全球頂級CEO的演講NOVELL(nice)

COMDEX Keynote: Novell CEO Eric Schmidt November 19, 1997
The New Face of Networking Transcript


When the folks from Softbank asked how long it would take to put together a network of 250,000 users, five million messages, we replied, "A few days, maybe a week." And we did it, with the help of Compaq. I'm very, very proud of that. We made a mistake, though; it's, in fact, 1.85 million messages a day because of what you guys are doing, which I think is incredible.
What I want to do in this hour is take you through all the things that you can do with networking, the promise of networking, and some of the exciting things that are going on in the industry.
Today, Novell has a much more sharpened focus. We're trying to balance new technologies with our existing customer commitments. We're managing the complexity of applications -- all the kinds of problems that everybody here deals with, and we're planning with our customers their long-term information technology strategy.
So we have a renewed focus at Novell to deliver the kind of technologies that people need to run their networks. That's what we're all about.
My talk is actually about the new face of the network. Now, the new face of the network is not an interface, and it's not even a single face. The new face of networking has a human face, and it has as many faces as there are users on the network today.
When we take a look at everything we've done in the last 15 years, we've been governed by Moore's Law, and it's been a good law for all of us. It's the thing that created COMDEX and made this as fun as it has been. And we're now moving to a situation where we're governing much more by the speed of light. But I think that's not the ultimate constraint. I think it's the speed of people's minds. I think it's the ability of organizations to deal with the complexity and creativity that networks bring to them; that solving that problem turns out to be the most important thing we can do; that we need adaptive systems that can work with this set of problems.
The industry that we're in is young. It moves fast. We deal with time compression and globalization; we're very proud of what we have accomplished. But what we're seeing now is a declining significance of Moore's Law, and an increasing significance of the problems that I'm going to highlight. They ultimately become the barrier to progress. In other words, the scarcityhere is of time: Time of the person, time of the customer, time to change these new models; that's the opportunity that's in front of us.
Now, how big is this? Well, there's a lot of interesting statistics. The size of the web: there are roughly 71.3 million web users, according to IDC right now, and 130 million are expected by the year 2000. The compound growth rate of these things yield tremendous numbers. We've often said that at the current growth rate, every human on the planet would have their own web site by the year 2004. But if you look at the number of trade shows we have to go to, every human will have their own trade show by the year 2004. It just goes on and on and on.



And there's an underlying transition here from static to dynamic, and that shift is something that underpins almost everything we do. It's not obvious until you begin to think about its implications, because the Internet and networking technology are bringing forth a whole new set of mediums.

There's a number of examples: creative pattern matching, or what computers can do that we really couldn't do before. It can watch what you do. It can figure out what you want. And it can suggest the next thing.

I don't know if you've noticed, but Amazon.com -- one of the services that I use on the net -- now remembers your books and then makes a suggestion, based on a perception algorithm, for books that you want to buy. When was the last time your bookstore did that for you?

There are lots of examples of this. American Airlines has a program that, for a certain category of customers who buy through electronic means, can give you instantaneous price updates so that you can absolutely know the marginal price of that middle seat that you didn't want anyway.

We see computers being used not just as substitutes for the physical world that we're used to, but in a very different way. And this whole thing is governed by a set of laws around network effects.

If you look back, we can see network effects that were a surprise to us 20 years ago. Most of us spend much of our time traveling around and dealing with the United States airline industry, and their hub cities. When the airline industry was deregulated, what was the competitiveness equation? Of course, it was the domination of hubs. That's the network effect. The more people you can get into one hub, the more you can make sure that they stay in your network. The same principle applies to what we do.

Technology adoption is driven by laws of increasing returns. When the first fax machine came out, it wasn't very useful. But when lots of people had them, all of a sudden, everyone else had to have them. When the first cellular phone came out, the same pattern unfolded.

These mass waves, driven by brands and people, are often confused by how absolutely crazy these markets are. One year you're a failure, the next year you're a billionaire. Back and forth and back and forth. It happens all the time. You announce a new product, and, in a week, you have a million customers, and, boom, you go public. You have to get the revenue later; ubiquity first, revenue later. That's what URL stands for.

My essential thesis here is that there are two intertwining issues. The human face of networking is interactive and changing; it's that interactivity. The network knows you, the intelligent network model. And that the adoption of technology is driven by increasing returns. The two together make our world extraordinarily dynamic and fascinating, and very, very competitive.

