22 January 1999
Source:
http://www.intel.com/pressroom/archive/speeches/pg012099.htm
MODERATOR: Our next speaker will be from Intel. And it's Pat Gelsinger, the group vice president of the platforms group for Intel. Pat and I have something in common, that we both started at Intel in 1979. Pat's career at Intel has really accelerated from being named the youngest vice president at Intel at the age of 32. Pat runs a wide-ranging group of all the platform desktop products within Intel.
As we learned earlier this week with the DES challenge, connected PCs provide us in the industry with a lot of power. And Intel is leading the march, creating an industry of a billion connected PCs. Pat today will talk about the connected PC and what Intel's vision and what Intel is doing to I think create the solution to the missing link, creating the trusted PC. With that I'd like to introduce Pat Gelsinger from Intel, group vice president of the business products group.
(Applause.)
PAT GELSINGER: Thank you, Al. It's a pleasure to be here at such a large and prestigious audience. As Al mentioned, I'd like to describe to you today the vision that we have for this idea of a billion trusted computers -- I think it's a very powerful vision for us and for the industry -- then describe to you Intel's commitment against delivering to that vision. And I guess as most things at Intel, we view things in a binary way. And security is now a "1". And, finally, is to invite you, the industry, to work with us in accomplishing this large and bold vision.
If any of you have heard an Intel executive speech anytime in the last several years, you probably somewhere in the course of that heard about this idea of a billion -- a world of a billion connected computers. The idea being very simply that we want to populate the world with lots of personal computers, every lateral surface of desktop, of business, every room of every home in the planet, you know, just get PCs everywhere, but then connect them, put them together, get them communicating to each other.
But what we've realized is that there's a significant gap in that vision. And what we're describing today is that we have to add a very important word to that vision, the idea of a trusted connected computer. And until we add this word of "trust" to it, there's no way that we can, in essence, accomplish that vision.
Let's take a look at the current environment that we're in. First, the good. As you can see by these variety of headlines, there's a lot going on. This industry, this Internet, this, you know, large installed base of computers, things are happening at a very rapid pace.
Obviously, the world was shocked by the amounts of Internet commerce that occurred over the fourth quarter, the Christmas selling season last year. We saw, you know, most people just simply astounded by how much it moved there.
Any of you who were holders in any of the Internet stocks have been rewarded richly by your investments there as a result of everyone realizing how rapidly things are moving to the Internet.
Also we're seeing a major change in their business practices. I include up here the Intel announcement that we made at our financial analyst meeting in November of last year, where we described that in our first quarter of doing e-commerce online, that we had blown away our objectives and had achieved $2 billion of e-commerce in the first quarter. And now we're operating at over a billion dollars of e-commerce per month between Intel and its customers.
So we're seeing a very rapid transition, a lot of stuff going on, businesses changing, new customers and new companies emerging in this space. So that's the good. However, we also have the bad. We have a set of fits, starts, and fears that are emerging concurrently with those good items.
We're seeing increasing amounts of threats and intrusions to corporations and businesses. The one quote here with the rise in attacks coming at businesses in meeting with many corporate IT managers, we see this incredible amount of concern that's rising on their part for the security solutions that are being deployed and if they're strong enough. And as they bring their intranet environment to their businesses, are they truly secure.
We also see amounts of concerns in the application models and business models. In the press last week, for instance, was the fraud concerns of eBay. Also issues around complexity and around ease of use, such as the glitches in the smart card startup, where it was just difficult to use, hard to understand. It wasn't a seamless, end-to-end solution.
So we have the good, we have the bad. And, finally, we have the very ugly.
All right. We're seeing that there were fundamental issues that have been seen in the industry as this online world is coming about. We've seen fundamental flaws detected in the implementation of a variety of security algorithms and solutions as they've been deployed, attacking the fundamental premise by which this new connected world is being delivered.
Not only that, but we're also seeing concerns in the worldwide connectedness and the deployment of solutions, the announcement here in red alert was out of India, where, basically, they issued an alert to the businesses inside of India not to rely on U.S.-based network encryption software because it was weak and tamperable.
We've also seen just this week, we're also excited that we saw the DES challenge. And in 22 and a half hours, all right, which is less than half the time of the challenge issued last year, it was cracked. So we're seeing this perpetual march to, you know, more -- what used to take literally weeks or years to crack, now we're accomplishing in less than a single day.
