The TCP/IP Pathway to Open Systems

Computing Japan: The obvious first question is: How did an American company come to he named Wollongong?

Dan Ladermann: When we founded the company in 1980, our products were going to be based on open systems, and at the time, the emerging open system was UNIX. We found some software that was being developed at the University of Wollongong in Australia -- version 6 UNIX -- got the rights, and came out in 1980 with the first commercially supported 32bit UNIX system in existence.

We were trying to come up with a name, and in 1980 there were a lot of companies in Silicon Valley with Compu-this, something-Tech names, and we didn't want to be one that you couldn't tell from anyplace else. We found out that Wollongong was an aborig inal name for "beautiful place by the sea," and we're in Pale Alto, California, right by the Bay, so we decided that would be a great name.

CJ: You mentioned the term "open systems. " I hear that a lot -- it seems to be a popular buzzword without a clear definition. What is your concept of "open systems "?

David Langlais: Let me give you the first and best example of an open system -- the IBM PC bus architecture. The reason that's an open system is that people are free to design, implement, distribute, and sell products based on that specification, witho ut royalties or payments to anyone. That, I and many others believe, is the single most important event in PC computing. It spawned an umpteen-billion dollar industry: disk drives, memory cards, video cards, controllers. Without it, the IBM PC would not h ave been successful.

TCP/IP is the networking communications parallel to the PC bus. Nobody owns the specifications of TCP/IP, and anyone and everyone is free to design, build, distribute, and sell their version of those specifications.

To us, open systems means exclusively those specifications that are publicly available -- freely available to anyone and everyone to build competitive products. Windows is not an open standard; it is not an open system.

CJ: Tell me a bit about your flagship product: Pathway.

Langlais: Pathway is an umbrella name for our products across all platforms, everywhere from UNIX to Macintosh, PCs to VMS. They all carry common components.

CJ: From a personal user's point of view, if I buy Pathway and put it on my computer at home, what's it going to do for me?

Langlais: To a large extent, that depends on what you want it to do for you. You can use it to dial in to your host at work, and do everything from home that you could do attached to your network at work log in to another host, use an application, tran sfer files, send stuff to printers. It gives you the ability to do at home everything that you need to do at work -- plus you get to have a little bit more fun.

CJ: Let's move up the ladder. Suppose I'm in a small office -- 8 or 20 PCs and a couple of Macs hooked up to a server running Novell. What will Pathway do for me in that type of situation?

Langlais: In that type of situation, users tend to make a choice. If they are going to become or stay servercentric, then Pathway is typically used if they need to go to a small UNIX system -- or if they need remote services, because NetWare provides v ery poor remote services. We actually position Pathway as an alternative to Novell; we believe that it is a much better long-term investment for a small business than Novell, because Novell does not grow very well at all.

What Pathway does is allow you to build user-centric networks, where your desktop is the center of the network. And through the use of the Open protocols, like TCP/IP and NFS -- you could use a Novell server, because it supports TCP/IP and NFS; but t hen as your needs grow, if you were using Pathway and needed a larger server-, more power, you could put a Sun server, or an LBM server, or an HP server into your environment and change nothing. You just add a new machine, and everybody gets the capabilit y, regardless of whether they're a PC or a Macintosh. Proprietary server-centric solutions won't let you do that.

CJ: How is your message received in Japan?

Langlais: The Japanese market does not believe that today, because Novell and Microsoft have invested tens of millions of dollars, with the same marketing message that they had in the Unites States in 1985. What they're saying is that the way for a sma ll business is to put 3 Novell server in there, because administration is centralized, and it's very easy. They use fear, uncertainty, and doubt to say, "TCP/IP -- that's what you use to connect to UNIX."

If you hear the marketing message of Novell in Europe and in the United States, it is not the same message they're doing here. In the United States, they realize that the marketplace has gone through that phase of building small, disconnected, server-c entric networks, and what they need now are large groups of interconnected networks. Even small businesses need that. I think Novell will continue to be very successful in Japan, but I also believe that their success is time limited; they will have to mak e the same transitions here in a couple of years that they have had to make in the United States.