If you follow the new face of networking idea, then every face is different, and everyone's needs for networks are different. Every person on the network has a different perspective. So the world of needs is really about millions of new faces coming onto the Internet. Networks are empowering.

When people think about computers, they really think about themselves, and they think about what they can achieve. This technology and the industry that it is making is an incredibly personal one. Ultimately it's driven by the desire for intrinsic recognition that we all have. And people want this stuff now. We're busy building this world and we're building in cyberspace replicas of our physical space. So, for example, it shouldn't surprise you that there are saints and criminals on the net. There are even banks to put your money. You have to have churches where you can pray for support








on the net. It makes sense that we're building all these new models.

In the middle of this, we have our chief information officer, who has enormous challenges. And in this new role that I have, I've spent a lot of time talking to them. The problem that they have is that they more or less had it wired like 10 years ago when they understood what everybody was doing.

We have a notion of a software distribution company to a distributor, that's the truck; to a CIO, that's the gentleman; to the user, that's the lady. What happens now is she goes directly, using the net, to the software manufacturer. You say, "No big deal; I do that all the time." The problem is, at least from a corporate perspective, serious, because of the support, asset management and legal use issues, the most important one being security. So you need a solution that works politically within the organization, that the lady, who is the end user, has to feel empowered, but the CIO has to be able to set up a framework.

So the correct way to think about this is that the services will evolve so that this end user will have permission granted through a set of directory-enabled solutions that allow her to get access to the things that she cares about. The CIO won't stand in the way; he'll make that happen. The CIO must embrace a view of the network that enables that.

Now, the CIO says, "What I really want is a network where I have a command central, a place where I have asset management, software distribution, bandwidth managing web access, security policies, those kinds of things."

The end user has a very different view. They want it all now; it's the world of me, they want to make it happen.

The key strategies in all of these faces are to use directory-enabled services to make those solutions available, and then give the end users and the CIOs the power to make those choices.

There's an enormous data explosion going on at the same time. The World Wide Web, HTTP, Java, etc., have made it so easy to publish any kind of data and deliver it in application web traffic, it's increasing at roughly 30 percent every three months. The World Wide Web traffic is approximately 31 terabytes per month. And the current expectation is it will grow by at least a factor of 100, which I looked up, it's 3 tetabytes, which is a really big number, per month.

You should expect that because of the latency inside of organizations, you'll see really innovative things, first on the web, and then they will move into corporations because of end user demand and so forth. So the web is this tremendous cauldron of innovation, which has awakened us from this sort of networking conundrum that we've had for so long.

It turns out that way because, with the interaction in the world of needs, there's an issue around being close to things. When you build these data networks, the most important issue is to have perceived proximity. You absolutely must have the data close to you. The technical solution for that is called replication. What you do is replicate caches and you make all that stuff work. By doing so, the network recognizes a person's identity and knows who you are.

The architecture that underlies these networks is built on a set of caches, a cache in a sense of a place where you keep local copy, and then complex algorithms that keep the caches up to date, because we all know that the computer performance and the demand from end users is outstripping the ability for enterprises to deploy bandwidth that they need. The combination of

the two makes a great deal of sense.

As an aside, we have a BorderManager, which handles approximately 50,000 concurrent connections, and 5,000 pages per second, to do caching and proxy analysis. The equivalent NT and UNIX versions are 1/10th of those. That's an example of a sweet spot that you can find when you look around on these networks at all of these different devices.

So, we find ourselves living in a world that's very different from what we thought. I first thought we would have a series of tiers; first tier, second tier.... In fact, the ISPs that we so much depend on are interconnecting, which is great. And by doing so, they're, in fact, building a whole new network.

One thing that's happening is that the TCP/IP technology, that's becoming the basis for the world's data networks, is also suited for voice. And although today the percentage of the voice and data are roughly equal to first approximation, the expectation, in some analyses, is that in the next seven years the percentage of data versus voice will go down to about 1 percent voice. That gives you an idea of how significant this change is.

Why does this matter to us? Because this is a deregulated world. It's an example of an inversion. Just when you thought you understood exactly how everything would be priced in a P times 64 pricing architecture, all of a sudden we have an opportunity to change it again. I don't know how that will change it, but I know that it will affect a $650 billion industry.

Reed Hundt, a friend, and former FCC Chairman, said, -- and you'll love this -- "What we need is a high-speech, congestion-free, always reliable, friction-free, packet-switched, big bandwidth, data friendly network that is universally available, competitively priced and capable of driving our economy to new heights." Makes sense to me.

