Speech Transcript – Jim Allchin, WinHEC 2002

Remarks by Jim Allchin
Seattle, Washington
April 16, 2002

JIM ALLCHIN: I love technology. I’ve loved it from the first time that I played with a Heathkit. Have any of you played with a Heathkit? Did you learn? I started with an analog computer. I loved those, too. I loved the first time I got the ASR-33 to just move just a little bit with the paper tape. And I really loved it. Do you remember the Diablo disk cartridges that used to be pushed in? Maybe you’re too young for this. But I loved it when I wrote the driver and it actually moved ahead. You know, I thought that was so cool. It’s why I’m still doing this. I think it’s why you’re still doing it. It’s seeing stuff that you didn’t know could be done. And this industry has come so far in such a short period of time. I was listening to Tom, this is their 11th year of doing that. I was asking backstage, well, what happened at the first one? Oh, well, it was just a few guys down in the San Francisco Airport area, you know, they’re talking about writing drivers for Windows 3.0. The industry has certainly come a very long way.

So, I’m going to talk to you about the opportunities I see. You’re going to see lots of demos during the next three days. Hopefully it will excite your imagination the way I’m excited about all the stuff I see that you’re doing, as well as that we’re doing together. And I titled the talk In Degrees because we all are interdependent on each other. If you make a system where the ambient noise is very low, you know, you probably did it by either working in some part of the thermal aspects, or perhaps that you have a chip, for example, that’s less, or perhaps you designed the box in a certain way. That has impact on other people. Or perhaps the fast boot, and perhaps you’re coming up with ideas for using MDRAM or some other type of MEMS or something in order to do the fast booting. Well, that gives other people all sorts of opportunities to do things, whether in software or hardware. So, we’re all interdependent, and I want to talk about all those opportunities that I see in the future.

I often times hear people say, oh, the PC industry, it’s commoditized. And I think, no, it’s really an opportunity to take advantage of economies of scale. There’s a lot more units out there, so you can do a lot more to touch people, which is another reason why I think you’re in the business, so you can touch people more, or you can take advantage of all those systems for sale of your product. I believe there’s plenty of money to be made, and we’re just now at the beginning of this next, if you will, decade as we’re starting this 11th WinHEC of all the things that are going to be possible.

I really believe in growing the market. But I don’t believe it’s going to happen by just cutting costs. And I think too often we focus on that. It’s about adding innovation. And if you add innovation, people will pay additional money for that value. Together, I think it’s something we have to work on to ensure that we are focused on innovation, so that you get paid the value for all your hard creative work.

I wanted to start with a sort of sobering slide, it’s sort of the state of the ecosystem from my perspective. First and foremost, the PC is still the platform. There’s no ifs, ands or buts. IDC just came out with a report that says for Microsoft products, the services products and software, hardware, software and services surrounding that, $200 billion in 2001. The next point is even more important, because the next point talks about how far we’ve come in an area where Microsoft, and frankly the PC industry was no place, which was in the — if you will — the datacenter area. PC architecture-based systems are now holding the top eight positions for TPC-c for scale out, the cluster solution. Scale up is next. And why this is so important, that same report from IDC talked about that the whole IT infrastructure revenue was a trillion dollars, moving to $1.4 trillion in 2005. The PC industry is going after that. So, we’re going to be able to, together, take this PC architected environment, and move it to the advantage of customers into areas that before just weren’t touchable. But we do need to add more value. Innovation to me isn’t changing just the color of some bezel, we have to do a lot more than that. That doesn’t fool too many people. Datacenters today still have to be managed by superhumans, and we have to do something about that. The complexity level is just much too high.

And some users still think their PC does everything they need. That’s one where we haven’t spent enough time on the value to understand what experiences that we need to really work on to get them to take that next step, buy their next machine, buy their next peripheral, buy their next piece of software, whatever it may be.

