You're going to have to focus on building a machine-to-machine infrastructure, one that goes beyond the one that we know today and contemplates the fact that there is a machine-to-machine infrastructure on the receiving end—in people's homes, in their cars, in their pockets and purses.
Networks will have to continue to be enhanced, beyond what people dream about today, because this revolution in usage will dramatically stress future networks.
Security is going to be a big focus. We're going to have well-executed public key cryptography systems as well as digital signatures and certificates, and individuals are going to be able to use them—just as they get driver's licenses and passports and credit cards today. That has to be integrated in the way in which our businesses interact with them.
We're going to need rich schemas to describe all these information and process components so that the machine-machine interaction can take place without explicit user administration. So the XML revolution will be profound because it's an enabling technology for the new class of services that people want to discover and interact with. If we actually do all this well, it will result in more direct access to these consumers for business and more direct access by consumers and all types of customers to the capabilities that we all have. It really is the beginning of a revolution. The Internet we know today is an adolescent and there is a lot of growing up to do. We look forward to working with all of you to do this.
Monday, February 11, 2008
Problems. Big Problems.
There are some big problems to solve. The first is just relevance. If people don't think that these improved tools or devices are really relevant to what they want to do, then they won't use them. Similarly, if they're not simple, intuitive to operate, and self-organizing, then they won't use them, either.
As we move more and more of our traditional businesses and applications into these environments, the issues of identity and privacy become very, very important. The Internet has already raised people's awareness about privacy issues. We don't have complete solutions in either the technology or policy space for that, but I think people are quite focused on it. But to do a lot of these things requires better identity technology than we have today. Today we have credit cards, our physical signatures, and picture IDs that are ways of documenting our identity, but as we operate in this cyberworld we'll need computer identities that are provable in similar ways. So we have to create the infrastructure that allows that to happen.
Reliability will be very, very important as well. For this to happen, the home environment has to be more self-healing.
Security is a big issue. If all our business assets, our relationships with customers, and our personal information are to be contained in this environment, we're going to have to consider a lot more of our business computing to be a national-security-critical infrastructure, even more than we have today.
The Y2K problem and the viruses problem are just early-warning indicators of problems that could occur when malicious people decide they want to attack society's new weaknesses in the form of computerized infrastructures. More emphasis is going to have to be placed on this, at the policy and law-enforcement level, but ultimately at the level of computer architecture as well.
We're all going to have more and more dependencies on computing now. In the past, once society has begun to adapt and depend on a new technology, whether it was cars, electricity, plumbing, or telephones, it really doesn't tolerate its loss very well. Computing in business has largely reached that state, but even there we occasionally teeter on the brink of not being able to keep it all together. As we diffuse new computing infrastructure into everybody's daily life, and we don't have professionals who can help every single person at home, we clearly have to make it something more reliable that people can depend on completely.
As we move more and more of our traditional businesses and applications into these environments, the issues of identity and privacy become very, very important. The Internet has already raised people's awareness about privacy issues. We don't have complete solutions in either the technology or policy space for that, but I think people are quite focused on it. But to do a lot of these things requires better identity technology than we have today. Today we have credit cards, our physical signatures, and picture IDs that are ways of documenting our identity, but as we operate in this cyberworld we'll need computer identities that are provable in similar ways. So we have to create the infrastructure that allows that to happen.
Reliability will be very, very important as well. For this to happen, the home environment has to be more self-healing.
Security is a big issue. If all our business assets, our relationships with customers, and our personal information are to be contained in this environment, we're going to have to consider a lot more of our business computing to be a national-security-critical infrastructure, even more than we have today.
The Y2K problem and the viruses problem are just early-warning indicators of problems that could occur when malicious people decide they want to attack society's new weaknesses in the form of computerized infrastructures. More emphasis is going to have to be placed on this, at the policy and law-enforcement level, but ultimately at the level of computer architecture as well.
We're all going to have more and more dependencies on computing now. In the past, once society has begun to adapt and depend on a new technology, whether it was cars, electricity, plumbing, or telephones, it really doesn't tolerate its loss very well. Computing in business has largely reached that state, but even there we occasionally teeter on the brink of not being able to keep it all together. As we diffuse new computing infrastructure into everybody's daily life, and we don't have professionals who can help every single person at home, we clearly have to make it something more reliable that people can depend on completely.