And, by the way, we're building it. We, together.

Put another way, what we need is a data network that can easily carry voice instead of what we have today, which is a voice network struggling to carry data.

By the way, yesterday AT&T announced WorldNet, a virtual private network service, with a business quality IP offering and end-to-end service guarantee. It enables remote access to LANs, extranets, intranets, that sort of thing. It brings down a huge complexity barrier for all of us as customers, and it uses a protocol called "radius," R-A-D-I-U-S, to authenticate into a directory, which happens to be Novell's NDS. It's happening, and it's happening very quickly.

There is a tornado happening. And this tornado is this change from this adoption of IP. I call this an intelligent network. I talked about it. This is as a replacement. In that model the subnet that you've built for so many years will die and be replaced by these interconnected TCPIP networks.

For us at Novell, Moab, the next version of NetWare, is the key part of our strategy, because that version of NetWare uses IP completely natively, and all the services that you as customers should know and love, and that have been best of class, will now be available on these open, interconnected networks, with extraordinary performance -- much higher performance than any of the other competitors in those spaces. That's the commitment that we've made. And it's in beta today.

Just saying TCP/IP, which appears to be a mantra in our industry, is not sufficient. It turns out that what's important is that you have the base case to TCP/IP and then all these services, which run on top. The enabled services turn out to be the differentiator. In a situation of open interoperability, where everybody has the same stuff, how do you differentiate? In our case, it would be to make this tornado be worth something to your business. Don't be afraid by the





destruction that the tornado means. Ultimately, it's a much better model.

The other issue that people are struggling with is a set of border or firewall issues. As people are building these networks, there are huge issues around security and access and those kinds of things. The original model that we pushed three or four years ago was that there was going to be this firewall, and there was the intranet, where the good guys are; and the Internet, where the bad guys are; and then there's the firewall, which is the Dr. No. It says no to everything.

The problem with that is it was an extraordinarily inflexible architecture; it was not the correct technical approach. What you, in fact, need is about 45 shades of gray. Because our customers -- and I suspect most of you -- are trying very hard to interconnect your customers, your suppliers, with different levels of service. Some are friendly. You're going to deal differently with your customers and your partners than with your competitors, but you may even collaborate with your competitors some. You need this different approach.

The point here is that security should be a business decision for you.

IDG, after we announced BorderManager, said, "With initiatives like BorderManager, Novell is, for the first time, truly repositioning its core competencies strongly within the Internet space."

You have an opportunity here to drive a very different and powerful IP model. The Internet service providers have all the border, all the caching, all the security and all the naming issues that I'm describing. So a combination of solving the ISP, caching, access and security problems turns out to be a huge sweet spot for all of us to drive literally right through -- mixing metaphors; you get the idea.

And we've heard a lot about thin clients, and they're a lot of fun. There's this thin client, and that thin client, and then there's -- well, there's not so thin clients. And then there's the fat client. I bet you all didn't know you had a lot of fat clients when you bought them.

If you take a look, it's actually happening. People expect over the next few years to build a lot of non-PC devices. It's very clear to me that one size no longer fits all. But I don't think it's going to be either/or; it's clearly going to be a mix. You want the promise of a network computer, for a lot of the reasons that we've talked about, but you also want to use your existing investments.

So perhaps a better approach is to NC your PC. In order to do that, there's a set of services that you need; in particular, you need a directory service that can keep information and hold the stats somewhere else than in your fat client. By moving the complexity from the PC client onto the network, you can pull that off. From our perspective, NetWare is the perfect model to do this. It has the dedicated performance and the scalability to handle many, many different devices and the right platform for directory-level kinds of things.

What's happening in this whole initiative is people are beginning to develop applications, and this, to me, is incredibly important. The whole Java phenomenon -- I'll talk about that in a few minutes -- is creating a new sort of network architecture. These are applications which are distributed by design, a little different from what we use today, multi-user by need, geographically dispersed by necessity. They derive value from using a thin client, and they often include work flow.





If you look at the players in the industry pushing this, Oracle, for example, has something called NCA, which is really data in the network. Sun calls this Java Computing. Netscape calls this Netscape ONE, really web computing. IBM calls this the "Open Blueprints," also network computing. Novell calls this an open solutions architecture. Microsoft has named this the Gang of Five, and feels it's a conspiracy. I'd prefer to think of it as being right.

