Five Questions With John Patrick, VP of Internet Technology at IBM
Thursday, August 3, 2000
The third of the “Five Questions With…” series is a conversation with John Patrick, VP of Internet Technology at IBM. He talks about peer-to-peer technologies, Linux, the potential of XML, the next generation Internet, and privacy.
By Jim Welte
John Patrick, the VP of Internet Technology at IBM (IBM,info), is one of the more lucid fonts of knowledge in the information technology community. He spoke with Business2.com on July 27, 2000 about some of the technologies he thinks will have a major impact on business, the next generation Internet (NGI), and online privacy.
What are your thoughts on the resurgence of peer-to-peer technologies, and is IBM doing anything at this point in that space?
In thinking about this in a larger sense, I think it’s inevitable that these kinds of technologies happen. Frankly, it’s a little surprising that the resurgence didn’t happen earlier, because when you think about the Internet, every computer is connected to every other computer, so that opens up a lot of possibilities. Exchanging information is one of the obvious things that you would expect to come along. So when I think about Napster and Gnutella and some of those technologies, and when you think about the various intellectual property and legal and those kinds of issues, I think of all that as just a blip on the radar screen. The real issue isn’t, ‘Does Napster win or does Napster not win?’ The issue is that we have the evolution of the Internet really taking hold and it’s putting a lot of power in the hands of people.
And in many ways the Internet is all about a really massive transfer of power from institutions to people, and I’m not talking about anarchy or about people marching in the streets. I’m talking about the power of a click, whether it’s a mouse on a PC or button on a cell phone. So the bigger picture and the impact on IBM is actually quite positive. Because people are becoming empowered, and they’re going to use the Internet to do a lot of things they don’t do today. I’ve been saying lately that the number of people actually doing something on the Internet right now as a percentage of the world’s population is a couple of percent. And if you look at, for those people using the Internet, the percentage of things they’re doing compared to the percentage of things they could do on the Internet, it’s probably a couple percent. Pretty minute. So we can surf Web pages and we can buy books and there are a lot of things we can do, but the list of things we can’t do is longer than the list of things we can do. You can’t renew your driver’s license for example. So Napster and Gnutella are raising the visibility of the fact that everything is connected to everything, and the impact for IBM is very simple: more transactions. And more transactions mean more disk space, more processing power, more services to help glue things together, and more middleware to help build the applications. So it’s all good news, and that’s why our Internet business is so strong right now.
What do you think of Microsoft’s .Net?
Well, to be honest, I don’t have much to say about Microsoft (MSFT,info). They’re doing a lot of interesting things, but the era of the PC as the center of what’s going on with the Internet is over. I’m not suggesting that the PC is over, but the era of the PC as the center of the Internet is finished. And we’re now in the era of an open standards-based set of things. Linux is the tip of the iceberg. So with regard to Microsoft I guess the question is, ‘Will they be able to play effectively in an open standards world?” and I hope they do. But that remains to be seen.
IBM has made a big push with Linux recently. Give us a sense of how you see Linux developing in terms of market share over the next few years.
Well, we think that Linux is a really, really big deal. And I guess I would start by saying that in the way I think of it, over the last thirty years there have been three really big shifts in the information technology industry. There are a lot of ways to talk about different eras, but to me there have been three really significant shifts. The first one was the PC in 1981, the second one was TCP/IP in 1991. IP was around before 1991, but 1991 was about the time it started to rear its head as a serious network protocol. And then in 1999, Linux.