Network Demand: Computers Never Sleep
As we move from a very static environment, just reading text and seeing pictures on Web pages, and start to blend this into our entertainment experience, it's clear that you're not going to do this by the same mechanisms that deliver Web pages today or allow us to send an occasional piece of data back.
While networks are getting faster all the time, the reality is that networks today are primarily only serving up those Web pages for us to read. The fact that we humans consuming the Web pages don't read or write nearly as fast as the computers do is why there isn't more congestion on the network. But when you get to get this world where computers are talking to each other all the time in order to help you out, the pull on the network is going to go up exponentially, because computers never sleep, they don't get tired, and it doesn't take them very long to read anything. So we're going to go through another big demand phase on the network side, and that'll force us back to creating a better balance between what we compute locally and what we'll actually ask the network and servers to do.
The impact of broadband and wireless networks is hard to estimate, but I think it's going to be quite profound. Virtual meetings, for instance, will change the way in which you can communicate with your employees. You're going to have new entertainment services—streaming video, enhanced television, the e-book. The Internet will allow us to have multi-player gaming, not only the kind that the kids play, but the kind of strategy and card games that grown-ups play.
While networks are getting faster all the time, the reality is that networks today are primarily only serving up those Web pages for us to read. The fact that we humans consuming the Web pages don't read or write nearly as fast as the computers do is why there isn't more congestion on the network. But when you get to get this world where computers are talking to each other all the time in order to help you out, the pull on the network is going to go up exponentially, because computers never sleep, they don't get tired, and it doesn't take them very long to read anything. So we're going to go through another big demand phase on the network side, and that'll force us back to creating a better balance between what we compute locally and what we'll actually ask the network and servers to do.
The impact of broadband and wireless networks is hard to estimate, but I think it's going to be quite profound. Virtual meetings, for instance, will change the way in which you can communicate with your employees. You're going to have new entertainment services—streaming video, enhanced television, the e-book. The Internet will allow us to have multi-player gaming, not only the kind that the kids play, but the kind of strategy and card games that grown-ups play.
Death of the PC: Greatly Exaggerated
As the browsing version of the Internet has dominated people's thinking in the last couple of years, there was some discussion that perhaps the world would just move back to centralized computing. Certainly there's some efficiency to a return to time-sharing, if you will. But it turns out that there are many aspects that ultimately preclude that from being the dominant way in which people will use computers—partly because we're going to diffuse computers into so many devices. The capacities are going to continue to increase not only at the server end, but also in each of these clients. Today's Pocket PCs are substantially more capable computers than a desktop personal computer was five years ago, and it's probably not much behind where it was two or three years ago.
The trickle-down effect—where we can inject technology at the high end in the commercial environment in servers and professional desktops, then have it trickle down into all the other types of devices in our life—will create an ability to exploit that power in terms of more sophisticated software and applications. Network performance is going up, too, not only in the backbone and enterprise environment, but starting this year it'll be in your homes.
It's clear that not everything is going to run on a server. Why? The world won't have a ubiquitous, persistent wireless infrastructure anytime soon. So offline operations are going to be important because all these other devices are primarily targeted at personal, portable, and mobile uses. Handwriting recognition, speech, gesture recognition—all these new means people will use to interact with machines—demand a level of computing power local to the device. We're stuck now at the tail end of what I think of as the mouse-and-keyboard generation of computing. As people start to interact with this environment in a more natural way, they're going to demand the ability to exploit these incredible capacities that will live at the ends of the wires, not just the center.
The trickle-down effect—where we can inject technology at the high end in the commercial environment in servers and professional desktops, then have it trickle down into all the other types of devices in our life—will create an ability to exploit that power in terms of more sophisticated software and applications. Network performance is going up, too, not only in the backbone and enterprise environment, but starting this year it'll be in your homes.
It's clear that not everything is going to run on a server. Why? The world won't have a ubiquitous, persistent wireless infrastructure anytime soon. So offline operations are going to be important because all these other devices are primarily targeted at personal, portable, and mobile uses. Handwriting recognition, speech, gesture recognition—all these new means people will use to interact with machines—demand a level of computing power local to the device. We're stuck now at the tail end of what I think of as the mouse-and-keyboard generation of computing. As people start to interact with this environment in a more natural way, they're going to demand the ability to exploit these incredible capacities that will live at the ends of the wires, not just the center.
