So with that, I would like to introduce Lars Rehder, who is the product marketing manager for wireless technologies at Apple.
Thanks for coming today to the Wireless Directions session. Hope everyone's having a good show. I think it was pretty amazing to be able to get one of those cameras. I wish I could have one. A little envious. I'm a big chatter, and I want to get on the video chat thing.
About three months or so ago, I was asked to put some slides together to sort of capture what's happened over the course of the last year in wireless and extract from that some of the trends or some key bits of information to pass along to you as developers that you might be able to use in your active wireless solutions that you're working on today. Or if you're sitting on the fence wondering whether or not you need to make the investment in wireless or want to make the investment in wireless, sort of a gentle nudge in the direction of getting started with your solutions.
The flow of the presentation today is really to start talking about 802.11 at a grander level, if you will, or more broadly. Then, with the help of two of our studly wireless engineers, Phil Kearney and Paul Rekieta, we'll talk about some of the updates to AirPort Extreme. Then, I'll come back in and talk about Bluetooth. Then, at the very end, I'll actually begin a discussion or a dialogue about where we could take wireless, how we can get it to the next level, and some of the opportunities that lie within that. Then, time permitting, we'll have some questions and answers at the end.
Flickr doesn't want to work. Before we actually get into 802.11, I want to talk a little bit about the matrix, the wired wireless matrix. This is a slide that we actually put up last year at the Developers Conference to sort of really assist in clearly positioning the wireless technologies against their wired counterparts.
And in the context of the time it was done, it was actually a very useful and important slide in that there was a lot of sort of confusion, particularly around Bluetooth at the time. So with this simple slide, we've actually done a lot for the industry to really focus folks into the right general areas. This is also something that's permeated not beyond Apple. There's folks, other major players in the industry that have adopted similar positioning, as well as organizations. Like the Bluetooth SIG or the Special Interest Group, which I'll come back to and talk about a little bit later.
The message here is actually fairly simple, but it's extremely powerful. And here, really, as a developer, what we suggest or recommend is if you're working on a solution that requires a lot of networking or requires the transmission of large amounts of data over the air, we really suggest that you gravitate towards 802.11. If, however, you're thinking about more peripherally oriented solutions to a computer or are not sending as much data over the air, and that data tends to be very bursty in nature, then Bluetooth is probably more appropriate for you to use.
So let's jump into 802.11, beginning with AirPorT Extreme. AirPorT Extreme is the third-generation AirPorT product that we introduced this year, just January, at this very same facility, the Macworld San Francisco. And wandering around here today, I've actually noticed quite a number of folks who have actually picked up an AirPorT Extreme-enabled system, which is really exciting to see.
And whether you know it or not, you're part of a very special statistic, which I'll come back to in a few minutes. And having had actually those units there, I'm sure you've actually firsthand sort of felt the power and the flexibility of the AirPorT Extreme product. But I'm not sure if you're aware or not, the actual design, the roots of the design for AirPorT Extreme go all the way back to 1999, when we built our first AirPorT product. At the time, the design goals for that... The design goals for that product were essentially to make it fast, affordable, and easy to use.
That was the mantra for the original AirPorT product, and packaging that all together was a very successful wireless launch for us. As last year, as we were looking at AirPorT Extreme, and coming up with our design goals for this product, we kept those three things at heart. Those were still part of the core aspects for AirPorT Extreme. But we didn't stop there. We actually went further. We've talked to a lot of customers, whether they're consumers or institutions, and captured from them their desires and their concerns as well. So we added in two more elements to the design of the product.
One is compatibility. This was a big message. Folks are telling us, "No matter where you go with wireless, how you're taking it forward, make sure you don't leave us behind, who invested with you back in 1999, when the first AirPorT products came out." So compatibility is an important piece to the AirPort Extreme solution.
The other is flexibility. We heard time and time again from various folks different stories. Some that say, you know, my Extreme network is great, but I can't quite reach out to the far room in a house or across the pathway, if you will, to another building on a college campus.
Or we've heard folks talking about, you know, I've got a lot of base stations that are up and running. I want to be able to dial back some of that power to prevent channel overlap. We did that. We provided those features in there as well. We also heard from folks that said, really, you know, it would be nice if we could share some more resources that were available then to all the clients that were on the network. So we provided USB printing as well.
One of the biggest changes actually that we made to this product was under the hood, and the transition from 802.11b to 802.11g. Now if you remember, or if you do a Google and go back to about a year ago today, and look at some of the articles that were written, it wasn't at all clear that folks believed that 802.11g would even come close to being finished.
And at the beginning of this year, as we were shipping the first products out, there were folks that were writing, saying, oh, you know, wait. Don't jump in just yet. You don't know if your manufacturer of your product is going to leave you behind and actually get you up to the final specification. Well, as we know, in the last week and a half, 802.11g is now a standard. And we've also introduced, a week after that, last Thursday, an AirPort 3.1 update that brings all of our users up to 802.11g compliancy in a free download.
So for those of you not as familiar with the A0211G, we really see this as our next generation high speed wireless protocol. It's what's going to take us to the next level. What it provides is 54 megabits per second speed, so data rates up to 54 megabits per second, which is nearly five times faster than the older AirPorT products.
And this is actually kind of interesting. This may be something of key use for you as developers, because now we're talking about bandwidth and data rates that will support more intensive media types. And there's some interesting solutions that could come out of that. Air2.11g, by its nature, is also compatible with B. It runs in the same frequency band, so it's available for use in all the same countries. It's essentially platform agnostic. It'll work on a Mac, on a PC, or even on a handheld.