All three of these things have five things in common. First, in all three cases we saw venture capital move into the area. Second, we saw some very smart people, thought leaders, from universities and companies of all kinds, quit their jobs and go into new areas. And third, there was a grass roots nature to it. None of these three things were top-down, orchestrated strategies of any company or any organization. They were grass roots situations. And number four, they were standards-oriented. When I say standards-oriented, I mean that Linux isn’t really a standard, the PC isn’t really a standard, but they were standards-oriented. And number five, and maybe most profound, is that in all three cases, established companies that had a stake in an alternative to these three things, all said, ‘Who needs it?’ You know, with the PC, a lot of IT companies said, ‘Who needs it?’ With networking protocols, I think back at that time there were probably fifty different protocols, and a lot of companies said, ‘TCP/IP? Are you kidding? I mean, okay for universities, but not for anything serious.’ And then in 1999, when Linux started to become very visible and people started to write about it, the same thing happened. Ed Zander, from Sun Microsystems (SUNW,info), said something I remember reading. I love his quote. He said, ‘Linux? Who needs it? We have Solaris, and it’s better.’ And Microsoft said, “Linux? Who needs it? We have Windows and it’s better.’ And at IBM, and we didn’t say, ‘Who needs Linux? We have OS400, or AIX or OS390.’ We said, ‘Everybody needs Linux,’ and I suppose you could say that only the greatest sinners know how to repent, you know?
Another way to say this is that we’ve seen this movie a couple of times before, so we really view Linux as serious. But when we say serious, does that mean we’re planning to replace all our operating systems with Linux any time soon? No, of course not. Linux does not yet have the scalability, the availability, or the reliability, so it’s missing some things. But again, that’s a blip on the radar screen, and with regard to our strategy, we plan on filling in the gaps that Linux doesn’t have. For example, we contributed the journaled file system (JFS) from AIX and OS2. We contributed that into open source, and our people, who are experts on JFS, are working with the community to get that moved to Linux. And we’re evaluating other technologies in the scalability and manageability area that have potential to be moved into the Linux open source community. We’re looking at various development tool ideas and we’re looking at a lot of things, and you should expect to see a pretty steady stream of contributions from IBM that help make Linux be all it can be.
In terms of market share; it’s hard to say, you have to look at the segments. Linux has become a very significant contender at the very high end. The University of New Mexico has put in what I think is the twenty-fourth largest super computer in the world, and it’s constructed from two hundred fifty-six IBM Netfinity servers running Linux. And then at the other end of the extreme, you have a lot of things going on in the embedded world with Linux. And the real sleeper here for Linux is Asia, because there’s a lot of appeal for Linux in Asia. A lot of creative talent and governments see Linux as something that’s not controlled by somebody thousands of miles away–something they can contribute to on an equal footing, and that’s inexpensive. It’s appealing in many different respects. So that’s a real sleeper.
The other real sleeper is the desktop. I wouldn’t be so bold as to predict numbers on this, but a lot of things are happening. A lot of creative people are doing things. To me, the Java language was very similar. A lot of people were saying, ‘Well, it’s just another programming language.’ But I said that if you want to know if it’s real or not, and if it’s enduring, all you have to do is go to Amazon (AMZN,info) and do a search on Linux books. Or walk into an Encore book store, or any bookstore, or go to a college campus where there are a lot of computer science students, and say, ‘How many of you people think that Java is really neat and know something about Java?’ And all the hands go up. And that’s what’s happening right now with Linux. Beijing University has Linux running on the OS390, as does Princeton, the University of Nebraska, and probably hundreds of other universities that I’m not aware of. They’re using VM to put native Linux under VM in a logical partition of a 390 and are able to give every student his or her own virtual Linux server. That’s pretty powerful stuff. So when I look at that collection of things, I say, ‘Hmm. Is this a flash in the pan?’ I don’t think so.
How far away are we from seeing industries adopt XML as a standard language and why?
XML is picking up a real head of steam. XML is to the Web like the Web was to email. That’s the way I think of it. In other words, it’s a really big step. And the simple way to think about this is that today the Web is a couple billion, more or less random Web pages, and they all have tags, and the tags describe what the pages look like. What XML does is provide a new set of tags that describe the context of a page, in other words, what the page means. Now that’s a big, steaming deal. Because now Web pages on one server can interact with Web pages on other servers. So servers are going to be talking to servers, not just clients talking to servers. And you’re going to see this in e-marketplaces. IBM has staked out a very significant leadership position in this space because we believe it’s going to be huge. A group of companies getting together in an industry to save ten percent on their purchases, that’s the tip of the iceberg. The real potential is to use XML to standardize on vocabularies that allow part numbers, bill material numbers, customer numbers, document types, shipping characteristics, postal codes, and all the things that let everyone use the same language. And when you think about the inefficiencies that exist in the industrial world with faxes. So XML is a huge deal and you see many examples of this. We’re doing a lot of things at IBM with XML. If you go to the XML zone at developerWorks, you’ll find a huge number of tools. We had the first XML parser in the industry, and we put it out on Alphaworks. We developed something called TPAML, the Trading Partner Agreement Markup Language, which allows contracts to be expressed in XML, so we can have common terminology–things like lease, lessor, purchaser, customer, and so forth. So yes, I think XML is not just another protocol. It’s the evolution of the Web.