This initiative around building these new categories of applications is one that all of us can play in, and it really does match the model that our customers -- and everyone's customers -- are really striving for, to move to this 1990s, Year 2000 kind of model.
If you look, for example, at PC ownership -- and you've heard these PCO arguments for years, and they never seem to get solved -- we've been analyzing this, and guess what? It turns out that the majority of the cost of PC ownership is the management of them. And you might say, "Well just don't manage them." That isn't a rational option for most corporations, although for folks in this room who are very sophisticated, perhaps.

What makes up your identity is your relationship to things. This way, you have a place in the hierarchy. You're identified by your company, your geography, your department, and yourself as an individual. And you can interact with things in the network world differently, depending on which department -- and I know you say to this, "Well, isn't this obvious?" It's obvious to us as humans, but not to our computers. Our computers are authenticated to the machine, not to you. So the solutions that we build, and the solutions that Novell is building, are based on this notion of the digital persona, that you get inalienable rights when you get yourself installed into the network.

From the technology perspective, you get access to services, you get through routers, you have modem banks, all those other kinds of things. The important thing is that you can authenticate to the person using the technologies that I'm describing. And you can end up in a situation where the network knows you.

So start thinking about, for example, when a new person joins your organization, what's their life like. What I thought I would do is take you through that a little bit and describe what happens when the network knows you. I want to demonstrate how this world might work. I'd like to do is ask three people who work at Novell -- J.D. Marymee, Kent Prows and Gary Hein to join me on stage.

MR. PROWS: The demonstration that we've put together is deceptively simple and yet extremely powerful. We're going to take two perspectives of this technology: one from the administrator, played by J.D.; the second played by Gary. The idea we wanted to portray is of a digital persona, and that with a single user we can expose this digital persona.

So we took a little bit different tack here, and we decided to create a new CEO inside of Novell.

MR. PROWS: J.D., you know, in this world, exposing this digital persona is kind of like, you know, you've got all of these different technologies: You've got e-mail, you've got database actions, you've got file imprints, you've got Groupware, you've got application management, you've got software distribution, you've got all of these things. And from the administrator's



perspective, J.D., what's it like to administer all this stuff?
MR. MARYMEE: Well, as an administrator, I'm responsible for thousands of workstations, hundreds of different sites and constantly trying to integrate new technologies into our existing systems with our existing infrastructure. Making all these systems work together as a solution is extremely difficult usually, but adding a new user to our company used to require coordination between half a dozen network administrators, each of whom was responsible for a piece of the complete enterprise solution.

I've got all these issues now distilled down to a single user create. So we're going to create Eric now; expose his digital persona. This is a flashback to when Eric got his desk. What's your last name, Dr. Schmidt? And I'm going to have one template that's going to create everything. And in addition to that we're also going to allow him to get access to e-mail over the web. I'm going to put a password in for that, so it's at least somewhat secure over the Internet.

MR. SCHMIDT: While this is happening, you're setting up this digital persona in the entire enterprise network?

MR. MERRAMI: Yes, I am. As a matter of fact, there I go. Now I've actually got him set up to do things like access to his workstation, access to e-mail, access to the Internet -- or limited access to the Internet, depending on how I want to govern that -- access to the Netscape enterprise web server, even access to file and print services, both on NT and on Netware.

MR. PROWS: Well, let's come over to our new CEO, played by Gary.

Gary, how does this affect you?

MR. HEIN: I'm a CEO, not really a computer geek -- well, maybe a little bit. So, I really don't care about operating systems and protocols; I just need to have applications on a desktop to do my job. I want to come to Novell my first day at work, log in one time and have access to all the resources I need to do my job, regardless of what platforms those happen to be hosted on.

MR. PROWS: Well, J.D. created your account over there, Eric. Why don't you log into here. You've got an NT workstation that's got the network client, and that's it. So why don't you log in and let's see what happens.

MR. HEIN: It looks like I got into my NT workstation, got to the desktop, and it's pausing a little bit; but it should probably start to pull down some information here. What's happening here?

MR. PROWS: J.D., from your perspective, what did you do to Eric?

MR. MARYMEE: NT Workstation is a secure desktop, which means the user has got to have a password before accessing a workstation. So I'm providing a single sign-on to the NT workstation, creating a user in NDS, actually granted the user rights to his desktop, in this case Eric, saving me, the administrator, the work of manually creating the workstation account on that particular machine. The benefit to Eric is he didn't need a second password to get into that system.