The good, the bad, and the ugly is the environment that we see ourselves living in. Out of this, though, we have a very powerful vision that we believe strongly at Intel and we think is appropriate for us as an industry, this idea of a trusted virtual world.
We see three major application categories that emerge from this trusted virtual world. First an idea of a secure virtual enterprise where I, as a large company, could trust the Internet with my business. I could have employees working on the intranet, I could have them spread across the world for remote work environments and remote offices over the Internet. I'm also able to work with my customers and suppliers in a worldwide manner over the extranet and integrate my business processes with them. So the idea of a virtual enterprise.
Second major category of application is this idea of trusted transactions, where I'm able to conduct both business-to-business electronic commerce, but also business-to-consumer electronic commerce. And I can have the sufficient level of confidence such as I have today in a face-to-face relationship in a paper environment, that I would have that in a virtual transaction environment as well, where I can buy, trade, or sell anything on the Net.
The third example is ubiquitous digital content. And when I say "digital content," I mean an embracing view of not only data, software, and true high-value content, such as audio and movies and the like, where I can receive high-value digital content anywhere, anytime, and deliver it to anyone across any infrastructure.
And maybe the most aggressive example of this would be a Disney* movie, or a high-value piece of content with a very long life associated and I would enable that from anywhere on the Internet to be made available.
What I'd like to do is to give one view of how this world might look. And by way of a brief demonstration, we'll actually take a look at what this trusted virtual world might look like. So what I'd like to do is come over here to my trusty next-generation trusted connected PCs and I want to be able to look into it. All right. And for those of you Trekkers in the audience, you might have a new user interface that looks like this. And rather than trying to remember passwords or losing smart cards, you know, it's hard to forget my thumb. So we'll log in using a thumbprint. Oh. It says not -- it says "rejected." Let's try this again.
COMPUTER VOICE: Access denied.
PAT GELSINGER: What do you mean, access denied? Well, the problem we have here is, this isn't my PC. This is in fact Victoria's PC. So it's -- didn't recognize my thumb print. So it's not going to let me access this computer. But let's bring Victoria's thumb up here and we'll see if Victoria can access this and help us with our demonstration.
VICTORIA QUINTANA: Hi, Pat.
PAT GELSINGER: Hi, Victoria. How are you today?
VICTORIA QUINTANA: You can access this -- so let's go ahead and start by logging in.
COMPUTER VOICE: Enter access code.
PAT GELSINGER: OK. OK. And we can see that it recognizes your thumb print. Must be different than mine.
VICTORIA QUINTANA: Hope so.
PAT GELSINGER: OK. So what do we have here, Victoria? I see a very visual, you know, next-generation user interface. What can you show us on this environment?
VICTORIA QUINTANA: What I have running right now is an application that keeps track of my personal financial portfolio.
PAT GELSINGER: OK. OK. And what we're seeing here with Victoria is that as we have the platform and her identification, she's logged into the PC. But now, actually, knowing her portfolio, it's now delivering her customized information, knowing that part of her portfolio in this case is AT&T and TCI, we're actually able to alert her of the appropriate information and then show it to her on the PC.
VICTORIA QUINTANA: Since this news affects my portfolio, we can go ahead and click on the video, see what's happening.
COMPUTER VOICE: Megamerger between AT&T and Telecom, Inc.
PAT GELSINGER: We see that we've taken this personality associated with Victoria, her profile, all right, her actual stock portfolio, and picked an appropriate news story and delivered it to you. So what else can we do now?
VICTORIA QUINTANA: Well, a lot of the new information -- I'm going to go ahead and take a look at my holdings a little bit closer. We can do this instantaneously.
PAT GELSINGER: OK. So it pulls up your full portfolio.
VICTORIA QUINTANA: And it looks like it's a good time to sell some stock. So I'm going to go ahead and do that. Do it simply by dragging and dropping from one account to the next.
PAT GELSINGER: OK. Great. So what we're actually seeing now is it's taken that identification and now, without any additional passwords or anything else, actually connecting online, authenticating that it was Victoria, and conducted this online financial transaction. OK. Half a million dollars into your account. You are one rich lady here, Victoria.
VICTORIA QUINTANA: That's right.