So I want to talk a little bit about momentum in the 802.11 space. I think this is sort of a key trend that's happened over the course of the last year, starting with-- the business. If we look here at this chart, we see essentially from 2001 to 2002 roughly an 18% growth in the revenue generated from 802.11 products. This is actually quite phenomenal if you think about it.
This would have happened in a time when the technology industry is sort of on a slump, if you will. And actually some aspects of wireless as well were struggling a little bit. But here, wireless land, 802.11 was actually on the rise. Last year, 11.6 million 802.11b clients were actually shipped out into the market.
This year, it's looking even better. So far, the first quarter, so through the end of March this year, nearly 5 million products have been sold. And if that trend continues, we're talking about 20 million products out there, which is nearly doubling what we had last year. The one thing you'll note actually on the chart is how fast and how impactful the introduction of 802.11g was on the market. In the course of three months, 17% of the units that were out there were 802.11g units. This is before the product was even sort of at IEEE standard levels, if you will.
It's actually very phenomenal to see an uptake that quickly in a new wireless protocol. We were very bullish on 802.11g, and we expect that to continue, although 802.11b is also sort of the mainstay, the workhorse of the 802.11 space. So they'll tend to coexist probably for the next couple of years before 802.11g sort of becomes more dominant.
So the other area of momentum I wanted to talk about was just product momentum in general. Historically, 802.11 has been something that has really been driven by the computing space, whether it's a notebook or, in the case of our products, desktops. That will probably continue to be the case for quite a while.
But one of the things we've seen over the course of the last year or so is new product categories coming online with 802.11. Whether they're handhelds or Wi-Fi phones or appliances that are as simple as allowing photographs to be sort of generated over a sort of picture frame type of a product or having wireless projectors.
So as developers, you may think to yourselves, well, you know, the computer space is kind of tight. There's not a lot from a hardware perspective I can do. There is a lot of opportunity now to expand 802.11 into new spaces and actually get development efforts going in different environments, different product categories than we've ever done before.
Let's talk a little bit about the Macintosh user adoption. We've sold over 2 million AirPort products since the inception in 1999. And in the first quarter of this year, it was 150,000. And as we heard in the keynote this morning, it's actually 300,000 as of today, which is pretty phenomenal numbers. In fact, our AirPort business is white hot. It's never been better. We're nearly doubling our business every year. There's a lot of momentum generated in people adopting wireless and going wireless with their products.
This is another one of those slides that's actually very interesting. This is that special statistic I was talking about earlier. If you take a look at all the 802.11b products sold, the 11.6 million that I had in the previous slide, we'll find that 10% of those products sold were to Mac users.
And 802.11g side, 18%. This is pretty phenomenal. This is actually very significant. If you think about that from a computer perspective, our users make up the top 5% of the computing market. We're talking about 10% and 18% adoption rate here. So you're really seeing a user group that is very pro wireless, willing to invest in wireless, willing to try out new solutions and actually get going in the wireless track.
So hopefully, this is something for you to take away and really consider as you're working on your business plans for your solutions to really consider Mac first, because it's the right user group to really get you going. And then if you need to, move on to another platform.
So the other thing I want to talk about is locations with respect to 802.11. Up to the end of last year, maybe just before that, when we talk about wireless locations, it's always been sort of those traditional buckets. It's the home. I can browse the internet from my couch. Or I can have multiple systems in the home share a single connection to the internet.
In education, I unwire a whole classroom and I allow kids to really be able to work where they feel most comfortable and still be able to connect to the resources, whether they're on a campus network or on the internet. And from a business perspective, it was about sort of extending the wired connection wirelessly to parts of their office environments. That's sort of been the traditional view of locations. And something strange has happened, something wonderful has happened actually in the course of the last, say, six to eight months. We're seeing more locations coming online than ever before and interesting in different places.
We have cities. In this case, the example is Paris. This is an effort underway by the city government in Paris to Wi-Fi enable, ideally, I think the goal at the end is the entire city, but starting with the metro lines. So as you're coming in to work or going home, you'll be able to get online.
Aircraft. The traditional, you know, verboten as far as any kind of wireless technology is concerned is now opening up to 802.11. Lufthansa has announced that they will be putting 802.11b access into their 747 aircraft, so as I'm cruising over the Atlantic or Pacific, whatever it might be, I can actually get wireless connectivity for $30 and be able to be productive throughout the entire flight.
We see in New York, Verizon is unwiring their payphones and providing hotspots for users throughout the city. So if you're on your way to work in the morning, if you happen to stop at the donut shop, you probably can get online if there's a payphone nearby. Same goes for coffee shops and bookstores through operators like T-Mobile, which currently have about 2,300 hotspots lit up in the U.S., where you can actually get on and be able to communicate, be connected.
So all these different pieces, and kind of, well, where's it all going? What's going to happen with all this 802.11 stuff? Well, firstly, let's review the 802.11 technologies. So we have 802.11b, which I mentioned before, is really the workhorse for 802.11.
[Transcript missing]
Technology of choice, or the standard of choice, if you will, for these public area hotspots, since it ensures the broadest range of compatibility.
AirTuttle 11G, fastest performing wireless, out the gates, fastest adopting wireless technology that's out there. There's really a lot of excitement around this. The numbers are looking really good. People are actually adopting it very quickly, and we expect it to be the next generation of wireless technology that will proliferate out.
