Interview with Netscape's Eric Hahn
DS: Where do we stand on the great stage of shifting paradigms?
CB: We're between modernism and postmodernism. To get an idea what I mean by these, read Sherry Turkle's book (Life on the Screen: Identity in the Age of the Internet, Simon & Schuster, 1995).
DS: Well, she offers a variety of definitions, but as I understand what you're saying about directory, modernism says certain things are true while postmodernism says anything is true.
CB: And that there are multiple truths.
DS: In other words, we hold these truths to be self evident and non-contradictory.
CB: Right. The traditional, or modern, paradigm, insists that there are certain truths and they are unchanging and monolithic. There are a variety of these, and they stand separately, like islands. Modernism is industrial. Second wave. A few big companies, each with their own sets of truths, try to dominate and successfully create competing and often incompatible monoliths. So the customer chooses an island and is forced to live there because the whole island is the choice — not parts of it. It's all or nothing. These islands were powerful, but now they're going away. Now we're looking at a sprawl of roads and intersections that changes constantly and nobody can control. We need a new paradigm for that. The postmodern paradigm works there.
DS: In the Industrial Age, we accredited great size, mass and especially capital holdings, with power. Wealth was in possessions. In the postmodern world these became anchors — balls shackled to the legs of corporations and governments. Impediments to change.
CB: Yup.
DS: If we look back at your notion of directory as a way to manage change, we may begin to see where the paradigm shifts. Because once we can manage change, we can to look at change and give full respect to change. In the process we give up our static views of computing — of files and sectors and storage and fully-formed tools in boxes, employed to create finished products. The notion of a changing computing world is as hard to hold in one's mind as the contradiction Heisenberg talked about: if you can say where something is, it's not going anywhere; and if you can say where something is going, it isn't staying in one place. Our point of view for a long time — our modern point of view — has been fixed in where things are... in the physical, as you put it. Even when things changed we had to deal with them in physical terms. And the physical world argued for single truths. Take occupation. If someone asks "What are you?" and I say "I am a carpenter," it's hard to say I am also something else. I have a single truth. But in the postmodern world that Turkle talks about, I can have many identities, and they do not necessarily contradict each other. We expose what Michael Ventura called "the myth of the monopersonality." We have many personalities, he says, and the only thing that keeps me sane is that I have managed to integrate them behind one proper noun, or one pronoun.
CB: What directory does is let you deal with the fact that there are multiple truths, in a sane way.
What we've had so far has been insane. Companies said, "Oh my God, the world's changing! We need to move to a new truth!" And we were forced to choose the IBM truth or the UNIX truth or the Apple truth or the Microsoft truth... And each was an island. There were roads in and out, but there were toll booths on them.
With LDAP, Netscape just drove a bulldozer through the directory intersection, and the toll booths are gone. The traffic is flowing fast and furious, and the guy in the toll booth is looking for ways to make money in a postmodern world, where his truth is just one among many. Microsoft won't like it because it's not their bulldozer, and it's not their API. But they'll adapt. They always do.
DS: What is that perspective?
CB: We asked, "how is it really going to look to get where we want to go with distributed computing? And the answer was, by reconceiving networks as services and OSes as commodities. We saw it had to happen, and now it has happened.
The question that everyone needs to ask is, "What is it that gives the customer choice?" Not "What is it that forces the customer to buy my product?"
DS: What does this do to change existing businesses and create new ones?
CB: It creates opportunities for directory service businesses to operate on the Novell model. They'll say, "Buy my services. Pay me per user for going into my universal space. Every universal existence of an individual pays a fee."
I don't know if I want to say directory services require a "directory server," though they might. But I can say that the directory services business will be no different than what Novell's model is right now. You pay for a 50 user service, and 50 users can log in for $49.95. Some customers pay license fees. Buy a license for 20000 people for a single fee, regardless of how many servers you set up. You have a hybrid of stairstep server-based pricing and flat user-based pricing that has nothing to do with the number of users you have. It all depends on how you buy software, and how you're structured, and what your business is.
The Novell model holds, and it's a good one. Novell invented the software services business. Nobody had ever made money that way before. But Novell only did it with file service. There are many other core network services that other companies needed to establish through the same business model. So now instead of hardware independent file services, you have operating system independent network services, including directory services, messaging services and the rest of them.
DS: Okay, we know what Netscape's purchase of LDAP does to Microsoft, but what does it do for the rest of us?