(Laughter.)
PAT GELSINGER: I think we need to go add some to that Intel stock account. What do you think?
VICTORIA QUINTANA: That's a good idea. I think I'll take your advice. Thanks, Pat.
PAT GELSINGER: Thank you, Victoria. So what you can see in this demonstration was, in a seamless way, you notice that Victoria never mentioned the word "security." Except for "Victoria Secrets" there at the start, it never was mentioned. And in this way, we took a seamless integration of a very rich visual user interface and we saw that her identification, the biometric as well as the platform identification, was used to allow her access to the system. But it was also used to customize information that was delivered to her. And, finally, used to authentication the actual stock transaction that occurred in this case.
So a seamless, smooth interface, right, that's delivered in line with this vision of a trusted virtual world. So against this vision, where are we today? A brief assessment of the industry that we have right now.
Businesses are getting online in a very quick way. They're changing their internal infrastructures very rapidly, and areas such as virtual private networks are taking off extremely rapidly. I mentioned before the Intel e-commerce environment. Also, Intel now has brought essentially all of our business practices online, largely eliminating paper inside of a large company such as ours. And everything is being done with online transactions, database, and e-mail on a worldwide basis.
Secondly, we're seeing a very rapid rise in e-commerce. Essentially, the "1-800" catalogue and phone business is very rapidly migrating to the Internet. And in this case, a set of low-value transactions, but still a very rapid migration, is occurring. And our challenge is that as we have rising value of transaction and number of users is to put in place solutions for the greater threats that will emerge.
And in the area of digital content, we see a different picture, one where, in spite of the lack of stronger protection, we have software companies putting their applications online and then to the industry, and at the other end of the spectrum, we largely see that digital content providers of movies and audio, etc., are very reluctant to bring that high-value content to online for lack of a sufficient security solution. So all categories are emerging, but all of them today are still very immature.
What do we need to do to accelerate this growth? Not just to maintain it, but to accelerate that growth to truly realize the potential that the connected world has to offer? First, the fundamental technology that we as an industry deliver must keep pace with the value and the threats that are emerging. We not only need to do what we have been doing over the last decade, but we have to run much faster. And I think the size of the audience today gives some indication that that's being recognized by the industry.
Secondly, we need proven interoperable solutions. We need from multiple product vendors on a worldwide basis interoperable solutions that can be seamlessly deployed to our mutual customers.
Maybe an example is SSL. It solves a very narrow problem, a very specific one, but, yet, as an interoperable solution that's available on a cross-platform, cross-environment way, it's had a very dramatic impact on the amounts of commerce that's occurred on the Internet.
Thirdly, we need consistent policies deployed on a worldwide basis. We need policies that have some level of consistency, that allow a strong degree of encryption and security to be delivered in a consistent way on a worldwide basis. We need a common recognition for those transactions to be legally binding. What's a legal transaction in one geography must be recognized by the other geography as well, because everything is connected.
And, finally, as security goes, one of the key underlying premises of security is some form of identification. And when I'm identifying myself, I'm giving some personal information of some form or another to a service application or a user. But as I give that information, I'm also raising privacy concerns. How is that information ultimately being used? So we as an industry need a consistent set of global policies that deal with privacy as well.
In 1980 or so, we recognized a computer industry that looked like this, a vertical industry. We had vertical provides to this industry. Think of this as IBM, DEC, Data General, and the like. And they provided us silicon solutions. They built those into computer systems. On top of that, they added their own operating systems. They had applications on top of it, and finally, they were the sales channel and distribution arm for them to be delivered and supported, the vertical computer industry.
Sometime in the mid-'80s, this started to change in a major way as the horizontal industry began to emerge. And we saw a specialization occurring at each level of this industry, where silicon providers such as Intel delivered a horizontal layer that was innovating and changing and evolving more rapidly in efficiency. We saw that at the platform level, at the software level, and applications and distribution as well.
And under this, this didn't change because any one company wanted it to. But there was an underlying economic dynamic, that occurred that this new horizontal industry was more efficient, growing more rapidly, and more effective at meeting customer needs from vertical to horizontal, and no company, no geography, no matter how strong, could resist the dynamics of this new horizontal industry.