You gave a presentation on the next generation Internet and the seven characteristics of it. The fourth characteristic mentioned that being natural was key to the Internet’s development. Could you elaborate on that idea?
Well, for each of the seven characteristics, there’s a lot, because we’re talking about basically all of IT. So there are a number of examples with each one. But just to give you a flavor of it, one of the things about the Web today is that it’s really not natural for a lot of people. It’s natural for you and me, we live there, but envision language barriers, for example. The Web is not natural when you have to go to customer service and speak the language that may not be one that’s comfortable for you. So when you think about combining instant messaging with voice-recognition, with text-to-speech, you really get a real-time, multilingual intercom. And the vision is that a person using instant messaging asks a question of customer service, say in Spanish. And the question gets routed to the most knowledgeable person for that subject area. And that person is Chinese and when he or she answers the question in Mandarin, the questioner hears the answer in Spanish. That’s getting more natural. And instant messaging, by the way, is a part of this as well. When you think about the adoption rate of instant messaging, even compared to the adoption rate of the Web, it’s tremendous. And why is that? The reason is that it’s natural. It’s like a walkie-talkie. It’s like two cans and a long string, like kids used to do. It’s very natural. And in fact, at IBM, we’re doing a lot with instant messaging. We’re creating bots, what we call buddy bots, that are actually software agents that live in the instant messaging buddy list. So on your list you have people, but you also have bots, and the bots do things for you. So you send an instant message to the weather bot and say, ‘weather, Boston,’ and it gives you the weather. Or you say, ‘What is NGI?’ and it comes back with ‘next generation Internet, click here for the URL.’ So instant messaging is going to make NGI a lot more natural. For a lot of people, even Eudora email software or certainly the Web, are not natural. But people can relate to a simple little TCP/IP client–an instant messaging client sitting on the desktop–like a walkie-talkie.
You recently co-wrote an editorial with Vint Cerf for The Financial Times in which you addressed the role of policy makers and law enforcement officials in regulating and policing the Internet. Can you speak to that issue in a little more depth?
I think the essence of what Vint and I were saying, is that the private sector really needs to take a leadership role on the key policy issues. I’m not saying that government has no role. I’m saying that the government should provide encouragement, seed money in very advanced research areas, and should be a role model in using the Internet themselves. By the way, I think the U.S. Government has done very well on all three of those areas. But when it comes to more operational things, like privacy, security, almost any area of taxation, or jurisdiction, the private sector really needs to take the lead. And if the private sector sits back and waits for somebody to pass a law to prevent spoofing, it’s not good for a lot of reasons. It takes too long, and once you get it, you’re going to be sorry you asked for it. Because then you have to implement very specifically, and likely that would be costly, and so on.
What we tried to do with our paper was raise the visibility about the particular issue of what was going on and encourage people to take steps. The biggest issue with security that I see has nothing to do with technology at all. It has to do with management policy and attitude. We all know how we feel if someone in one of our companies rips us off on an expense account. We know how we feel about that. How do we feel about an employee who puts his or her password on a Post-It, and sticks it on a PC, or under a mouse pad? How do we feel about a firewall administrator when we find out that the password of the firewall is the word password, which one study showed that seventy percent of them were? How do we feel about that? So the private sector needs to take these issues really seriously, and adopt very strong policies and audit them. And the same goes for ISPs. ISPs can’t be responsible for all known problems, but there are things that ISPs working together can do in terms of headers and things in emails so that when somebody is trying to gum up the system, it’s possible to find out where the problem came from.