Universal Plug and Play
These technologies are very important because, if we're going to see this adopted in small businesses and home-offices, we need to give them the same facility that we're giving the more technologically naive consumer. So there will be a spillover back into the enterprise-computing environment from the technology that we're being driven to create to deal with this simpler environment.
One of those technologies we call Universal Plug and Play (UPnP). It's a way of exploiting these networks and the intelligence in the devices, but it's also a great example of what XML provides. Universal Plug and Play is completely built around XML technologies and basic Internet protocols. Why? Because the devices have to be able to describe themselves. Before I can take my WebTV and ask it to call up my bank record, or call up my financial or health record on the PC in the office or out on a secure site on the Internet, the devices have to be able to find each other first, and then find the services that can give them access for those things. And we have to do this without having an IT professional come in and do some incantation over it.Machine-to-Machine CommunicationSo we've taken XML and these Internet protocols, and created an automated way of dealing with addressing and naming so that devices automatically connect themselves to networks. Then we use the same technology to describe the device and its capabilities, and to advertise the kind of services it provides and the kind of services it would like to consume. This is the beginning of an infrastructure—what I call the machine-to-machine communication of the maturing Internet—so at the end of the day you perceive a simple experience. You want to sit down and ask it to give you your medical record or your portfolio information. You don't want to think about the plumbing that's underneath.
About 125 companies have joined the Universal Plug and Play Forum. What are they doing? They're coming together with their engineers and agreeing on the words that we are adding to the XML language to allow the machine-to-machine description of all these devices. Companies like GE and Maytag are in there, describing washing machines and dishwashers, and companies like Sony and Matsushita are describing VCRs and televisions. Industry by industry, we not only have to agree on the way we're going to describe the data that goes back and forth, but we're ultimately going to have to describe the services we want to provide. And we're going to have to describe how all these new computerized devices come together. So XML is an important technology because it gives us a structured way of describing all of these different kinds of capabilities.
One of those technologies we call Universal Plug and Play (UPnP). It's a way of exploiting these networks and the intelligence in the devices, but it's also a great example of what XML provides. Universal Plug and Play is completely built around XML technologies and basic Internet protocols. Why? Because the devices have to be able to describe themselves. Before I can take my WebTV and ask it to call up my bank record, or call up my financial or health record on the PC in the office or out on a secure site on the Internet, the devices have to be able to find each other first, and then find the services that can give them access for those things. And we have to do this without having an IT professional come in and do some incantation over it.Machine-to-Machine CommunicationSo we've taken XML and these Internet protocols, and created an automated way of dealing with addressing and naming so that devices automatically connect themselves to networks. Then we use the same technology to describe the device and its capabilities, and to advertise the kind of services it provides and the kind of services it would like to consume. This is the beginning of an infrastructure—what I call the machine-to-machine communication of the maturing Internet—so at the end of the day you perceive a simple experience. You want to sit down and ask it to give you your medical record or your portfolio information. You don't want to think about the plumbing that's underneath.
About 125 companies have joined the Universal Plug and Play Forum. What are they doing? They're coming together with their engineers and agreeing on the words that we are adding to the XML language to allow the machine-to-machine description of all these devices. Companies like GE and Maytag are in there, describing washing machines and dishwashers, and companies like Sony and Matsushita are describing VCRs and televisions. Industry by industry, we not only have to agree on the way we're going to describe the data that goes back and forth, but we're ultimately going to have to describe the services we want to provide. And we're going to have to describe how all these new computerized devices come together. So XML is an important technology because it gives us a structured way of describing all of these different kinds of capabilities.
Honey, I'm Home. Where's Our CIO?
But if we're going to build this environment, we have to recognize that there can't be a CIO for the home. And as consumers and businesses interact with each other in the Internet environment, there can't be a CIO for the world either. Enterprise computing has had a big assist, in terms of how we use and deploy the technology, from the professional IT organization. There's always been a dynamic balance, almost a tension. Before we could add more technology, we had to have more trained people to use it, and that becomes a limiter to how fast we can deploy technology. It's easier to add technology than it is to get people to adopt it. As we take this out of the enterprise, where there are IT professionals to help with that process, and put it in everybody's life, it's clear that people will look at the network and all these devices as they look at the power grid today: there's just a jack in a wall, you buy a lamp or a fan or a toaster, you plug it in, and it just works.