CJ: The president of Novell Japan recently said the company will grow 50% over last year's sales.

Langlais: Yeah, 50% ... going from a dime to fifteen cents is impressive growth, but at the end of the day, you still have fifteen cents. Novell is investing in Japan, that's what they're doing, and the investment is starting to pay off.

Japan has always had a focus on open systems, a focus on client/server. But Novell is not client/server: it's server/server. Everything happens on the server.

Client/server really means client/servers, because in a real client/server environment, there's not just one server. You have your file and print server, which is probably Novell. And that's OK, it's a great file and print server; but it's not a great database server. So you have your database server, which is probably a UNIX machine or a symmetrical multiprocessing UNIX server. And you've probably got your applications sitting out on an IBM mainframe, or maybe on your UNIX box, or another type of serv er. So the user ends up using many types of servers. That's an extremely cost-effective way to build a network.

CJ: Is part of the reluctance that Japanese management doesn't understand how easy this is

Langlais: Dan and I have both been traveling to Japan once or twice a year for the last -- I don't know how man!: years. We definitely understand that it is a very long-term sale, because we're saying something that really has not been said before. And it really does seem to involve an incredible amount of risk, because instead of having a very close relationship with one vendor, you now have a variety of relationships with a variety of vendors.

But if you want open systems, if you want open networks -- with that freedom comes responsibility. If companies want to move away from proprietary environments, proprietary systems, they have to accept the responsibility for themselves to choose, rathe r than letting someone else choose for them. That is the essence of an open system: you have the opportunity to choose, and you have the responsibility to choose. Making people feel comfortable about choosing is the hard part.

CJ: Japanese managers are not used to being given so many options.

Langlais: Yes, although you can see the cracks in the wall, and they're getting wider and wider every time come over. I meet with some fairly high-level people, at companies like NEC and Toshiba, and these guys are beginning to understand it.

CJ: How no you see the Japanese market developing over the next five to ten years?

Langlais: We've been in the Japanese market since 1984 with Data Control, and it has undergone very similar evolutions as the US market did. What is different about the Japanese market is that, while all the minicomputers were being connected in the US , Novell was starting to come along very, very strong. And so you had a prepositioned base of TCP/IP already there. When the PC LANs reached a critical mass -- where they needed to get to the VMS applications, or the UNIX databases, or the mainframes -- t here were already widespread networks of PCs; they just weren't doing TCP/IP.

Japan went through all of the minicomputers and the UNIX workstations and the technical markets. So TCP/IP is there, but unfortunately there is very little networking of the PCs right now. That has not followed the same lines as in the US. We believe that over the next couple of years, the networking of PCs, in any type of network, will continue to accelerate far in excess of 50% to 70% per year. What took the United States four to five years to achieve will probably happen in Japan in two-and-a-half . In two or two-and-a-half years, you will see the adoption rates of TCP/IP for desktop start reaching an almost vertical slope.

The TCP/IP market in Japan now is restricted simply by the number of PCs that actually get connected on networks. That is taking longer than I think almost everybody anticipated.

CJ: What type of efforts do you go to localize your products, particularly to Japanise them?

Langlais: We partner with Data Control; and it's really a partnership, though it's not a joint venture or anything like that. Not only are they a distributor, but they are also, you could say, an OEM of our products. We've licensed them to do the two b yte support and localization of our Pathway products for Macintosh, Windows, and DOS. By trial and error, we are getting much better at architecting and designing our software to be easily localized into the Asian languages.

It's relatively simple to localize our products for Latin languages, but double-byte languages cause significant difficulties. Data Control has localized two versions of our DOS and Macintosh products.

Ladermann: And we're on our third generation of localized products.