And finally, we have Air 2.11a, which, you know, I was thinking about this last night. How do you kind of explain this without, you know, you can't really sugarcoat it. It hasn't moved much. In fact, analysts have told us that last year a grand total of 100,000 802.11a products were sold. So there's really not been a lot of growth there. And I think they really missed the compatibility message, and that's really hurt them in that camp. So there's really not a lot of movement there.
So what are the takeaways from AIDA 2011? First off, business is good. Money is there to be made in this space. It's clearly the case that there's an upward trend while the market is going down. There are new product categories coming online. So again, if you're not particularly interested in joining in on the computing side of things, there are opportunities to create and work on solutions that sort of peripherally work around the computing environment that would take advantage of this technology. Thank you.
There are more places to connect than ever before. So now you can actually work on a solution and kind of be assured that more often than not, when you go out throughout your day, your users go throughout their day and wherever they sort of stop and pause for a moment, they will be able to get online.
There's lots and lots of momentum, so I would definitely encourage you to really think strongly and go forth into developing 802.11 solutions. So from here, I'm actually going to hand it over to Paul. He's going to talk a little bit about the actual updates we did to AirPort last week.
Thank you, Lars. Thank you for that studly introduction. I'll try to live up to the reputation there. I want to talk a little bit about the software update that we've done and what we have planned for coming up into Panther and beyond. As Lars mentioned, it was a year and a month ago last year when we first started talking about 802.11g and we delivered that to you in January.
We did a software update in the spring and did a pretty significant software update just last week. And is everybody in here who is running wireless, have you updated to 3.1 yet? Not had any troubles? I hope so. So good, that really helps the show network, and we'll talk a little bit about some of that as we go through.
So I mentioned the 3.1 update. The goals there were, first of all, compliance with the IEEE spec, but once we get into the software, we can't just leave one thing going, so we have to work on multiple things all at the same time. And so we kind of took the top half dozen or so things that we heard back from customers and feedback that we got to try to address some of those. So with the draft specification that we released that we had in January, there were some issues with performance and robustness, so we addressed those.
AirPort Extreme introduced this concept of a WDS, which is the ability to extend your wireless network wirelessly, so people were asking for that. In large homes, or especially in schools where you want to go from a main campus to a remote classroom without having to run a wire across a parking lot or a baseball field, WDS is ideal for that solution.
But the first implementation was somewhat complicated to set up and generated a lot of questions, so we tried to address that. The other thing that comes a lot... from developers especially, are requests for SNMP support, so SNMP to get information about the network, get statistics about the network.
So we incorporated some of that in the firmware update, and one thing that would come up every now and then that we felt like would be a good thing to do is UTF-8 character support. So that's really helpful for our non-US friends who want to publish network names, publish names of base stations using their native language. And so... We allow you to do that now and I'll show you some slides about that as we go through.
First was performance. So one of the advantages of delivering a draft standard, if you will, is it allows us to get the product out into a real world situation and see how it's really used, how it's used beyond just the theoretical specifications that engineers like to talk about when they're in closed rooms. So we saw a lot of usage patterns and scenarios We addressed some of that. Early on, there were a couple of common myths that we wanted to dispel. One was that a single 802.11b client would drag the entire network down to 802.11b speeds.
We obviously didn't want that to happen. The other thing that people would say is that a base station next door, a base station down the hall, who's an 802.11b base station would, again, negatively impact the performance of your G base station. We didn't want to do that either.
A couple of technologies that our card firmware vendor delivered to us is directly directed at improving that performance in this environment where you have mixed B and G clients and networks and lots of other interference. One is a technology called frame bursting. Every vendor will call it something else. Some might call it Nitro this or Turbo that.
It's just really the ability to take that amount of time that you have able to send data. In the case of ADF11G, since you're sending it five times as fast, you might as well just send more packets during that time. That's kind of what frame bursting is and the other RTS/CTS protection mechanisms.
I always wonder what the translators are going to do with that. But basically, that's again, so you have a G-Bay station in your house, your neighbor has a B-Bay station, and you don't know, you don't know who your neighbor is, you suspect it's Linksys because his network is called Default or whatever. But if you both happen to be on the same channel, you're going to be interfering and you don't want to go over to your neighbor.
But you probably could, you could probably just connect to your neighbor's bay station and type in "admin admin" and I'm going to reconfigure it for him, for his protection. But this protection mechanism is really just a way for the two base stations to cooperate and better share the bandwidth. So when our G-base station wants to send some data, they'll say, wait a minute, I'm going to send some data. The V-base station says, okay, go ahead and send it. And then when it's done, things go back to normal.
The other is improving the robustness of the connections. So again, when NATO 211 first came out, it was highly controlled, and it would be in these enterprise shops, and everybody would coordinate what channels are being used, and you wouldn't really have problems with interference. But now that we're selling into more and more homes and classrooms and small offices, well, there's not any kind of integrated, coordinated effort.
There's just no way to know what the base station next to you, what channel number they're on, what other kind of parameters they're using. So we added the ability to look for the best channel. So when the base station starts up, it's going to look around, scan the network, and say, oh, I'll operate today on channel four. So that's just one of the improvements to simplify your connections.
The other is kind of a more low level thing that happens in the card firmware. So as you know in this 2.4 GHz space, it's unlicensed, so there's a lot of devices that operate there. 802.11 networks are one, cordless phones are another, video distribution systems, so if you're beaming satellite something from your living room to your bedroom, all those things are going to interfere with the network.
one too many. So anyway, so at the card level, again, we just look for sources of that interference. And when we see that going on, we tighten up some filters inside to better determine whether to accept or reject packets. And that's going to have a little bit of impact on your performance. So when that source of noise goes away, then we go up to a faster speed. And unfortunately, in this particular software update, we can't do that automatically.