And we also have quality issues that we need to stay on top of. Together we’ve come a very long ways, and I want to thank you for all the effort that you did, certainly in Windows XP as well as in our server products. We’ve seen the difference in the type of work that you’re doing. We need to up level, again, that’s a statement to Microsoft as well as you, to ensure that when you add a driver, application, or add anything, it’s obvious and it just works. I had somebody tell me, well, you know, I have to wait to install my PC, you know, it’s going to take me a weekend to do it. And that’s just horrible, and it’s something that we have the power between us to do something about.

If we do these things, make sure that we don’t lose focus on the PC being the platform, adding more value, and we focus on the quality, that will catalyze the next wave. And what do I mean by the next wave? Well, I have four things here that I consider to be really quite fundamental for the future.

First is, from the past, the world was pretty much words and numbers. We’re moving to sights and sounds. That has a profound impact across the board in terms of the storage that you might need, in terms of the bandwidth that you might need to get to memory, to get to bandwidth, to I/O systems, the interconnection connections and what you want to do. It means quality of service is important in the space of wired or wireless. It’s a whole bunch of parameters. But it also opens up new opportunities for the PC industry to go after entertainment like we’ve never done before.

New devices, there were lots of devices before, but they were mainly isolated devices. Now those devices are all going to play together and we want to have a vision of more connected, more adaptive, automatically synchronized those there, these devices are a benefit and not more a burden, so you’re not draped with a kajillion different devices that you have to worry about synchronizing and the like.

Another key theme for the future, I think, is automation. And I mean automation in the biggest sense possible. When I talk to businesses, the thing that’s on their mind is automation of their business processes, and it could be within their company, or it could be to other companies. But there’s also the problem in a microcosm within the company. And this is about automating their datacenters. These things are just being too complicated to manage. So, we must focus on being able to make systems that are self-tuning, that are self-manageable. We need to take the focus of reliability and move it up to where it’s going after availability. And, you know that Microsoft is trying to lead the way in Web services. We also think that’s really important because it takes it from a dumb Internet world to where you can programmatically do things on the Internet, or programmatically do things within your company. So, automation is a key driving factor, and together we can up level the complexity

down level the complexity problem that people have today in their datacenters.

The last part is that we’re an IP believer. One, interconnected networks throughout the world. You’ll see our investments in things like IPV6, in technologies of IPsec and the end to end argument, and distributed firewalling and the like, and whether it’s on wireless or wired, we see this one interconnected world with tremendous opportunities when you interface it with the other areas that I mentioned.

So, those are key underlying themes. I believe the big change that were going to see over the next five years is more life immersion. People are exercising with little devices, theyve got lots of smart devices in cars, and you can go on with the set of devices and how theyre taking advantage of the power of information, music, video, or whatever. I think this trend is going to continue, and the PC itself, the thing thats in the home, we need to just make it a part of the fabric of the way somebody lives their life. When I get up in the morning, I dont read the newspaper, I read the Internet. Theres a machine in the kitchen, I flip it on, thats how I get my news every morning. We need to make those types of things very, very simple.

So what are the things that we need to focus on in my opinion? The PC has come a very, very long way. You cant really prove it, but of all the devices perhaps that have been created, the PC is probably the most, or close to the most, important device. But, were still in the early stages of simplicity, ease of use is something that were going to have to really work on. I mean, having to know what a driver is, or much worse, what a driver version is, those are things that should just go away. And I take my laptop and I put it on my lap, and it starts to burn, you know, it gets hot, you know, we need to work on things like that.

I have a home studio for music, and I really wanted a quiet PC, theyre quite hard to find, and many people have said, well, besides it being too noisy to my bedroom, its just too ugly, I dont want to put something ugly in my bedroom. And of course, it starts too slowly and has too many reboots. The TV doesnt have to be rebooted, and youve never installed a service pack on the VCR. These are just a long list of issues that together we need to work on. Microsoft cannot do this alone, we need your help. The result will be a healthier ecosystem, happier customers, more sales.

The thing that I want to emphasize at the bottom of this is that people really do want immersive experiences. Weve really learned deeply with Windows XP how much they love these end to end scenarios. And working together with the hardware peripherals companies, weve really learned how much that helps them. And when weve gone out and surveyed after the fact, its amazing. So we are going to drill harder on this in the future with immersive experiences, end to end, for entertainment, communications, creativity, productivity, and I wrote down here utilitarian, in the sense of people are going to have these things immersed in their refrigerators, and screens around their house, and their cars, and the like.