And before Eric can start to do his job, though, he needs applications on his desktop. The directory knows who Eric is, and we're using this information to install and configure software that he needs to do his job.

MR. PROWS: Wow, it looks like everything came down. You know, your bitmap, the new corporate bitmap, and I know that you're the CEO, but, even at that, I doubt you can even change this bitmap.

MR. HEIN: Yeah, I know NT; it's easy for me to change this. I just go under properties, click on properties, and I can change -- well, nothing, really.








MR. PROWS: So I guess we know who really controls all this stuff.

MR. HEIN: What happened. I don't understand.

MR. MARYMEE: Because you're known in the directory, we can actually configure applications on your desktop and everything with information from your digital persona that is stored in the directory.

MR. PROWS: It looks, Eric, like your communicator was downloaded there. Why don't you start that up, and let's see what kind of work you can do?

MR. HEIN: Before I do that, we can do a lot in GroupWise and so forth, but I really need some more applications down on my desktop. We just purchased a lot of licenses for some productivity apps, so I'm going to ask my administrator, J.D., to set up some more applications on my desktop for me.

MR. MARYMEE: We're going to add a couple applications here to Gary; not just to Gary though; anybody that happens to be in that part of the tree will also get access to these applications. So put in Access, Powerpoint, Word, those kinds of things.

And just with a couple of clicks, I've actually delivered availability for those applications down to Eric's platform.

MR. HEIN: Before I can do any work, I need to customize my computer and my applications so it understands my identity, my persona. So what I'd like to do is configure my preferences in this Netscape Communicator with my information, my identity. Well, it looks like it's already there: My e-mail address is in there, and my name. How did that happen? This application isn't necessarily NDS aware.

MR. MARYMEE: Remember that as we bring this information down, we go and read your digital persona, and we can customize any application as it's distributed to the desktop to match your exact needs.

MR. HEIN: Let's go out and browse the web. We have a little issue, being from Utah, and our engineers are relatively frustrated and spend a lot of time out on the web. And I've asked my administrator over here to make sure that this doesn't happen any longer. So I'm going to go out to see if he's actually doing his job here, and go out to Playboy, you know, just for the articles. And, hey wait a minute, it looks like I couldn't get in for some reason, although I used to be able to get in -- I mean, users used to be able to get in here.

MR. PROWS: J.D., why don't you explain to the new CEO why this happened.

MR. MARYMEE: BorderManager makes Internet access, actually that kind of management simple. I just create a couple of rules that control access to content on the web, such as non-productive sites like the one that Eric just hit. Once again, the directory manages who these rules affect.

So, Gary, instead of sitting around browsing the web, let me give you a URL from the new on-line realtime sales database. And, of course, since now you're a user in your e-mail system as well, I can e-mail this to you. And we'll just say "new URL; sales.novell.com."

MR. HEIN: I spend a lot of time on the road, and, I'd like to access my e-mail anywhere. I don't always have a connection to the phone system, believe it or not, but I always will have a connection to the Internet. So, I'm going to actually log in and access my e-mail right out of my web browser.

MR. MARYMEE: The URL that I sent Gary points to an application running on a Netscape interface server running on Novell's NetWare, produced by Novonyx. This application is actually running server-side Java script that natively connects to our Oracle sales database.







MR. HEIN: So we can go ahead and go to the URL here. This looks like the front-end to our new Oracle database. I'd like to go ahead and check sales for a couple of the different regions. And it looks like North America is a little bit behind their quota. So I want to check on which of our sales reps is in the North America region, and how close they are to their quota.

MR. HEIN: What's basically happened at this point is I've used a standard web browser, talked to my Netscape sweet spot server that's running a Java server side scripting that talks natively to the Oracle database, posted on the intraNetWare server, and, I can go through and browse the sales figures from any web browser anywhere in the world.

This is wonderful. Everything I needed to do my job happened the very first day.

MR. PROWS: Well, J.D., I also noticed that one of the requirements that we've put up there was access to that one operating system, oh, NT. And I was just wondering, you had access there; how did you do that?

MR. MARYMEE: Let's go in and add them into another domain. I can manage my domain information now using NDS for NT. In NDS I'm going to go in and take Eric's account; with that single create we actually gave him access to a domain already and then add him into another domain; it's just a matter of choosing a domain that I'm interested in. In this case we'll go into the Seattle domain. That's how simple it is to manage user access to NT domains now using NDS for NT.