Let's compare this to the security industry of today. Guess what? The picture looks the same. We largely have a vertical industry. We have hardware solutions which are uniquely bound to the operating systems and the frameworks associated with them, with the applications, services. And at the highest level, you'll think about our government policies, where they try to control export and use at the vertical solution level. We have a vertical industry.
But as the Internet, the PC, and many other things are emerging, we're in the process of seeing the first signs of this becoming a horizontal industry as well, where we see that we have common building blocks, where we have common middleware and frameworks, applications, etc. And we need to adjust that through every one of our businesses. Every one of us needs to adjust to a new horizontal business model to deliver the full capabilities of an Internetworked world.
And as we saw in the computer industry example, there were two kinds of companies: those that embraced that change, those that got in front of it; and those that tried to fight against this fundamental dynamic of the industry. But as a new economic dynamic, it cannot be reversed. And, really, successful companies only have one choice, but to get in front of that new horizontal industry model.
Now I'd like to describe what Intel's going to do to deliver, our commitment against that vision.
First is, we are -- as I said, security now a "1" at Intel. As part of our core CPUs, chipsets, motherboards, platforms, software building blocks, we will be enabling with those building blocks trusted connected PCs, we will be adding security into the fundamental building blocks that we deliver. Security is only as strong as the foundation upon which it is built. And, fundamentally, hardware is a firmer foundation. Hardware, by its very nature, is highly resistant to tampering.
Software, at best case, you can disguise secrets in software. You can try to, you know, lay them out in the bits in some very creative way, the actual software application, and you can disguise the actual operation against those secrets. But hardware, in comparison, is very good. You can put it at places in the hardware where it can't be seen and can't easily be monitored in its execution. So it's good at hiding and processing secrets.
It's also very hard to duplicate. Any one of us can start up a CD-ROM duplication shop, go down, you know, to our local computer store, hundred bucks, add it to our PC and we're in the CD-ROM duplication business and I'm able to copy those secrets and create exact replicas of them. But to replicate hardware, you need to go build a multibillion-dollar fab. Not quite as simple or easy an investment to make. And also, hardware can process secrets more rapidly than software can.
Our point here isn't that we are replacing software. No, no. Far from it. What we're doing is building a foundation for trusted software to use, and as such, delivering that -- and as such, hardware is improving software solutions for security. So first, building it into the building blocks of the PC and into our actual silicon products.
Second is leading the development of a common security architecture for the platform. The first we'll describe in this area is CDSA. And we'll have some specific things we'll discuss on that shortly. Also, finally promoting common policies for security, both at the government and at the public level, and working with the industry to get common policies in place.
Let's take a look at the road to the connected -- trusted connected PCs. In '99, we're going to begin by our first set of building blocks that we're going to be delivering. These will enable platforms and the users that are on those platforms to be better identified. We're also going to be delivering some of our first enhancements for cryptographic processing into the platform. Our first building blocks in the area of security.
In 2000, we're going to add to these significantly, building on identification and allowing now trusted access, adding authenticated permissioning to PCs, increasing levels of capability in the underlying building blocks, and, ultimately, to, in a large way, deliver this platform that businesses can manage and communicate in an Internet, intranet, and extranet environment for the virtual enterprise.
In 2001, yet again, adding more building blocks, you know, those building blocks in the platform, also addressing areas such as platform and peripheral integrity. And in 2001 is when we see that we're accomplishing the trusted transactions as well as a platform strong enough to bring all forms of valuable content to the PC.
So with this road map in mind, let's take a more detailed look at 1999 and what we'll be doing. Let's start by looking at public policy. First, government policy.
Intel is a large global supplier. We have some $30 billion, some 65,000 employees on a worldwide basis. And, fundamentally, to keep that business running successfully, we need to do four things. We need to be able to ship our products on a global basis. We need to be able to run our business on a global basis, our communications, e-mail, all of those kind of things. We need to be able to conduct e-commerce on a global basis. We have some 50,000 companies that we do business with on a daily basis. And we want to be able to conduct e-commerce with them worldwide.
And, finally, we need to be able to manufacture our products on a worldwide basis. We have some 20 different locations where we have manufacturing facilities. So those are our four needs. And we need a government policy.
(Lost audio.)
PAT GELSINGER: An example of this was what was just announced two days ago by the French government, where Prime Minister [Lionel] Jospin announced a major -- deregulated up to 128-bit encryption inside of France, a dramatic reversal from the policy that they had before, as most of you are aware of. Furthermore, and this is what we really mean --
(Applause.)