We do some things automatically, but you as administrator or you as a user will have to control whether to use this interference robustness. So if you've gotten the 3.1 software update, you would notice in the admin utility, there is a checkbox to turn it on. And... In the menu extra of the client software, there is a menu item that says, "Use interference robustness." Looking at some of the bulletin board discussions over the past week, there's a lot of confusion about when to use that, when not to use that. Basically, you shouldn't use that unless you really need to use that.
How do you know when to use that? We don't know. You have to tell us. Generally, if you think that you have this source of interference, for sure if you have some kind of, like I said, a video distribution system where you're sending video around your house, you almost certainly want to use it in that environment. But in most other places, you probably wouldn't.
So it's going to reduce your range. It's going to reduce your throughput. So don't turn it on unless you really need to. But if you're in a place where you can't get to a network, you might as well try to turn it on and see if that helps. But just be aware of what that control is for.
The WDS setup was the wireless distribution system, so I mentioned that was a little bit complicated to set up in our first release. Under the covers, it's really very simple. It's just that the base stations have to know the MAC address, the AirPort ID, of all the other base stations that are participating in this network.
So in the first release, you probably saw, if you tried to set this up yourself, that you had to go around to every base station, pick it up, turn it over, find out what its MAC address is, go into configure one of them, enter the MAC address, the channel, the WEP key and everything, and then go to the other one and do the same thing, and then go to the other one and do the same thing, and make sure that they were all coordinated so that they were all on the same channel, all with the same WEP key.
And we thought, well, that's kind of silly, because if the base stations are all there, if the program can see them all, if the network can see them all, then why not just go ahead and configure them all for you? So that's kind of what this panel does, is two things.
One is that if you are connected to... what is known as the main base station, which is the base station that's connected to the internet,
[Transcript missing]
You can tell who the IT administrators are around here. So that gives you some statistics about what clients are connected to your base station, and statistics on the physical interfaces, how much data is going through, what kind of error rates are there.
And we'll eventually expose that through some tool with the admin utility, but right now it's just available through SNMP. And then the International Character Support. So this is something new, and we've been wanting to do this for a while. We always had to worry about some compatibility issues, but we just decided just to kind of bite the bullet and go with it.
So 8.2.11 has this notion of what is a network name, or what we call an SSID, which is just the name that your network is known by. And in this particular slide, we're showing Home Network, which is the U.S. version of that. And then we're also showing Home Network, which I think is the Korean version of that. So the radio doesn't care. The radio just cares that there's 32 bytes of data that says what my network name is.
It's just how it gets presented in these kind of user interface elements where we have to decide whether it's how to show it. In this case, we're standardizing on UTF-8 everywhere. Just be aware if you do this that some PC clients and other clients may not be able to associate with your network until they also update to do this. But that is an option.
And I think finally, how many people here were at wireless, were using wireless at last year's WWDC and using wireless at WWDC before that, and wireless at WWDC before that. And so maybe these guys are the ones that are setting up these IBSS networks. So we talked about this for the past couple of years, but finally with AirPorT Extreme, we differentiate now between an infrastructure network, so in this case, home network, which is kind of above the line, or a computer-computer network, which is showing home network below the line. So in these kind of environments where there are these mixture of infrastructure and ad hoc networks, an AirPorT Extreme user can choose which one of those they want to associate with. And so that's a great way to do it.
In some cases, you might want to do that. But I think before I go on, just a note about that. So this is the first day of a big show, and you probably had some good experiences and some not good experiences with the wireless network. And just be aware that there are a lot of people working on it, working behind the scenes to make all that happen. At one point this morning, there were over 1,800 simultaneous users of the network.
So that's putting a big strain and stress on everything. And so I think that's a great way to do it. And in particular, Everybody in this room can help. You can help with a bit of social engineering, which is to, first of all, not do this yourself, but more importantly, if you see other people, ask them not to do this, which is do not create, do not turn on internet sharing and use the same name as the show network. That helps no one and hurts almost everyone, so don't do that. And likewise, don't create a computer-to-computer network with the same name as a show network.
[Transcript missing]
So just be aware, and again, so this is a couple things here. One is the international character set. The other is this differentiation between infrastructure networks and computer-to-computer networks. And this is the interference robustness menu item that I was talking about.
With that, oh, well, the one more thing. One more thing is the Windows admin support. So we had a Windows admin utility that worked for the original Snow Bay Station, and now we've updated the Windows admin utility for the Extreme Bay Station. So it's very similar to the Mac utility in its usage. And again, just that's up on our website. It's not on the CD, but if you go to apple.com/airport, you will eventually find your way to it.
And the last thing, since all of you are developers here, what you're most interested in is what kind of opportunities there are for 802.11. We've talked about APIs off and on through the years, and we're starting to expose some of those APIs. And we have available now under NDA if you want to do anything that adds value to the client support, like turning power on and off, looking for wireless networks, choosing wireless networks, getting status information. We have some APIs that are available for that.
But you have to go through DTS, so just send a note to DTS at Apple.com and tell them what you want to do and why you think this API would be useful to you, and we'll see about getting it to you. And with that, I want to introduce Phil Kearny, who's going to talk a little bit about AirPorT security and some of the other things we have planned for the next few months. Thank you very much.