We do see many different types of PCs, and you may be concentrating on one particular area or another. And of course, theres no hard science between these boundaries, but I thought I would show this in preparation for showing my next slide, just to show you we pretty much think that there are premium PCs, there are mainstream PCs, and theres low-end, low-cost PCs. We think there are opportunities in all of these, with different margins that youre going to be able to get, depending on where youre selling into this model. But, there are improvements that we need to do to continue the growth in all of these spaces.

So what we drew out here was, together what are some of the things that I think that we should work on? Again, the categories here arent super rigid. And some areas that I mark here may belong to others. Lets start at the low end, better power management, weve come a long ways, weve still got a ways to go. We really want the fans to knock off if the system is in a quiet state. Support auto-updating, very important for customers. Deliver instant on, this has been a goal, a dream for a very long time. We still havent done it. Thats one that together we have to do. Eliminate reboot, people dont want reboots. This isnt necessarily

dont think about this just on the server side, this is on the client side, is my thoughts here. People dont want to have to reboot, and were spending a great deal of time working on this within Microsoft, but you can either help that or not, depending on how you make your devices, and your software.

Inkless input, you know a lot about our tablet, I know that weve shown it. We think that thats going to become a mainstream when we go out a few years. Implement 3GIO, theres going to be sessions on that, youre going to hear about that, some of your companies are implementing that. We think thats a very big step forward. Uncluttered user experience. We know from our usability studies that people just dont like complexity, and lots of stuff blinking or in their face. And they also dont like it in terms of the hardware. In fact, one of the demos were going to show, not in my session, but later in the conference, is being able to just plug a wire, whether it be a microphone or speaker, into either jack, and the system figures out what it is, is it a speaker or a microphone. Thats the level of making it uncomplex for the experience for a consumer, so its more resilient.

Cool and functional industrial design. I personally think there are very few designs today that just blow me away. And thats something that I want to encourage, whether its in the mainstream or the premium, I want to encourage industrial design, it does work. The number of people who have said

I wont mention the name, I was walking around with a particular laptop, and they said, wow, thats really cool, they want to know about it. That means future sales. You have to invest in order to get the return.

On the premium side, audio and video is going to come up dramatically in terms of quality, and its just going to become standard, particularly as we push into the entertainment space, into the living room, but also just on standard level PCs. Weve come a long ways, but I want to see a lot more. Arrayed microphones and cameras, the PC is the ultimate communications device. The way I talk to my mother, the way I talk to my friends is all through the Internet using RTC, the Real Time Communications Windows Messenger, and Windows XP. And if you have broadband its a first step. We can do much better than this, but hey, the arrayed microphones cut out more of the problems with the echo cancellations, if you had arrayed video then you wouldnt always look like youre looking at a different place. The camera is looking at you, youre looking at the screen, the camera is looking at a funny angle at you.

All these opportunities, and Ive already mentioned low ambient noise, this is a particularly difficult one from my perspective, because we want more processing power from the graphic system, from the CPU, we want that, but we also want a quieter system. And thats one together that I think were going to have to work on. And at the bottom of this slide is the key point, enthusiasts, power users, they will pay for more value.

One of the things I wanted to do is contrast a few pieces of technology from Microsoft to make sure its clear, because youre going to see all of these technologies through the rest of this conference. So I thought I would just set the stage right now and clarify it. The first thing I want to do is talk about Mira and the Tablet. Now, Mira is sort of like your cordless phone in your house. You cant really use it outside your house, because it needs that base station. Mira needs the PC, its a remote display that you can take with you up to a certain distance away from your base station. The Tablet is like a cell phone, because you can take the Tablet with you. You take your storage with you, you take your CPU processing with you, you take many of your other peripherals with you. So I think some people have been confused about that, think about Mira, remote display, think about it as the wireless phone in your house, think about Tablets as the PC that just happens to have a high resolution digitizer, and smart pen support that you can take it with you.