Langlais: We actually will be expanding the exchange of our technologies. In the past, when Data Control did the double-byte extensions, they kept them over here for support and maintenance purposes. But in the future, what we're going to do, when they get double-byte support done, is actually take the double-byte support back and embed it into the future versions of our products. So not only will it be much easier for Data Control to do the next versions, but it will also be available for us to go int o the Korean and Chinese marketplaces.

CJ: Aside from the language difficulty, are there any particular problems localizing for the Japanese market?

Langlais: No more difficulties than you have going to any other foreign language, where the language itself constitutes a different culture. Our manuals and our messages to the user are in American English, and many of them don't make sense even to som e of our European cousins; they just don't translate well into French or German, or Japanese. So, developing the text communication with the user, as they're using the product, is particularly difficult and requires a very high level of sophistication. Th at is something we will achieve over time.

But from a very low-level technical perspective, TCP/IP and the TCP/IP applications are very adaptable for double-byte use. In fact, much more easily adapted than your line-based communication protocols or your screen-based communication protocols.

CJ: How do you rate your sales success in the Japanese market, compared with the US or European markets?

Langlais: Sales in the Japanese market have traditionally been very strong, although over the past couple of years we have undergone a significant transition. We hold a dominant market share of TCWIP on VMS machines in Japan -- many, many more than Dig ital themselves have for their products and so those sales have traditionally been very good, although the last couple of years, with all of Digital's problems and the aggressiveness of IBM, HP, and Sun in DEC accounts, that revenue growth has really slow ed down and actually has started to decline.

At the same time, in PC sales, our growth rates are very strong, both here in Japan and the rest of the world. The problem is that a much higher number of units is needed to achieve the same revenue growth. So our revenue growth here in Japan has not been as strong as it has been in other areas of the world. We know that the growth rates are there, though, and that as the adoption of networks in Japan grows, our revenue growth over here will move back up to what it has been in Europe and in the Unite d States.

CJ: How do you foresee TCP/IP, and your products, developing?

Langlais: The foundation focus is always to stay current with the changes of TCP/IP. TCP/IP is not the same as it was 29 years ago, when it was developed; it won't be the same three years from now.

One area that we're going to concentrate on is providing products, tools, and capabilities that allow network managers and administrators to manage not the network, but to manage the information that is going on the network. That is not done today. T he key point is that they have to be able to manage one server connected directly just as easily as managing a remote server that they never see. We have some technologies we are working on that, over the next couple of years, will provide some unique cap abilities.

Another area of focus is providing users with advanced ways of using information. I know -- that sounds like a nothing sentence. But some of that information is graphical, or it's voice, or eventually it may be video. The combination of those in what we consider today normal, everyday electronic mail is the aim. And being able to have compound documents easily exchanged anywhere. That means everything from voice mail combining with your electronic mail, to the ability to do business over the networks without the unnecessary complications imposed by EDI.

There's lots more to do in TCP/IP. The TCP/IP market is not threatened in any form by any other company like Microsoft bringing out a portion of the stack. All it's going to do is grow the market -- and the market opportunities -- by orders of magnit ude.

CJ: Let me wrap up with a general question: How do you view the Internet? There have some people who claim that it is "the sum total of human knowledge.... "

Langlais: [laughter] That's not exactly a ringing endorsement for human knowledge.

CJ: Others argue that the infrastructure isn't adequate, like when parts of the Internet almost came to a standstill because everyone was trying to get In and download photos of the comets hitting Jupiter.

Langlais: We were the first commercial company to be put on the old Defense datanet -- back in 1984. And we were actually on the military side, because of the work we were doing with the Department of Defense. We were out there when the Internet was 45 0 computers attached to the ARPAnet data network. We've seen it grow all the way through.

It's not the sum total of human knowledge. It's a somewhat disorganized, limitless library, and it's got all the problems that a library has. If you don't understand the Dewey Decimal System very well, it's pretty difficult to find what you need to fin d. So you waste a lot of time cruising up and down the aisles, looking at the books trying to find your way around. For the people who really understand how the Internet is organized, though, you can not only find information, but you can take pieces of i nformation from the library and combine them together to do yet other things. And that is the promise and potential of the Internet.