In the home, you can't demand that people install new wiring infrastructure. So we have to piggyback the networks on the wiring infrastructure that's already there, or use technologies that allow wireless connectivity. Occasionally there will be new wires added, particularly for very high-quality video distribution in the home. But, by and large, the power line, the phone line, and the air will be the dominant ways in which most of these devices in the home get connected.
So personal computers and other specialty devices will become a bridge between the broad Internet environment and this quite sophisticated set of devices that people will have in their homes.
In the home, you can't demand that people install new wiring infrastructure. So we have to piggyback the networks on the wiring infrastructure that's already there, or use technologies that allow wireless connectivity. Occasionally there will be new wires added, particularly for very high-quality video distribution in the home. But, by and large, the power line, the phone line, and the air will be the dominant ways in which most of these devices in the home get connected.
So personal computers and other specialty devices will become a bridge between the broad Internet environment and this quite sophisticated set of devices that people will have in their homes.
The Next PCs: Office or Home?
My view is that the personal computer will continue to evolve, of course, but it will bifurcate into, on the one hand, devices that are optimized for professional use, and on the other hand, devices that are optimized for personal and home use. These devices will link to these new networks—the broadband connections to the home, the wireless and wired connectivity within the house. These are the things that will create the next generation of the Internet.
All the devices that people already have in their homes are going to get microprocessors embedded in them. As a result, the companies that make them will change their relationship with the consumer, providing consistent support and maintenance of the devices, and gaining the ability to market additional services and features to consumers.
This computing milieu is going to redefine the way businesses have to deal with consumers. The buyer is getting more and more of the power in this. All these devices are going to create enough points of presence for consumers to redefine the way they deal with entertainment, communication, and shopping.
All the devices that people already have in their homes are going to get microprocessors embedded in them. As a result, the companies that make them will change their relationship with the consumer, providing consistent support and maintenance of the devices, and gaining the ability to market additional services and features to consumers.
This computing milieu is going to redefine the way businesses have to deal with consumers. The buyer is getting more and more of the power in this. All these devices are going to create enough points of presence for consumers to redefine the way they deal with entertainment, communication, and shopping.
What It's Meant to Microsoft
For Microsoft, as a software company we have had to deal with a dramatic expansion in the requirements of our product line. We've always been in the business of selling people platforms that have tools and technology to build products. In the past those things have primarily been personal computers. We also have sold experiences in the form of application programs, so when people buy Microsoft Word, Microsoft Excel, and Microsoft PowerPoint® as Microsoft Office, they're living an experience in those programs that allows them to communicate or write documents. But the world is expanding quite a bit. The applications are becoming more diverse, and they're spreading out to cover all these other new devices, like Pocket PCs. As the applications become more sophisticated, we need the middleware environments of back-end servers and browsing technologies to allow people to build these experiences on top of platforms. And, as we move into other devices that are not all of the same level of capability, we have to take subsets of the technology—for example, the media technologies—and allow them to be put into many different devices.
The Internet has created another radical change. There are service components of these experiences and, in the future, the platforms themselves that the company is going to sell. For us it's a very important change, and we've learned a lot about this in the last few years. For example, we have the Microsoft Money application, which people use to balance their checkbooks at home. But through MSN™, Microsoft MoneyCentral®, and Microsoft Investor, we've realized that people want a linkage between the application that they use to manage their personal finances and the online services that allow them to pay bills or manage their portfolio in real time.
We've also learned, for example, with Microsoft WebTV® that it's important to have platform technologies that have service components, too. Why? To eliminate the complexity of managing the software. That's a very vibrant environment in all these emerging technologies. You want platforms that can be self-organizing, self-maintaining, self-healing.
The Internet has created another radical change. There are service components of these experiences and, in the future, the platforms themselves that the company is going to sell. For us it's a very important change, and we've learned a lot about this in the last few years. For example, we have the Microsoft Money application, which people use to balance their checkbooks at home. But through MSN™, Microsoft MoneyCentral®, and Microsoft Investor, we've realized that people want a linkage between the application that they use to manage their personal finances and the online services that allow them to pay bills or manage their portfolio in real time.
We've also learned, for example, with Microsoft WebTV® that it's important to have platform technologies that have service components, too. Why? To eliminate the complexity of managing the software. That's a very vibrant environment in all these emerging technologies. You want platforms that can be self-organizing, self-maintaining, self-healing.