The other thing I wanted to contrast was Freestyle and the work were doing on a media focused PC, and Xbox. Now, Freestyle is a set of technologies, part of Windows, that will allow you to use remote control, and be able to move the PC, make the PC more a center part of your entertainment experience with a remote control. Mike Toutonghi later, possibly today, is going to be talking about that, and showing you that. Xbox is more focused on gaming. So yes, Xbox can play DVDs, and yes the PC can play games, but those are the centers of gravity for these different devices. The PC is one who were morphing to take advantage of a lot more entertainment capabilities, Xbox is primarily focused on gaming.

So I wanted to just do a little demo. Youre going to see a lot of demos during the next three days. The one Im going to show you is what can be really done in audio for Windows. Now, I guess the PC got its voice ten years ago, give or take, when the first media extensions for Windows came out. And since then the system has gotten better, and better. But, weve made some huge leaps up in terms of our quality just during this last year. So what Im going to show you is some technology thats coming out later this year that gives you six channels, at basically the same bandwidth usage that you would expect on dual channel MP3, and this is streaming across the Internet, this is how we expect it to be used. The clip Im going to be playing here is local, but you could stream this across the Internet. It comes as part of a new update that were doing to the operating system, this technology called Corona that youll hear more about in some of the other presentations.

So around the room are a set of speakers. This is beyond 5.1, this is a full six channels. So imaging that youre watching a video, and Im going to pull up a little video so you can see what that sounds like, and just imagine what its going to be like sitting, watching this, streaming across the Internet.

(Video shown.)

JIM ALLCHIN: I love technology.


JIM ALLCHIN: Let me give you an idea of what music sounds like when it’s converted into six channels. This is just a set of different clips.

(Music clips.)

JIM ALLCHIN: You get the idea. Bill later will be showing you what we’re doing in terms of video later this year, Bill Gates, on Thursday. This requires hardware support, as you might expect. It requires speakers, which can get very intelligent. It requires processing power in terms of the sound card. But it’s an opportunity, and it’s just the next step for us to be prepared to move into a richer experience in the living room.

I want to change and talk about enterprise servers. I talked about life immersion, and the servers, the heart of it is toward automation, as I mentioned. Now, some of the areas that are at the top of our mind are virtualization. You want to take a rack of computing modules and be able to treat those things as a scale and variant set. So, we want to be able to have a Web service, or something on there. If you need more power, you can just be added. And this might be in a scale out environment, or it’s possible that it might be in a scale up environment. But we want to make those computing modules virtualized.

Another thing we want to do is not just stop with the CPU. We actually want to take it to the networking and storage. The storage is quite difficult and an area where we have just invested quite a bit of resources to try to work with a set of partners to try to address the virtualization of the storage space. And when we take it from being an extremely proprietary, just notch it back a little bit, a lot more of you can play in an area that the PC industry hasn’t been able to play in yet. But I think the results for consumers and in the datacenter is going to be dramatic. I mentioned management. This is an area where we have an effort underway to really focus on self-configuring systems, self-tuning, auto-updating and the like. This doesn’t mean just Microsoft, though, the problem of BIOS updates and firmware updates is a real problem.

I bought a naked machine, and the server, which is a raised server, and the like, and I installed it all. And, you know, it’s complicated. It’s complicated because you have to have the right firmware, and you’ve got to load the right driver in the right order, and hit this function key at the right time. Not the way it should be. Diagnostics is something that we need to get better on. People never want these machines to go down. We need to be able to figure out what’s going on with them in a much simpler way.

You’ve all read about, I’m sure, the Trustworthy Computing effort that we have underway at Microsoft. I would like to suggest that that applies to the entire industry. Together, we need to work on resiliency, and security, and reliability. I also think we have to make good on the promises that we say. If you’re talking to a customer partner, we should try to make absolute good on what we’re saying. We have a statement by Microsoft, we want to be more transparent. Well, we have conferences like this, we try to tell you exactly what we’re doing. Sometimes we may have to change our minds due to marketing conditions, but we’re trying to be as open about what it is that we’re doing as we possibly can.