PAT GELSINGER: Yeah. But maybe down further, that announcement that came out is maybe even a more significant statement. Because what he said was they were also increasing the level of funding, right, of the underlying technology in France so that they could provide to governments and law enforcement agencies in France tools that were staying ahead of the bad guys. So fight technology with technology. You can't control technology. The best you can do is invest in getting ahead of technology.
Second thing is, no secret back doors. Intel is out to deliver trusted building blocks that are used by everyone in this audience, everyone in the world for our PC platforms, trusted building blocks for billions of PCs that will be scrutinized and analyzed by millions of users.
All right. A policy of back doors is in stark contrast to that. The only way that we can build trust into our platforms and have the industry build upon it is with a policy of no secret back doors. And even if we attempted to, the industry would see them. The only way for us to successfully accomplish this horizontalization is a policy of no secret back doors in our products.
And, finally, we need -- we, as an industry -- need to lobby the governments effectively to improve and change their policies on a worldwide basis. As a U.S. company, we're doing that through the Americans for Computer Privacy. We encourage every one of you and the companies that you represent to aggressively participate and join with the ACP in our efforts to relax and improve and accomplish this vision for government policy in the United States.
As we have said, as security is built on identification, you're giving some of your personal information in some way to the application or services that you're accessing. As such, we need to be a role model for responsible use of user information. What do we mean by that? Well, this is Intel's policy that we're going to first, you know, notify and give user control of the information that we gather from them, so they'll be aware of the information we gather and they'll be given the opportunity to control how it's used.
But as a building block supplier, we have to go a step further. We have to build the core capabilities into the products that give those users choice as well. And we're committing to build our products in such a way, but, again, as a building block supplier, we have to go a step further again. We have to evangelize the customers of those products to properly use them and deploy them. And we have active programs under way such that we can evangelize and deliver solutions through our ultimate customers that are in line with those privacy policies as well.
And, finally, we want to be able to communicate that. We want to increase end user trust and confidence in this, and as such, we're engaged with industry certification bodies to help communicate that in an end user-visible way. And we [are working toward] an agreement with TRUSTe to assist in accomplishing that objective. So that is what we're doing at the policy level.
Let's talk about the application framework or the middleware layer. Again, as a building block supplier and adding security capabilities to our core CPUs, chipset, platforms, we've taken an approach where we want to enable every infrastructure onto that platform. As I jokingly say, we love every piece of software as long as it's running on an Intel platform. And our approach here is to do exactly that, to work with CAPI, CDSA, RSA, and the other major suppliers and get them to use those hardware features to increase the level of security by taking advantage of those hardware capabilities. And then develop and promote a cross-platform solution through CDSA.
Let's look at each one of these in a little bit more detail.
I was very excited about the announcement on Monday where, with RSA, we announced the relationship that would result in the RSA frameworks and tool kits, be safe, taking advantage in supporting the Intel security road map starting this year and beyond.
So as part of the tool kit, there will be releases coming from RSA that enable and take advantage of these hardware primitives that we're putting in place. We're also working together to accelerate the industry's deployment and use of those capabilities, deploying that to the ultimate applications that use those tool kits, and as such, to enable them to a higher level of security.
We also, through the relationship, announced a fairly broad set of cross-licenses and technologies that are being licensed between the two companies. A very significant announcement for both RSA and for Intel.
In the area of CDSA, we announced our activities in CDSA in 1996 at the RSA Conference. Right here three years ago, we announced the formation of the common data security architecture. And as you can see by this slide, the many significant names on it, we've made tremendous progress toward accomplishing the vision of this -- of this technology.
We are very happy to have The Open Group adopt CDSA as the standard by which it's going to drive and deliver to the industry. We're also very happy that, yesterday, we were able to announce a relationship with Baltimore. And Fran and our friends at Baltimore in the audience, you can wave at us here. Where are you guys? You're here somewhere.
We're very excited with the relationship that we are able to announce with Baltimore/Zergo as the first international provider of CDSA. And they'll be building us into their products and also providing tool kits on an international basis in line with the CDSA architecture.