How's it going? I seem to be always running up and down the halls of Apple muttering all these acronyms and things. And they said, "Hey, nobody understands what you're talking about, so why don't you explain it to us and explain it to the developers, because this stuff is coming down the pike, and it'll be good if everybody knows about it, because there are some assumptions that sometimes people make in their networking applications that you can't make when people secure the network." So the first thing I'm going to talk about is something that is near and dear to everybody's heart who does IT and wants to actually secure their network, and that's 802.1X. 802.1X is an IEEE standard for Layer 2 network security.
What that means is you can, in a wireless access point, and also on a wireless network, you can do authentication on that network before the person is allowed to access network resources. Basically what 802.1X says is that, restricting physical access to the medium is not sufficient for security.
So any of you who have worked in the DoD or any other government sites know that they've had 802.1X on their wired networks for a long time, and 802.1X has become the basis for the new security infrastructure of wireless networks that I'll talk about in a little bit. As you can see here, if you run your Panther CD, you will find an 802.1X supplicant.
As part of Internet Connect, one of the tabs that you'll see is 802.1X, and you can do your configuration either for a wireless -- the airport interface or else for the wired network. We encourage you to try it out. We've tried it out against servers from Funk, Microsoft, Cisco, and others, and it works for us, but we really would like your feedback on what works and what doesn't work, so that by the time we ship Panther, we have a really good 802.1X supplicant in there that everybody is happy about. about.
As you can see, doing 802.1X authentication, you enter in what port you're doing, your username, password, or if you're doing an EAP type, EAP is the Extensible Authentication Protocol, which is used in concert with 802.1X, for what type of security you're doing. The supplicant that's in Panther supports TTLS, TLS, Leap, Peep, and MD5.
If you don't know what those are and want to know what they are, you can go look on the web or come find me sometime during this week and bend my ear and I'll explain them in excruciating detail to whoever wants to understand each one. For any of the authentication types that don't use passwords, stuff like TLS, there's full support, keychain support for digital certificates and stuff in there. So again, I wholeheartedly encourage all of you to try it and send us feedback and let us know how it works for you.
That leads me into a thing that we call WPA. What is WPA? Well, I call it a strong interoperable replacement for WEP. Now, everybody here who's familiar with wireless should know what WEP is. WEP stands for Wired Equivalency Privacy. It's the security, and I use that term loosely, that's on your wireless network today.
The problem with WEP is it was originally used just to secure the radio link at a time when the only place that they had wireless networking was on machine shop floors and things like that, and people weren't actually using it for data. The problems are it uses 64-bit shared keys with only a 24-bit initialization vector.
This means that as you've read all the horror stories, the keys are easily recovered. There's no forge re-protection, there's no replay protection. The same key is shared by everybody using that access point. So once the key is cracked, they have access to everybody's network traffic that's on that base station or network.
Access Point. So an industry consortium called the Wi-Fi Alliance, or Wi-Fi, which is where the term Wi-Fi or wireless fidelity comes from, they got together and said, okay, what we want to do is we want to try and make things more secure. We need a software firmware fix for existing devices out there that's cross-vendor compatible and operates in the enterprise, the home, everywhere, and can solve all these problems. So they came up with this thing called WPA, or Wi-Fi Protected Access.
It's actually a subset of 802.11i. I know there's every letter in the alphabet for 802.11, and it's hard to keep them straight, but again, if you want to come talk about all of them, I'm more than happy to. The differences with WPA, improved data encryption, and it says TKIP there in parentheses, basically what that means is instead of a 24-bit, it uses a 48-bit initialization vector. Now, 2 to the 48 is a big number. You know, that's like over 500 trillion. So it's a really good place to start when doing your encryption.
It has user-based authentication, so in enterprise and SOHO situations, use 802.1X with the AP, and you authenticate every user before they get access to the access point, and people who don't have authentication can't get on the access point. It uses TKIP, which is the Temporal Key Integrity Protocol, to make per-packet keys.
So unlike the static key that's in WEP, it's generating new keys all the time. So every time it feels like, what's going on? It's generating new keys all the time. So every time it feels like, what's going on? It's generating new keys all the time. So every time it feels like, what's going on? It's generating new keys all the time.
Well, things might be getting a little bit insecure here because there's about 20,000 packets or less. Maybe we'll change the key. All this 802.1X stuff is all radius-based, so if you have radius servers, you can hook them up to whatever directory system you want, and it'll all just work great.
Testing that I've done with WPA has been wonderful. If you have any questions about WPA, you know, because I know there were a lot of people who were familiar in the wireless industry saying, "Hey, when is Apple going to make a WPA announcement?" Well, there you go. Panther will include WPA support when it ships by the end of the year, whenever Panther ships.
Beyond that is 802.11i, which is the future of wireless security. WPA, or Wi-Fi Protected Access, is actually a subset of 802.11 that they brought the pieces in that they need for legacy equipment. 802.11i is going to be an IEEE standard, but it's not likely to be ratified before 2004. The folks in Wi-Fi decided they couldn't wait until 2004 with everybody yelling and screaming that their networks were insecure.
So we've taken all the pieces. For those that didn't know, Apple is a member of the Wi-Fi Alliance. We've taken all those pieces that are there for legacy equipment to make sure that everybody can have good interoperable security in the home. In the home, there's a mode of WPA called pre-shared key, which is 400 million times more secure than WEP. And for the enterprise, you can use 802.1x with EAP. Wi-Fi is currently investigating a thing called WPA2, or Wi-Fi Protected Access. Which replaces the RC4 algorithm with AES.