Collectively, what do I think we need to work on? Well, I think we need to make sure that we build the right size servers, whether they be high end 64-bit, or whether they be small rack-mount units that are 1U factors.

We are committed to dynamic partitioning. This is something that we worked on a little bit in .NET Server. So, there’s a lot more to do. NUMA support, again, that’s something that we’ve worked on a little bit, there’s a lot more to do. Dense server form factors, you’re going to end up with all sorts of heat problems and the like that together we’re going to have to concentration on, not to mention the management problem.

Robustness, we need to be able to robustly and quickly fail over machines, or build it into the hardware so it’s totally transparent, something like what Stratus has done. Watchdog timers, device hardening, those are all efforts that together we need to work on in order to bring up the level of robustness that customers want in a datacenter.

Easier to deploy, we added the ability to add memory in Windows .NET Server. But there’s a lot more to do if you want to do hot swapping of the CPUs and the other modules.

Secure remote boot is something that we’re working on, working on with some of you. And that’s very important so that customers feel comfortable that if a naked machine, server or client, ends up booting that the right things will happen to the right servers in an authenticated way. ACDI 2.0 and EFI are replacing legacy BIOS, a lot of advantages with that.

Interconnect, well, PCIX is still around, it’s going to be around. The future is 3GIO, one that the industry is going to move to, Infiniband, and we’ve really appreciated the work that we’ve done with many of you on this. It’s been a great collaboration. Interconnects will get very, very fast. Maybe I should have put greater than 10 gigabit Ethernet here, because that’s really where our mind set is.

Multitasking I/O is something that takes a little bit of work on your part, but the benefits for customers is very high, and one that in some of the sessions we’ll talk about exactly how to do that.

I want to bring out Valerie See to do a little demo of what the effort that we’re doing about dynamically scaling out datacenter. We have a datacenter here, and Valerie is actually bringing out her little datacenter with her.

VALERIE SEE: Hi, Jim. Well, I brought this little server so that we can show you the full range of our management technology that we’re working on for .NET Server in the future. So, let me go ahead and hook this guy up. And one of the problems that we’ve heard people telling us a lot about is, they hate how long it takes to deploy new servers. So, we’ve been working on that problem, and this is an example of some of the future technology that we have that will make it really, really fast to deploy.

So, what I’m going to do is go ahead and turn this guy on, and I’m not going to plug the net cable in yet, because what I want to do is show you, this is a barebones box. We don’t have anything on it right now. It’s going to boot to a network prompt, and then fail because I’m standing here holding the networking cable.

So, once we get through this and you get to the net boot prompt, we actually have a timer, because I don’t think you’re going to believe how fast this is really going to boot once we get it going. We’re almost finished.

What this does is, it basically allows administrators to build an image that we want to be able to just blow down with absolutely nothing else happening. If I had plugged the networking cable, we could have skipped this. And when I turn it on, we can just skip this next step. But if you want, why don’t we time this, if we can count down from three. Three, two, one. Okay.

Now, what’s going to happen is, you see we’re starting a load from the network, and we’re going to basically pull a preconfigured image that the administrators would have ordinarily just been able to plug this in, turn it on, and it will start loading Windows. And before 60 seconds is up, we’re going to have a full load of an appliance operating system on this. This is kind of the low end of our management technology.

But why this is important is that it’s not only for customers who are telling us they want to be able to turn their datacenters over like this, but it’s also for our OEM partners, for people who are putting loads on their systems, the faster they can do this on the line, the more margin they get out of the platform, and the more opportunity they have for expanding their business. And, as you can see, we’re just almost ready, and we’re almost booted, and there we go, and we had seven seconds left. Under a minute, full operating system load for an appliance system.

JIM ALLCHIN: And this could be any type of system. We just put Windows XP embedded here, but it could be any of our appliance characteristics inside Windows XP, or a whole Windows Server.

VALERIE SEE: That’s right.

JIM ALLCHIN: So, how about the big guy here?