I said it's a somewhat disorganized library, because there are no librarians. There are not a whole lot of rules, so at various times people are whispering, people are shouting -- it's somewhat of an anarchy, though an anarchy that has a definite set of social mores. It is somewhat self-regulating, and needs to be; we certainly don't need the same type of government organization on the Internet that we've had in almost every other facet of our lives.

CJ: Can there, or should there, be limits?

Langlais: When we start limiting access to information, we limit the ability of people to do things with that information. The Internet is unique -- it has incredible amounts of really good scientific data, and other types of information are complete g arbage. It allows everybody to pick and choose what they want. That's really important. The Internet needs better technical infrastructure because people will start using graphical user interfaces or the browsers like Mosaic. Traffic on the Internet will grow not geometrically, it will grow exponentially, and by orders of magnitude per month. And that's going to be very expensive to maintain.

CJ: What do you foresee in terms of government intervention Will governments try to regulate the Internet?

Langlais: Sure they will. Maybe there's some things that aren't good, from your own personal perspective, about the Internet. And so somebody, somewhere, will try to regulate it. On the pornography issue, the government will try to regulate it in Japan ; the religious conservatives in the US will try to regulate it there. The Europeans will simply shake their heads and laugh at us. "Why are you bothering? There are lots of other problems to solve than whether or not somebody wants to look at pictures of naked people."

But there will be all sorts of limits and changes that the government will want to have. And that's not necessarily bad; it's all part of growing up. When you're an adolescent, you have to learn what the rules are in order to become an adult and set your own rules. And right now the Internet is in that extremely explosive, youthful exuberance where it's trying everything. Some of the things are good, and some of the things aren't so good. One of the problems on the Internet is a very small group of p eople who are very, very vocal, who don't want any type of government intrusion, any type of control by anybody. The problem with that is -- that's not reality.

CJ: Recently, one of the more vocal groups has been those decrying "commercialization" of the Internet. What do you see as the future of corporate use of the Internet? Langlais: [chuckles] I'll have to make sure that whatever I say is quotable.

I have some pretty strong views on those types of people; what they need to do is essentially grow up. The largest vocal group is centered in the United States. Dan and I are taxpayers in the United States. We pay for the Internet; they use it for fr ee. The Internet is not a right, it's a privilege. It should be treated as such.

They don't own the Internet. The beauty of the Internet is that you're free to participate however you like. Just like in everyday life: don't impose on me what to do or not to do. And that's what those people are trying to do; they want you, or compan ies, to live by the rules that they feel are the right way to go.

I don't think that's right, and it really doesn't matter. These people are a vocal minority, so they get a lot of press -- Newsweek writes about them, Time writes about them. You get the flame wars out on the Net; they turn around and try to break so mebody's server, or something like that, because they advertised on the Net. Fine, that means companies like us will simply build better and better protection products for servers, and pretty soon these people will yell and scream their way out of existen ce.

Ladermann: The Internet is a collection of networks, of which there are now a lot of commercial pieces. A lot of corporations now build their own private networks, and also go to public carriers to build these networks, which ... they're not public net works; they are services like the phone system. Long distance service isn't free on your phone; same thing with the Internet and the carriers. And now you're getting all the major carriers providing Internet access service, which is good because it's also building an infrastructure that can grow.

CJ: So you would oppose any attempts to control commercial use of the Internet?

Langlais: I think commercial use of the Internet needs to be organized -- not controlled, but it needs to be organized. It can do so many different things that other types of commerce can't.

If we don't find out how to use the Internet as a commercial tool, we will not be able to sustain its use, because it is incredibly expensive to run. And we won't be able to afford to expand it unless there is some way of gaining commercial value. Co mmercial use is vital to the Internet's future.