The Third-Wave Web
But the real important thing is the maturing of the Internet. Now we have XML (Extensible Markup Language), which gives us the ability to describe many things. That's very important because in this world where we're trying to get computers to come together and help people do things—whether it's have a meeting, or place an order, or access a medical record—we need to have ways for the computers to understand how to talk to one another.
This maturing Internet will be characterized by a few things. One will be broadband and wireless connectivity, whether it's on a broad-area basis or a local-area basis. The connections more often than not will be persistent, as opposed to occasional. One of the things that is going to be very profound, though it's a little hard for people to imagine, is that the next generation of the Internet will be about machine-to-machine communication through the Internet, not just about people reading Web pages. If we do that well and have adequate software, then we should be able to have a significant improvement of our quality of life at work and at home—because these computerized devices are going to help us get things done.
Whose Work Is It, Anyway?Ultimately it's all about creating richer experiences, where the machines can do the work so people don't have to—or at least not as much as they do today. And by work I mean not only the task at hand, but also the work of learning how to use these computers and to master the complexity that has grown up around them.
We're going to have many new modalities of man-machine interaction. We're going to have the capability for these machines to have vision as well as speech recognition and synthesis. The machines and the networks have to become more self-configuring; the use of XML and some of the new technologies give us the ability to do that.
This maturing Internet will be characterized by a few things. One will be broadband and wireless connectivity, whether it's on a broad-area basis or a local-area basis. The connections more often than not will be persistent, as opposed to occasional. One of the things that is going to be very profound, though it's a little hard for people to imagine, is that the next generation of the Internet will be about machine-to-machine communication through the Internet, not just about people reading Web pages. If we do that well and have adequate software, then we should be able to have a significant improvement of our quality of life at work and at home—because these computerized devices are going to help us get things done.
Whose Work Is It, Anyway?Ultimately it's all about creating richer experiences, where the machines can do the work so people don't have to—or at least not as much as they do today. And by work I mean not only the task at hand, but also the work of learning how to use these computers and to master the complexity that has grown up around them.
We're going to have many new modalities of man-machine interaction. We're going to have the capability for these machines to have vision as well as speech recognition and synthesis. The machines and the networks have to become more self-configuring; the use of XML and some of the new technologies give us the ability to do that.
The Integrated Computing Experience
The First-Wave Web.The early Internet, which lasted about 25 years, started with academic and government work. Computers and workstations were the dominant tools that people used in developing this technology. The telephone connections were all dial-up lines. Basically, computers were masquerading as people on the world's telephone network. Occasionally people would have leased lines and ethernets that they connected. The protocols were for simple applications. The earliest use of this consisted of just transferring files and moving packets around at the level of TCP/IP (Transmission Control Protocol/Internet Protocol). The applications were FTP and SMTP, which stood for, respectively, File Transfer Protocol and Simple Mail Transfer Protocol. This was the beginning of the Internet we know today.
The Second-Wave WebNext came what I call the "adolescent Internet." In this environment, HTTP (Hypertext Transfer Protocol) and HTML (Hypertext Markup Language) grew up to allow us to take browsers and implement what I call the "Web, Phase One." People were using computers to dial up and access the Internet. Here you had people reading Web pages. We get very excited about the Internet; it's already started to change everybody's life. But at the end of the day, the Internet you know today is just about people clicking their mouse on a link and reading a Web page. Now that has evolved to the "Web, Phase Two," where, with PERL (Practical Extraction and Report Language) scripts and hand-coded mechanisms, you have the ability to create some two-way interaction through those Web pages. This is the beginning of the ability to have this connectivity.
The Second-Wave WebNext came what I call the "adolescent Internet." In this environment, HTTP (Hypertext Transfer Protocol) and HTML (Hypertext Markup Language) grew up to allow us to take browsers and implement what I call the "Web, Phase One." People were using computers to dial up and access the Internet. Here you had people reading Web pages. We get very excited about the Internet; it's already started to change everybody's life. But at the end of the day, the Internet you know today is just about people clicking their mouse on a link and reading a Web page. Now that has evolved to the "Web, Phase Two," where, with PERL (Practical Extraction and Report Language) scripts and hand-coded mechanisms, you have the ability to create some two-way interaction through those Web pages. This is the beginning of the ability to have this connectivity.
Subscribe to:
Posts (Atom)