VALERIE SEE: So, using the C management infrastructure, what we’re going to look at is some more of the automation issues that you were addressing earlier. One of the things that we know is, you’ve got to be like a rocket scientist to drive a datacenter. This is a very easy interface. It’s task-based, but it’s fully configurable.

So, what we can do is take a look at the devices in our datacenter. We can see what their activity level is. We can tell everything is up, running, it’s all healthy. We can also, by our choice, order these systems into various sets. We can do it on the basis of functionality, or we can do it on the basis of physical location. We brought two racks of servers with us. Our far one is a brand new rack, and we’re going to show you some things on that rack. We’re going to do some things with it.

The near one is divided up into a group of Web clusters, and we’re going to look at one of our running Web clusters that currently has two nodes in it. And if we can take a look at the performance monitor on that cluster, you can see it appears that it’s pretty loaded right now. We’re taking a lot of requests. We don’t have much reserve capacity. We really need to add some capability to the cluster.

So, that far rack, the upper-most server is barebones. We don’t have anything running on it right now. So, if we could switch to that, and show that it’s sitting in

— that’s a little lightweight operating system kernel that we build as basically a basis for this management automation. It boots up from the network, and it’s sitting there right now waiting for us to give it a command.

So, from my remote Web-based interface, I’m going to tell it, I want to deploy an image to that box. And I’m going to just run it, and you’ll notice up on the screen that we will have received a command, and it’s started to load.

JIM ALLCHIN: So, that’s all activity here.

VALERIE SEE: Right. Once that thing is finished, I could then add that to my Web cluster, and give it some more resources. Now, in the interest of time, I happen to have a preloaded image, one of those other 15 servers

JIM ALLCHIN: Just like you did there.

VALERIE SEE: Just exactly right. And we’re going to go ahead and just show you how easy it is to dynamically add resources to a cluster. So, let me go ahead and add that. If we could go back and take a look at the performance, you’ll notice we’re still fully used up.

Let’s go ahead and add that new node to our cluster, and what’s going to happen is, as the node gets added, you’ll see the performance monitor trace take a real nose dive. And then once it reconverges, it’s going to come back up and it’s going to be running at a level that shows that we’ve got reserve capacity now.

So, there we did it. Capacity on demand, fully remote, I didn’t have to walk over there, you didn’t have to walk over there, nobody had to do anything except right here.

JIM ALLCHIN: So, what about a patch?

VALERIE SEE: Now, that’s a good question, because I just told you those are all new servers. We ought to be checking what level of security we’ve got them set at, look through all our QSCs. So, let me do a quick query on that. And we’ll go ahead and see if we managed to get all the right security caches in because we know that customers really hate that. It’s a really arduous process. And, okay, we’ve got to fix it. We don’t have it on any of them.

So, let’s go fix that real quick. Here we’re going to apply a security patch, and just to show you this is real-live stuff, let’s take a look at the wizard that will walk you through these preconfigured tasks. So, what we do first is, this is where you’ve actually defined what the task is that you want to do. And this is a real-live query straight from Microsoft.com, nothing funny about it. We have a straight download. And we’ve applied some parameters so it will run silently and install without any intervention on our part.

The next thing is, we can choose where we want to deploy this. We can either deploy it to a certain device, or since we happen to need to deploy it to that whole rack, we’re just going to blow it onto the whole rack.

So, that’s all there is to it. Let’s go ahead and do it, and you’ll notice that the lights on that far server rack over there are flashing. You can see we’re pulling the code into it, flashing over the network. We’ve got the job in progress now. And, in fact, if we go ahead and do that query again that we did the first time to see if we actually got the QSC installed on those servers, now we should come back and we’ll see yeses instead of nos.

JIM ALLCHIN: All right.

VALERIE SEE: There we go.

JIM ALLCHIN: This is great. Right now, you’d still have to reboot those?


JIM ALLCHIN: And one of the things that we’re working on is a technology called Hot

Patching, so our goal is to be able to do most patches while keeping the system running.

VALERIE SEE: Exactly right.

JIM ALLCHIN: This is great.

