What I am going to do today is open it up to a - we've got some media questions over here on my right and I'd like to open it up to our computer members as well. I might just say that at our previous lunch our computer members were very much into one way information flow, they didn't seem to be too keen to actually stand up and ask questions. If that's because you don't want to admit you're from Microsoft, they don't mind. You don't have to identify yourself but feel free to if you'd like to. But I would like to encourage everyone in the room to feel free to ask questions.
The first one though comes from Bill D'Arcy.
Bill D'Arcy: Mr Lane, Bill D'Arcy from PC Week. Welcome to the vertically integrated city. By way of background, we've got a process under way here at the present time by the Federal Government, conducted through the office of Government Information Technology which is examining - it's got what is called a scoping process under way. Part of that process is going to look at the outsourcing of data centres - IBM data centres et cetera. From what you had to say today it seems to me that we're fiddling at the margins, that really there are more fundamental issues associated with what governments should be involved in, then looking at, for example what it should be doing as far as data centres are concerned, as far as insuring that government departments can speak to each other. Am I correct in saying this: what effect is the network computer going to have when it comes to the delivery of government information services and what do you think can be learned out of what you've had to say today. I notice that you say we haven't reached the age of information yet. What can we learn out of that as we go about the process really of re-engineering government that hasn't - you talked about 30 years - I don't think it has ever been measured in this country.
Ray Lane: Well, you asked three questions I think. The first was out sourcing. I agree with you. Outsourcing is I believe a marginal issue. Outsourcing, although it's a huge monetary issue, it is marginal. The art of managing data centres is quite mature. Most large companies and large governments do a very efficient job and many run what is called a lights out data centre now, so you don't have many people that are needed to operate it. These large mainframe computers run pretty much uninterrupted and the computer itself and the disc drives and all of the heart of the system runs in a room where you literally turn the lights out and you have somebody sitting at a console that could operate it remotely.
So the data centre, you can't - it is a scale game so certainly medium size operations can benefit from adding their transactions, their work load into a larger work load and because of the development of microprocessor technology we do get scale advantage, by going into larger and larger data centres, larger and larger operations. But once you're very very large that becomes a marginal issue as well.
Now outsourcing to me is not perfect evaluation. In other words it doesn't make sense to say outsourcing makes sense or it doesn't make sense. It very much depends upon the operation or the customer, the company that's considering it. If you've traditionally not been a very good operator no question you're a candidate for outsourcing because there are those that may know how to do it better or have more scale. If efficiency is not what you seek then it may not make sense to outsource because typically the large outsourcers are not the first to understand new technology. It's not innovation that you get. It is large scale transaction processing and typically outsourcers are able to make money by using older technology rather than newer technology. They use operating systems that they design themselves for efficiency as opposed for innovation. So you can't - there is not a kind of one answer fits all here, and certainly it is a growing trend to outsource but it remains to be seen whether when a company makes a decision to outsource whether that has much benefit some companies will tell you yes, because they've done it in the last three or four years - don't think they really have a decision yet. The people who are saying yes, the people to made the decision, of course it was a good decision to them. After I think we know after seven or eight years whether it makes sense because these are all long term contracts by EDS or CSC or ISSC from IBM and remains to be seen.
But certainly the company doesn't feel like this is our core competency and we constantly have a need, an open head count for technical skills. I know every customer I talk to has open head count for Oracle skills and cannot keep enough skills on board. And so we've actually given some thought, not about outsourcing data centres but providing applications, our applications, financial, manufacturing, HR types of applications, with the network computer so that customers can basically install network computers on the desktop and not worry about installing the applications but just simply take advantage of a server that is run by somebody else. So that was question number one on outsourcing.
Number two , what will the network computer do for government? I think I mentioned some of the things in my remarks but I think it will allow certainly the major thing is lower the total cost of ownership. Today's desktop is anywhere, you've heard the figures, is anywhere from probably 10 to $15,000 Australian to manage that desktop. The cost of the computer, the updating of the computer with new software, the help desks and still we cannot manage that with any kind of accuracy. Most CIOs will tell you they cannot manage that environment because every PC is configured differently and so it's , if you can get that software running back on a server, put an NC on the desk top, you have exactly the same environment, you don't give up anything other than persistent storage, and if you're on the network all the time you do not need the persistent storage in the desktop. So you lower the desktop to about 3 to $4,000 Australian, so a tremendous drop in cost and if you got you know thousands and thousands of desktops this is a tremendous saving, so cost of ownership is a major benefit and governments surely have to be concerned about that, without giving up service and maybe increasing service, because if you can provide more NCs than PCs - you know, why doesn't everybody have a PC? The answer is simple, because they cost too damn much. We can't give everybody a PC.
So if you give everybody an NC, get everybody online, you could have a much more efficient government and manage it centrally. Manage it in a more precise way. So I think you can increase services a great deal. Now it will be a long time before every consumer has an NC that is able to take advantage, but web TV, other kinds of appliances, will allow same kind of access so that the faster we get that kind of penetration of low cost devices the more opportunity there is to offer services in electronic fashion. People standing in lines, filling out forms, doesn't make happy taxpayers.
Andrew Birmingham, Computer World: Two questions - first of all what do you see is the biggest impediment to the adoption of the network computer? And second an issue of core competencies. Isn't there a danger that you would be isolating your own core customers by talking too much about network computers and not talking enough about databases in applications?