We're also happy to announce that Compaq is the most recent adopter of CDSA. Compaq plus their units of DEC and Atalla. And this announcement and the actual licensing of this will be formally announced shortly.
Late last year, we also announced the relationship with IBM, where IBM was bringing multiple operating systems, their major middleware solutions, such as Lotus* and Tivoli*, to the CDSA architecture.
Also, Intel and IBM announced a set of developer programs by which we would be mutually working with the industry to bring applications to the CDSA environment. The first of these developer days will occur in Washington on March 30, and in San Francisco, April 1, will be the first developer forums that we'll be jointly hosting for CDSA. And we're very happy that IBM is very committed to the CDSA architecture, a key contributor to The Open Group, and the first company to ship products using CDSA.
So what I'd like to do now is describe in a little bit more detail exactly what we're going to deliver as part of the platform in 1999. In '99, we'll be adding the processor serial number as part of the Pentium® III family of processors. Additionally, we'll be adding to the chipsets and the core silicon a random number generator directly into the core silicon of the platform.
At the Intel Developer Forum in November, we announced the Wired for Management 2.0 specification. One of the key pieces of that was the ability to have an integrity services at bootup as part of that. And that's what the BIS architecture is. And that will be delivered to developers in the first half and be shipping with products in the second half of this year.
And, finally, a commitment to the IPSec technology. And you can see this through the acquisition or the pending acquisition of Shiva by our networking group. So be bringing these four capabilities into the core platform in 1999.
Let's talk about the processor serial number in a bit more detail. So the processor serial number will be added into the CPUs at the time of manufacturing and every one uniquely burned in as we go through our normal manufacturing process.
You can think of this processor serial number like the equipment serial number of your cell phone, a, you know, identifier that's part of that piece of equipment that's then being used by the services that run upon it.
We're very happy and actually rather surprised by the amount of enthusiasm we've gotten from application developers for the processor serial number capability. We have some 30-plus applications today that have committed to take advantage of this. And that number is rising very rapidly.
We see both consumer and business applications for this. And some of the three categories of applications that we've identified, one is applications using security capabilities. You think about this maybe as a chat room, where unless you're able to deliver the processor serial number, you're not able to enter that protected chat room and providing a level of access control.
Manageability, the major enterprise management suites will be taking advantage of this to allow management control. Also, it can be used as a stamp for information manager, maybe such as an enterprise data backup solution, where you know exactly where -- what platform this information came from as it's being backed up.
As we mentioned earlier, we have a policy of end user privacy that we will be supporting. And we've also built into this capability the facilities for the end user, the control to be able to enable and disable the processor serial number. This capability, again, shipping as part of every Pentium III processor family. And the shipments of those will begin later in the first quarter of this year.
Next capability in the platform is the hardware random number generator. The -- My favorite quote here is -- you know, I really like this one -- is -- and this comes from a gentleman that some of you might have heard from, John von Neumann -- and his quote is: "Anyone contemplating generating random numbers using software is in a state of sin."
So, since we don't want any of you to remain in a state of sin, right, we're evangelically calling you to repent and providing you means to do so.
So as part of the platform later in the year, we'll be adding a hardware random number generator. This random number generator will use the thermal noise of a semiconductor resistor to generate high-quality random and nondeterministic numbers. We see this as a complement to software, where software can use this as a seed for their random number generators or for the random numbers themselves. And as a result, it provides a stronger underlying foundation for key security applications and software such as cryptography, digital signing, and security protocols.
In conclusion, we see our opportunity is to recognize this horizontalization that's going to occur in the industry, an economic dynamic that none of us can resist, none of us can stop. What we can do is recognize it and embrace it.
We can do that through our products, by delivering trusted platforms and applications in a horizontal industry model. We can build upon common architectural frameworks. Every layer of that horizontal industry needs horizontal -- needs standards and a common architecture to connect it with the layer above. So we need common architectural frameworks at every layer.
And, finally, it's to build on a stronger security foundation, take advantage of the capabilities that we'll be building into the underlying security building blocks. And as we've described, we think, as an industry, we have a significant challenge in influencing both government policy as well as privacy policy, and we ask you to join with us in influencing those policies in a way that allows our industry to continue to grow and thrive.
In summary, please join us in making the trusted virtual world a reality for all of us. Thank you very much.
(Applause.)
*Third party marks and brands are the property of their respective owners.