VALERIE SEE: Now, it turns out, while we were talking, remember that server we were initializing, we can see it’s just booting right now, it’s almost ready to add as our next Hot Spear for the cluster, and that’s all it took to deploy a new server from barebones metal just through the remote interface.

JIM ALLCHIN: That’s great. That’s exactly the type of vision that we have in terms of automation.

VALERIE SEE: Yes, sir.

JIM ALLCHIN: Thank you.


JIM ALLCHIN: We are very close to releasing the .NET Server family. It will be later this calendar year, from at least a manufacturing perspective, probably be in customers’ hands some time next year.

I tried to write a slide to show you the relationship between all the products that we’re making available there. There will be a Web server-focused dedicated system; standard server, mainly oriented towards file and print and simple collaboration; the Datacenter; the Enterprise Server for high end infrastructure capabilities; as well as the super high end Datacenter. Separately we have a small business server, which is trying to work on a complete solution, at least in terms of the infrastructure pieces for a small business that, again, applications can be written on top of it. So were making very good progress, we appreciate all the feedback that youve given us, its going to be a rocking release when it comes out.

In terms of networking, one of the other areas that we talked about earlier in terms of themes, where do we see it going? Were hardcore believers in 802.11x, as well as Bluetooth. And we think that GPRS supplements the WiFi. We see interconnections between those as being really important, and youll see a little bit of that throughout this conference. We also see 1394 as being of growing importance, because we want to be able to attach to the consumer electronics devices, and those will have 1394. Its possible that 1394 could become more than just connecting to those consumer devices, if you apply IP on top of it, and use it for other technologies, other devices being connected to it.

We see IPV6 as the real catalyst, to be able to handle all the devices, to be able to get through NATs, in order to deal with the complexity problem and get to the simplicity that we need. We could have done work on IPV4 for mobility, and we have opted to go for IPV6, because we think its such a stronger solution there. And of course, weve already had IPsec for some time, and that will continue to be evolved. IPV6, though, we consider to be the strategic direction sort of across the board. Now, we wont be running it as default coming out of .NET Server, but it will be there, and we will encourage people to turn it on, and start to move to it. Collectively, whether it be in the hardware or software, you need to look at all the ways that you can add value to IPV6, whether it be offloading, which is probably one of the things that comes to mind, or whether youre doing things in the mobility space, how you can fit into that, because that is going to take the Internet to the next level.

Together in the wireless space theres a whole set of things, weve come such a long ways in wireless, but we still have a long ways to go. Although Windows XP is pretty good at zero config, there are still more things that together we need to do. Bandwidth aware applications, to take advantage of the optimizations, the interfaces that we have, that were publishing, seamless roaming between hot spots, say at airports or hotels, and your home, at work, and we also see an opportunity to take advantage of the CPU processing to do more in terms of the processing that has typically been offloaded. And depending on how you look at it its either a good or a bad thing, it seems sort of inevitable. So its something that youre going to hear in one of our sessions later on at the conference.

We do see the PC being able, as a supplement, being able to used for very small, little, home environment, as an access point. Were working on some technology there. We see that wireless is going to come in lots of different shapes, and colors, and sizes, so were hardcore lovers of 5G, we just need to get going, and get the cost of that technology down. Bluetooth and WiFi, as I already mentioned, and the WAN space, and WiFi, all of those things need to be seamless as youre moving between them, thats the point of that bullet.

In terms of the backing and the interconnects, theres been a lot of work done in Infiniband, and other high speed technologies. Well, Winsock Directs, and Sockets Direct protocol are what we really encourage you to use for doing the interconnects, the benefits are that so many applications will be able to take advantage of your high speed interconnects if you do that. So I wanted to do a little demo of some integration that weve been experimenting with, and we feel very happy about the progress there. Its some technology form Intersil and Silicon Wave that have made a combined Bluetooth and WiFi dual-radio system. So on this part of the stage, you can see that there is a Toshiba machine here, with that dual radio, and you can see the wireless connection is life, thats an 802.11 system. I happen to have a Bluetooth mouse, a Bluetooth keyboard, I happen to have an HP Bluetooth printer, and I have a Sony Ericsson Bluetooth phone. So two years ago the rage was, you couldnt make this stuff work together. One year ago somebody wrote a paper and said, I think I know how to do that. Well, let me just grab the mouse here and bring up the Internet.