We don't know if we're going to get there before 802.11i is ratified, simply for two reasons. Currently, with WPA, we do the best possible fix that we can to minimize performance degradation on existing equipment. And as everybody knows, AES requires a little more computing resources. Unless there is another flaw found in this implementation of RC4 or in one of the other pieces of WPA, like TKIP or MIC, they're probably not going to do WPA2 anytime soon. So, you know, I'm just here to tell you, don't worry.
We're tracking all the security. We're tracking all the standards. We're implementing all the standards for you. It's going to be part of Panther. There are developer opportunities if you want to put this stuff into legacy operating systems. I would encourage you to do so. There is a big push for people who have Macs in the enterprise to be able to support this stuff. And I think... I think it's a lot of fun. With that, I'll turn it back over to Lars Rehder, and he's going to talk about Bluetooth. Love my proximity detector.
All right, let's transition out of wireless Ethernet, move over to wireless USB, which is what we refer to as Bluetooth. Bluetooth, for us, is really the peripheral solution of choice. It's the way to unwire the peripherals that connect to your computer. And it does so in a way that's actually very powerful. I'll kind of go through some of that.
It's very simple to use. We've made it very simple to use for a user to be able to connect to another device. There's security built into Bluetooth as well, so you can have an encrypted connection between the two devices, 128-bit encryption. There's also the concept of pairing. So as an application developer, you can decide what is appropriate for whatever the task is the user is doing. You can limit what the user can actually communicate with and actually force a trusted connection. You can also have a lot of flexibility there on the security side.
Bluetooth is also extremely versatile. You have a core specification, and around that, and sort of what fans out from that, are various profiles. And essentially, the profiles sort of map to different types of devices and allows you to unwire various types of peripherals around your computer. And the connections are also very reliable. At the radio level, you can always make a connection. It's always very reliable. There's no real issue there as far as that goes. So these all sort of contribute to a very good user experience around Bluetooth.
Now, as far as our development is concerned, you probably remember from last year when we first introduced Bluetooth as part of Jaguar, we introduced profiles that supported serial, so basically a serial cable, virtual cable type of replacement, object push, dial-up networking, and synchronization. And since then, we've added in two additional profiles, FTP and HID support as well.
Now, I'm not actually going to go into those in detail here. But I would encourage you, if you're all interested in this, to join in the session tomorrow morning where they're going to talk about Bluetooth in much more detail. I've got a couple of guys from the Bluetooth team on the engineering side who really dive down in a lot of detail.
What I want to focus on instead is talking a little bit about where Bluetooth has come and kind of where we are today and how we're moving forward. Essentially, if we look at... If we look at the years from '98 to about 2002, those are really the early adopter phase, or the early years, if you will, where there were fewer units sold.
Last year, there was 20 million units shipped out into the markets. There was a little bit of momentum there at the very end, but not a lot of initial movement. The solutions themselves were mainly external in nature and some integrated, the integration coming primarily on the mobile phone side of things. From a software perspective, the connections were pretty basic.
While the connections at the radio level were pretty reliable, I think they were still cutting our teeth on UI in general. For any of you who've actually played around with Bluetooth in those years, you probably remember a lot of fiddling around and throwing switches and levers to actually get a phone to talk to a headset or a handheld to talk to a phone to get to the Internet, etc. Also, in this timeframe, there was a lot of ups and downs in terms of where Bluetooth was going. The original objective was to be a cable replacement technology, the premier wireless cable replacement technology.
That was all fine and good. Somewhere in there, there was a bunch of folks that came on board at the special interest group level and decided to move into networking as well, and thereby causing this confusion around what is the right thing. I mean, is it Bluetooth networking technology or is it cable replacement? And that's one of the reasons why we put up that matrix slide, which is the first slide I showed you, to sort of help clarify some of that.
If we fast forward or come to this year, essentially what we're seeing is we're seeing some momentum now in the number of units being shipped. We're probably not mainstream yet, but we're getting really close to sort of crossing that chasm and getting to the point where we're going to see Bluetooth in more and more and more types of devices. From a hardware perspective, it's much more integrated now from a solutions perspective and less external dongles that are connecting into devices.
The UI, we sort of figured that out now, I think, generally speaking. And it's much easier now, much better integration across the different devices in terms of a user being able to engage and find Bluetooth within whatever it is they're doing. And networking, thank God, is gone as far as Bluetooth is concerned. It's strictly a cable replacement technology, which is what it should be.
Now, if we project forward to next year, we're in a position to really dramatically increase the number of units that are going to ship out into the market. We'll most likely see only integrated solutions and maybe onesies and twosies in terms of sort of more externally plug-in type of solutions.
But the main transition I think we're going to see or we need to see, this is kind of where you guys all come into play for any of the application developers that are out there. So we need to achieve a level of transparency in wireless that allows the user to sort of engage with Bluetooth at the task level within the application.
So I wanted to sort of talk a little bit about kind of some of our design goals and show you some examples of that from a Bluetooth perspective. We set out to really make Bluetooth an easy, elegant, useful solution for the user. I think we've done a pretty good job. What we've done is provided a Bluetooth menu extra.
So this is something where, you know, basically stealing a page out of the AirPort book and providing a one-click access for the user to the most important Bluetooth features that they need, be it on-off, discoverability, and being able to set up another device. This makes it much easier for a user to actually approach and engage with Bluetooth.