MR. PROWS: To summarize, the challenges that we face with daily are both from the users', as well as the administrators', perspectives. They need access to applications. They need to be able to manage the desktop. We have groupware needs, and e-mail. We need access to the Internet. We need to be able to control access to the Internet as well. We need database access to do our jobs and also file and print.

We've demonstrated that when a user is created, this digital persona is also created. And based on the information you put into the directory. The directory then manages that relationship and provides you with the ability to get access to the information that you need.

DR. SCHMIDT: The important message about that demo is that those products are shipping, except for the Netscape web server from our subsidiary called Novonyx; that ships in beta Thanksgiving week and is downloadable from the net at www.novonyx.com. Quite a departure from what many of you think Novell is doing.

Another important point is there's a much broader story about what we can do. The current sorry state of our networks would be represented by a human face is wrapped up in wires. It's enormously complicated. The average number of directories that a person in a Fortune 500 company is in, in the Fortune 500 company, is about 120.

We at Novell took our eye off the ball, but now we have a solution for all of this. We have distributed network services in the mid tier -- that's what the Moab and NetWare stuff is all about; a platform for management of these services -- that's the directory, our ManageWise product, etc.; and a new and renewed attractiveness to the developer community.

The directory simplifies because of one of the most fundamental principles in computer science, which is that every problem can be solved by one level of indirection. What we've done essentially is taken all of the complexity that you all deal with every day and put it in a place which is in the network. It's intelligent, it's replicated, it's cached, and it's an area where we have a significant competitive lead.







The Aberdeen Group wrote that "NDS for NT, the same product running on NT, even obviates the need for Microsoft's Active Directory," which is good, because it doesn't exist yet.

People are doing user account management, desktop and network management, collaboration and application management -- all made possible through this model.

What you saw in the demo is the principle that we call application redirection. The old-timers here know that Novell made its name because of DOS file redirection -- a unique skill within the company. By bringing out a set of libraries and services that are on the wire protocol compatible with everybody else, we can now take all the applications that you use and have all this happen without your user or systems administrator intervention. That second phase is a huge change in the way you all use networks. But by delivering this common access management security framework, you don't need to rewrite your apps; you get all sorts of benefits, shared data, less hardware; everything's more secure. The total cost of ownership solutions -- we've got white paper after white paper from the marketing folks giving you all the data -- is incredible. I think it's intuitively obvious how strong that stuff really is.

By the way, we're going to do it for NT NetWare and also for UNIX.

Java is also part of our strategy, for obvious reasons. To give you an idea of how well Java is doing, Java today is very much a huge success. There have been more than 1.1 million downloads of the JDK 1.1. IDC says there are 450,000 Java programmers, and 700,000 projected for 1998. This is for a language that is two years old. There are 800 books, showing where the real revenue is. There are 400,000 websites using Java. There are more than 1,000 applications shipping. And at the Java1 conference in March, there were approximately 10,000 attendees. This is a very, very significant phenomenon in our landscape.

To date, 87 percent of the application developments has been on the client. And that's where you have all the fun. That's where all the battles over who controls what and who has this library and the lawsuits and so forth are really being fought out.

But people are missing the real story behind this. The real opportunity for Java will turn out to be when it's implemented on the server. Middle tier -- remember, there's the middle tier where a lot of the competition goes on in networks today -- is where server-side Java will take off. We call this again the intelligence in the middle. It's very powerful. By its very nature that tier is not dominated by any single monopoly. There's lots of different choices. They're not going to go away. Customers have lots of complexity problems.

So server-side Java plays to one of Java's strongest suits, which is platform independence: faster performance, portability of applications. It has the right architecture to build these new network-aware applications. This is a core part of our strategy.

Now, it's incredibly important that the APIs in the industry not get split. And if you read sections 2.6 and 2.8 of the Contract, you can form your own opinion about what 117 licensees have all agreed to do.

Java is more than a dancing bear on your website. It's really the language of electronic commerce. And where are we in all of this? The Java language will replace C and C++, and the 100 million JVMs that are out there really are an enormous asset for universities, new language development and things like that. Key computer companies, key software companies -- companies like Novell, universities, etc., are going to build new dedicated platforms that use









Java to provide new value. And those dedicated platforms will eventually -- and eventually needs to be underscored -- become major platform players in our industry. It will take a while because of applications and robustness and things like that.