So were browsing

good, I like that.


JIM ALLCHIN: Im not to be

somebody tell me which one? It doesnt work out there either, so the wireless network is down is what youre telling me.

Well, maybe we wont be demoing this. The mouse and keyboard are working.


JIM ALLCHIN: Somebody in the audience asked me to say that. And this actually has a sheet of paper in it for the printer. So Im not going to be able to do much of the demo that I planned to do here. This does work without interference, its very cool. Its a mini-PCI card that can fit in any of these machines that you can build it in. I think its super cool, and youll hear about this later. I expect lots of different Bluetooth peripherals.

One of the things thats very cool, because we are trying to listen to you about innovation that you can do, its my point about commoditization. We did a base Bluetooth stack, and you can do your own set of extensions on top of it, so whether it be sync between, say, the address book for your phone and the PC, you can add that all on top. And so you can add your own proprietary value there, and still get the value, if you will, of the economies of scale. So enough of that demo.

In terms of the Windows roadmap. Let me just highlight a few things. In the middle of this year well be coming out with Windows XP Service Pack 1, well have the Tablet Edition coming out just a little bit behind that. We have as a mirror, which is the remote display capabilities, as well as freestyle, which is the technology that I mentioned about remote control being able to move into the living room. In terms of the servers, Ive already talked about those coming out, RTMing later this year.

In terms of embedded, we came in with an Embedded XP system at the same time that we came out with XP, give or take a little delay, and we also came out with the Embedded CE system. When the .NET Server family comes out, there will also be a .NET Embedded Server capability, so if youre building capture devices or server appliances and the like, youll be able to take advantage of that.

The next big wave for us is beyond 2003, which were working on very hard, and Longhorn is the code name. Were going to synchronize a tremendous amount of our technology behind that. A significant number of people are working on it, and throughout the conference you will be seeing little pieces of that technology being shown.

I want to end by just sort of repeating what Ive said in terms of our philosophy, and how I see the opportunities going ahead in the future. PC architected systems will continue to morph. You will see systems the size of the Pocket PC running Windows XP. Youll have all the advantages of taking with you your storage, and all your capabilities. Youre going to see that very soon. Ive seen prototypes running of this. Youre going to see it. Youll see us move up, even higher, so whether its in the ultra-mobile form factor, or the extreme end, the PC architected systems are the ones that give customers the most advantage, and the ones that we have proven time and time again that we can continue to morph.

New experiences will inspire additional PC sales. Think about just the sound system that I demonstrated in here. Think about what you could do if you had great array microphones, and doing video conferencing and the like. Automation is the key to datacenter agility. We want to get in the datacenter, wed better make it simpler, better continue to scale up. And seamless, secure connectivity is a requirement, not a luxury, thats something that seems pretty obvious, but Im not sure that everyone understands that when I say seamless it needs to be able to move between all the different technology types. Thats what customers are expecting us to do.

Im a hardcore believer, economy or no economy, innovation is the way to profit. I appreciate the fact that were able to work with you, and how large this is, and Im very proud of our industry and all the things weve done. We need to continue working on quality. If we work on quality, we work on solving real customer problems, then you will be rewarded, which is the last thing I wanted to say.

I hope you have a very good conference, there are very great speakers coming throughout the days. I just highlighted here the executive overviews that are happening. Mike Toutonghi later today, Bill Veghte covering the Enterprise Server, Windows drivers, you should go to this one. Im counting on you to do a great job in terms of the quality. Were working on the tools that youve asked us for. Rob Short and Brad Carpenter are doing that. The Windows client with Joe Peterson, youll get a vision of what the future looks like there, and what were thinking about, interoperable networking. And youll get even more demonstrations in the capabilities of moving seamlessly between the roaming in your home, and other roaming capabilities.

Again, invest and you will be rewarded. Thank you very much.

Related Posts