We also added in a setup assistant. So one of the things that the Bluetooth Special Interest Group still is in the process of figuring out is how to make the overall pairing experience easier to use. And that's something that will take a little bit longer as these sort of larger bodies take their time and sort through that kind of stuff.
And in the interim, what we've provided is a very easy way for a user to be able to set up a connection between a computer and a peripheral device. Essentially what we've done is we've broken up this series of steps into much more sort of digestible pieces for the user to be able to go through and work their way through the process. And it's been tremendously helpful for them to get started and get going.
The other thing that we've provided, which is actually also very useful for the user from a user experience perspective, is device filtering. So if you were to go out there and search for devices, say you've got an application where you want to send a message out and you want to be able to do it through a phone, that you can actually filter by the type of device, and the user is only exposed to the things they really need to choose from. They don't see other types of devices pop up on the list.
This is particularly useful at larger conferences. Granted, this is sort of extreme, but last week I was at the Bluetooth Congress, and you do a discovery there and you come back with 75 devices if you don't filter, which is a bit extreme for a user to have to go through and find the right one.
We've also built Bluetooth in to the hardware, starting with the 12 and 17-inch PowerBooks, which were introduced in January, and shortly thereafter introduced models of the iMac and PowerMac, where you can get Bluetooth as well. So we're very bullish on Bluetooth as a technology to really unwire your desktop or from a travel mobility perspective to be able to unwire the cables that you normally have to take with you when you're traveling.
So now I want to talk or show you actually some examples of Bluetooth solutions that are out there from a synchronization perspective, messaging, remote control, and basically turning your mobile phone into a modem. And one of the things you'll notice as we walk through these is sort of the common denominator across all the examples is the level of transparency that exists so that the user is removed from the complexity of Bluetooth and is really just thinking about the tasks that they're working with.
The first one is iSync. iSync is the application that we developed last year, which allows you to take your most important information with you, be it calendar or contact information. It follows you around wherever you go, across the different types of devices you would carry with you. One of the main benefits of this is being able to send your contacts back and forth from a computer to a mobile phone. The transport for that is primarily Bluetooth.
If you look at this example of the screen, as a user, all I really do is worry about checking the phone, selecting the phone, and pushing the sync button. There's no Bluetooth engagement beyond that. We've actually built that into the setup process so that when you actually set up the mobile phone, you can opt to have synchronization as part of that.
This is the address book. This is actually a really fun application from a Bluetooth perspective. If you haven't used it, I'd urge you to try it. It's really great. It actually allows you to move forward some of the voice applications that you'd find in a mobile phone over to the computer.
I could be working, and when I have an inbound call, I could see a caller ID come up. If I need to send someone a message, I can do that via SMS, or I could dial from the address book. These are all done over Bluetooth. Again, the level of Bluetooth complexity here is limited to the one button that's there in the upper left-hand corner as far as the user being able to turn that on. Beyond that, it's sort of handled.
This is also a great application, a lot of fun. It essentially takes your Sony Ericsson phone and turns it into a remote control for your computer. Extremely powerful in that it's based on Apple scripts, so you can extend it any which way you'd like. It does come with a number of scripts that allows you to navigate through iTunes and different songs and albums and so forth, as well as DVDs, or run a presentation from Keynote or PowerPoint.
But it's sort of, again, if you look at the UI here, the user is focused in on what the remote control function should be, what the action should be, and there's really no reference to Bluetooth in here at all, which is the way it should be. Bluetooth should become transparent to the user and very sort of seamless in the overall experience.
There's another solution that's out there called Mobile High Speed. And this provides sort of, if there's three elements to the phone, one is synchronization, the other is sort of voice-oriented solutions for voice applications, the other is data. And removing the complexity of being able to get onto a 2.5G or GPRS network from the user, so it's a very seamless experience. This application actually handles that and provides, again, that connection over a Bluetooth link and makes all those levers and switches that are sort of required vanish from the user and they just get on, which is great.
So what I'd like to do is actually provide you with some tips if you're working on or thinking about Bluetooth solutions. These are generated from a low sleep cycle, high caffeine sort of type of environment that we're working in for a while. But I think they're pretty useful to use as guidelines.
One, the challenge has been is that for a long period of time, Bluetooth has been extremely hard to find for a user. In fact, if you go to some type of devices, you could go five layers deep on the UI before you even find any kind of reference to Bluetooth at all.
And having this kind of barrier to the user just sort of inhibits or prevents users from actually engaging with the technology and using it. So if you're working on a solution, we really encourage you to move Bluetooth way higher up in the UI hierarchy, if you will, so the user can quickly get to Bluetooth and at least be able to turn it on.
One of the other challenges is that discovery is an easy concept. People get that, but it's not always straightforward. And again, this goes back to the filtering. Show the user only the types of devices that they really need for the tasks that they're working on and not a whole list of everything that's out there in the room.
Pairing is difficult. I mentioned this before. There's a lot of different steps involved. They need to be done in a sequence. And again, this is something that the Bluetooth SIG is looking at fixing in the longer term. In the shorter term, what we really recommend if you're working on Bluetooth, unless the data you're sending absolutely needs to be secure, avoid pairing.
Rather than doing pairing, remember where you've been. Allow connections, remember that connection, expose that to the user first so they can choose off of a cache list, and then they can get right back into where they are. What we found is essentially that most users will wind up connecting to about the same six or seven types of devices.