CB: First, it means that none of us, from Microsoft on down, are going to be forcing users to buy and use our exclusive services. Take Novell. Before their approach was, "If you want NDS (NetWare Directory Services), you have to buy it from us. As soon as Novell puts LDAP on the table — which they just did, at the same time as the Netscape announcement (see the Novell press release) — they give us a real choice. We can buy NDS from Novell or some other directory service, from Netscape or some other company. And the reality is that, since all of Netscape's applications are speaking LDAP, they have no idea whether they're talking to NDS or Netscape. They are independent of each other. This is the network services model at work. There are abstraction layers between every level of service and functionality in the infrastructure that are completely independent of everything else. So I don't need to know what platform you're running on, or whose service it is, or what application it is, or what client it is. Now you get complete choice at every level of network computing, from the client to the operating system, to the browser, to the services to the access to those services. It's a flat field. Anybody can play.
DS: And Netscape's standardizing on LDAP is what flattens the field.
CB: Right. It standardizes at the right layers. So, for example, in the client I can say "I'm going to do HTTP and LDAP." That means I don't know whose HTTP server I'm talking to or what platform its running on. They're not tied together. They're not dependent on each other. They're completely independent. On the service side, I'll say "I service HTTP and LDAP. Anybody doing LDAP talking to me gets LDAP. Anybody doing HTTP talking to me gets HTML. If I'm doing a directory lookup, I'm doing LDAP. I don't know whose X.500 server is out there, or if it even is X.500. All I know is that it speaks my protocol. So we have choice. Now what vendors compete on, instead of locking customers in and forcing them to pay a toll, is selling better features and functions and prices to anybody on the road who wants to stop and buy those things.
DS: So the rule used to be "Survival of the Most Restrictive" —
CB: —And in the Network services world, it's now "Survival of the Best Featured." That means you've got to be running like hell, all the time.
DS: What will this do on the customers' side?
CB: Customers have to change as well. They can't keep saying "I want complete freedom of choice and I also want one vendor to do everything." You know what customers are saying to Novell and Banyan and Netscape right now? "Don't bring products out so fast. We want everything now... just not that fast." Well, customers are going to have to belly up to what they've been fighting for, which is complete freedom of choice. Figure out how to put into place the fact that things are changing and anybody can get stuff down off the Internet — rather than trying to restrict that.
DS: For many years it was as if every vendor tried to create an island on which customers could live. Now you're saying everybody has to share the same land, without restriction to what anybody can or can't buy.
CB: Yes and no. In Microsoft's case, MAPI and WOSA were ideas of trying to provide at least island separations that Microsoft controlled.
DS: Today, the day after the announcement, the press coverage of the story is minimal. In fact, there is no story: a few paragraphs here and there about Netscape's stock prices, which went up a couple of points. But there are some real stories here, and I'd like to talk about those.
At the bottom level, which I call the sportscast level, we have "Microsoft vs. Netscape, Round 2." This is the pure journalistic play: stories are about conflict, and these two companies have chosen each other off as opponents. While this may be a valid story, it is also limited in the scope one can cover in writing about it.
CB: Absolutely.
DS: There is a higher level story, a dull but important one that says, "This will change networked computing because it introduces capacities for intelligence and memory that were not here before."
Then there is another level, perhaps higher again, where this news effects business strategies. It says, "Business will change fundamentally because companies can cooperate in ways they never could before." It says, "Finally, a free, open and cooperative marketplace is finally here. The Third Wave finally arrives."
CB: These two higher-level stories are the ones I'd like to see told.
What we're getting down to here is that distributed computing requires a set of core infrastructures to work. It just can't happen without infrastructure. The way the vendors addressed this was by saying you do that by throwing out what you've got and buy my stuff and voilá: infrastructure.
DS: The forklift upgrade.
CB: This is a real business story. Here we are in 1996, and less than 15% of the entire market has directory services, and even fewer actually use it. There is no infrastructure.
And so, from the higher level, in spite of all the distractions about scoring between two rivals, you see this event goes a long way toward moving the game to a field where the fight isn't about who owns the field. It's about other choices that the customer would rather face than the one that requires he either keep or throw out everything.
The important point here is that we've moved. Finally, we've moved.
To the top
To Craig Burton Interview, Part 1
To Eric Hahn Interview
To WEBsmith
To Reality 2.0