Java on the server will be platform-independent, and the client depends a lot on things like how the litigation plays.

The technical issues will be fixed independently of the battles between Sun and Microsoft.

The future for Java breaking into the mainstream with the server components, and an incredibly scalable JVM on the server -- in our case one that is going to be probably higher in performance than its competitors -- is a key part of that value proposition going forward.

To put that into perspective, we've got an architecture that solves the final piece of the problem. If you take a look at the architectures that I mentioned earlier, they tend to be client or applications category specific; they're all important, but they're missing the component that provides them the manageability and the distributed Java solutions. By bringing out something we call the open solutions architecture, we can deliver the back end of the services that make this whole value proposition happen, better than anyone else.

The key is to build an infrastructure that is targeting networks, developers and ISPs. Then we can provide the complete management infrastructure on top of them and make the rest of this stuff work.

We are building all of the pieces. This is a diagram which shows you each of the different layers -- and I know it's a busy slide. The important thing is that it's based on the principles of the Internet. Each of the set of interfaces is open. Anybody can adopt them. In most cases, they're already industry standard, and, in those others, we will work with the appropriate partners. By doing so, we have a strong commitment to open interfaces on the server side, -- which, again, is where the action is going to be -- to build these business critical applications across the net.

Now, how do you know how you're doing, against this sort of broad panoply of choices and direction that I've laid out? I tried in one slide to sort of say, well, what's the waypoint? A waypoint is a progress point to the network age. For example, e-mail; everybody in your company needs to use e-mail. You need to use internal web stuff. The security policy has to be there. The bandwidth has to be there, and you have to manage it in an interesting way. There has to be a directory. The applications have to be integrated with it. And the various other enabling solutions all must be present. Most customers, when they've used this, struggle with getting all this deployed. This is ground zero in the networking industry today. Getting these things implemented is the next stage. These are the things that are on everyone's very short-term time horizon. And, again, Novell has solutions in each of these cases.







So, where are we and how do we grow our business? How do we make our company really go? I am fortunate to be involved with a company where most people in the industry are already customers. So we begin by saying, "You're probably already a network customer, and NetWare is going to be completely TCP/IP capable in the next six to nine months." And everybody says, "That's a good idea."

Along the way, we're deploying this directory called NDS. Many customers don't know this, by the way. They are aware that there's a directory, but they don't understand what it's good for. They use it for, say, sign-on. But once you have this directory, you're enabled in all sorts of the ways that I described. By having the directory-enabled applications that I described and others that we will be announcing on a monthly basis over the next six months, we can deliver network-centric applications. The progress for our ManageWise business and our GroupWise business is all part of that. The BorderManager products are all integrated with that.

And because we do dedicated kernels and because we do networking software that does networking and nothing else, our products are more -- they have the abilities; they're more scalable, they're more reliable, they're more available. All of those abilities turn out to make sense if you have a best-of-breed approach.

That story is sufficient to grow Novell's business, to restore its leadership position in key parts of the networking industry. But there's a twist, and the twist is Java. By adding the ability to run these Java applications that I highlighted briefly for you, we can actually become a significant need to your server vendor as well.

So those are the stages. That's how we put all this stuff together.

In taking the lead, we as a company are now going through the first phase -- think of it as a stabilized phase, a learn-how-to; things like Management 101, deliver products on time, solve your customers problems of one kind or another. That phase probably goes on for another six to nine months until this next generation of products is shipping and customers are excited about it.

And then there's an opportunity to drive and an opportunity to thrive. The drive part is focusing on innovation in the standards world, driving new applications categories along the lines that I described, and really driving the network services ISP integration. There's enormous growth in those markets.

Once we deliver on that, it's possible, for example, to get to this interesting ultimate destination that is very much Java-based. I think it will take a couple of years for all of that to happen. And to the degree that leadership is inventing the future, Novell will lead and others will follow in this new world.

I want to go back to my earlier theme, which is that the new face of networking has a human face. It has your face and my face on it. I've taken you through a clear understanding of where networking and Novell are going and what our product portfolio looks like. We will win because we have a product strategy in place that will reconcile the drive towards this new face of networking with the landscape, the reality that many of you deal with. We are one of the companies on the verge of one of the most exciting new areas of network




computing. Network computing is the interesting thing to be doing. I'm glad to be part of it, and I want to thank you all for being part of it, too.