So there's a good chance that if they go in and you remember that connection, they'll go right to that same spot again, and it'll make it a very easy experience. Now, if you do need to pair because the data is sensitive, then go ahead and do that. But again, I wouldn't make that a default, if at all possible.
So the other sort of challenge for Bluetooth is that there's limited bandwidth available over the Bluetooth link. And you really need to be cognizant of what you're sending over the air. And think of Bluetooth as wireless USB. It's a small pipe. Don't try to send large amounts of information over that. Be very selective of what you do with it. Otherwise, the user experience will suffer, and people will get a bad feeling about the overall solution. You don't really want that to happen.
Okay, so on to the last section here. What I really want to talk about is where we can, or start a dialogue on, is where we can take wireless from here. I'd be interested to get your reactions at the end in the Q&A process. So from a user perspective, what we're seeing is that users have more and more wireless technologies that they're using, whether it's a mobile phone or cordless phone, Bluetooth in a computer or in a handheld, all different types, whether it's 802.11, et cetera. And there's different things that they can do with that.
For them, it's harder to keep it all straight. They're having a difficult time now to remember which steps are which for which wireless technology. Really, what they're after is, again, a level of abstraction from the technology. They want the applications or the developers in their apps to take care of that for them, so it's a very seamless and transparent experience for them. Okay.
So what I think needs to happen next is really we need to get to a stage where applications take more control over that experience and provide that level of transparency so that connections from a wireless perspective are done, are sort of task-driven and done at that level rather than at the system level. So the user can focus in on what they're trying to do, which is to create or share or collaborate, and not really worry about, I've got to connect, I have to associate, or I need to do some discovery to be able to get this to work.
The other thing, too, that I think applications should start to drive towards is being more efficient about their bandwidth and know what kind of connection they're on. And based off of that, adjust their features accordingly. And the example I always like to use is email. So if I'm on a pretty wide pipe, whether it's a wired Ethernet connection or an A0 to 11 connection, I can probably bring down my attachments and all that pretty straightforward, easily.
It's not really a bad situation for the user. But if I try to emulate that over a narrowband connection, whether it's a Bluetooth to mobile phone GPRS type of connection, you kind of almost want to have the attachments stay up on the server and the user to bring those down as they choose. But this should be something that happens transparently for the user and not something they have to remember to go do.
So one of the things I want to talk about here was really, or as an example, is talk about the meeting place and kind of what it represents today from a wireless perspective, and then perhaps some of the things we could do going forward to really take it up to the next level.
If we think about the meeting place today, it's really about people coming together where they can, again, they can share, they can collaborate, they can create stuff, and having come together, they leave having done something better at the end of that. And in that process, typically, they'll wind up connecting to different types of equipment or networks and different devices that they take with them into the meeting.
So I wanted to show you some examples of what's available today that sort of address that. Today you can actually go into a meeting and wirelessly present whatever slides you might have to an audience using an 802.11 connection. In this case, the example is from InFocus. You can also, at the end of that presentation, work on a whiteboard type of environment and collaborate and work on next steps and so forth, and then wirelessly be able to move that back over to a computer and then have it distribute via email or whatnot. E-Beam is working on something like that.
In a meeting room, you'll typically find a lot of handhelds. This is something where you could actually synchronize or send a business card over a wireless connection. Even between handhelds, they can actually communicate and chat. These are the types of activities that typically go on in a meeting with respect to handhelds.
Then there's a mobile phone. Most users will come in with a mobile phone. And here again, with using an application like AddressBook, you can filter the calls that come in, respond via SMS, and do some of those things as well as any synchronization that needs to happen while you're in that meeting.
But if you really step back and look at that, these are different kinds of things that are in that meeting room. It's really about connecting my stuff to something. And I think that the next level of wireless solutions really come from an environment where we have achieved sort of this transparency and application awareness of what you're connected to. And to transition that model so it's not just about my stuff connecting to something, but I can connect my stuff to your stuff.
So take that virtual cable from a Bluetooth connection, unplug it from my mobile phone, and maybe I plug it into your mobile phone, and we can do something interesting that way. But it's a whole other environment, I think, that's sort of untapped at this point that really allows and fosters collaboration using wireless connections in the various resources that are available in that room.
So I think this is something really to think about and sort of drive towards or work to in your plans as you're working on your wireless solutions. So in summary, what have we learned over the course of the last hour or so? Remember the matrix. So again, this is probably one of the more important positioning slides you'll see from a wireless perspective. 802.11 equals wireless Ethernet. Bluetooth equals wireless USB.
Mac users love wireless. We can see that from their early adoption of 802.11g and from the market share positions that they hold, which is much larger than what we have on the computer side of things. 802.11 has a lot of momentum behind it from a business perspective, from a product perspective, and from a connectivity perspective. So it's a really good environment to start working on solutions.
AirPort 3.1 is a great product, great update to the AirPort Extreme platform. Bluetooth is on the rise. It's getting wings. There's more and more products going out there. It's getting easier and easier to use. And it's really a great environment to start some development in from a peripheral perspective and following along the lines of some of those examples I showed you in the different applications that are available via Bluetooth. I think longer-term wireless transparency is the key to move the whole category forward for the user and really generate greater adoption and really drive it beyond where it is today.
So with that, there's a slide here that shows the places you can go to get additional information. There are three other sessions that sort of are complementary to this one that are available throughout the course of this developers conference. We encourage you to go to the one tomorrow. It is on Bluetooth, one that I mentioned earlier. It will be very interesting for those of you who want to get more detail into the profiles and such.