Ray Lane: The biggest impediment - well, there's two and it's hard to select one because there are two markets here. There is a market, if you want to call it a market, of people that have personal computers. Now if they're making the decision, many times they're not, it's consolidated in an executive who will make the decision for them, because most personal computers are bought by companies, they are not bought by consumers, but if the consumer is allowed to make the decision of whether I have a personal computer or an NC then they do have an emotional hurdle to get across. That emotional hurdle is the giving up of persistent storage. My goodness, security, people will see things - yet they seem to be comfortable if you introduce other models like banks - where do you keep money? Do you keep your money under a mattress, in a box on your desk or do you keep it in a bank? And if you look at, and of course everybody that I know of that has a PC backs up their data every day, every single day and so I understand that giving it up to somebody who is a professional manager that might back it up even more frequently is some difficulties as well. But those that have gotten used to the PC model it's a very emotional thing. I know anybody that touches my personal computer, I bring it to Australia, I make some last minute changes to slides I present to an audience, and immediately I get here all the people on the show want to take it away from me an download it and use it and put it online and I said no, no, you do not touch my personal computer. Not because I'm concerned about my data, I'm concerned that when I get it back they will have changed something, and I can't use it, and it's an extremely fragile device. So I think that's one hurdle with those that have PCs.
The other hurdle with those that don't have PCs is probably a band width issue. Getting the band width in the networks. We have got a lot of homes passed b cable now, there is a lot of fibre going in, certainly local area networks through either Ethernet or ATM Technology will make the NC very popular in government institutions, schools, or companies that have local area networks. Not a band width problem there. You can use the Intranet very effectively, but when you are starting to bring consumers in or send E mail over the Internet then we will have a band width issue and that needs to be solved. So I would say those are the two impediments. I think your second question was about if we talk too much about the NC. Well the NC is all about - we look at it as we are a server company. We look as the database application meeting up with a more efficient distribution channel. So the Internet is all about distribution. So we would like to see more users, more people online because it creates a need for more data bases, more servers, and the NC is all about looking into those data bases. So I consider it all part of the same business, database, applications, NCs, that if we end up distributing databases and desktops with NT operating systems it's going to break, you just can't take it much further. So the NC is actually part of the database business so far as we are concerned.
John Hilvert, The Australian: Just following my colleague, two questions if I may. Am I correct in understanding that Oracle is actually moving its resources, its development of its future technology away from NT, if I heard you correctly you said NT was not a good platform for you to focus on and secondly, or maybe you'd like to elaborate a little more on a comment, and secondly there's been suggestions that the solution that Oracle, the actual architecture that you've been propounding is falling short in terms of not having a gateway to mainframe legacy applications. That's been asserted and is a comment in today's Australian. Do you want to comment on whether there's any merit in the fact that Oracle is actually turning its back on legacy applications?
Ray Lane: Just the opposite. First, I said just the opposite. NT is as important to us as UNIX. So it is a core development platform. What I mean by that is we develop all of our software on UNIX Solaris from Sun and on NT from Microsoft. Those are the two core, and then we put it to about 80 different computing plant forms, so HP UNIX, IBM AIX, the Macintosh OS, and so we consider NT a very important server platform, we do not consider it a very important desktop platform so there's two environments here. There's the desktop, what we call the client and there is the server. NT will be a very popular distributed application server and it is a very very popular platform for us to develop and we own about 40 per cent of that market. But on the desktop NT has no role in desktop as far as we are concerned. It is too complex, it is too heavy it is too fat. Unless you're an engineer or you're a software developer, then you would obviously develop using NT and you need it on the desktop. But for 90 per cent of users today there's just no reason to have NT on the desktop but because there's no alternative, you have banks, governments, all deciding well, because Microsoft says to do it, let's put it on the desktop. But it would be a very very complex environment to manage.
Your second question was about gateways. We provide more gateways than any other company in the business. We are committed to an open architecture, Microsoft is not. If you want to write applications you must write them to Active X, okay, not Corba, Active X. We can go to either. You must use visual basic or C++, doesn't matter what language you use for our architecture so and the gateways to legacy systems are absolutely established, back to DB2 to IMS to all sorts of legacy applications through the Oracle database we have probably got 50 or 60 different types of gateways to legacy applications. So, yes, total openness, and we hope we can compete on time to market because obviously if it's open everybody has the same opportunity to develop.
Norman Peters, Director of Systems of the Joint House Department of Parliament: Where does the issue of charging for the use of applications provided over a network fit with the scenario of the NC?
Ray Lane: For what charging?
Norman Peters: For click charging, for dollar cost for the use of a software suite or an application.
Ray Lane: Transaction charging? Well, we are trying to deal with it right now. We don't have the answer yet and because I think we've got time to do it but today most software is bought in a licence and user manner - you count the number of users and try to charge by that so that typically determines the size of the agreement.
That won't work in a network computing world so we are developing pricing economic models that are usage based as opposed to user based so that if you - all sorts of things change for us in this model. We have a sales force that pushes our technology in the market, educates the market and if they don't do a good job then most people don't hear about the software.
Now, it will be pull, a pull system where you can go on the Internet, you can go to conferences, you can participate electronically in some way to find out about Oracle software, you can pull it down on the Internet, you can get it, order it and have it arrive to you via CD and then you've got customers of yours that might use Oracle that we have no way of charging for - casual users that have Oracle embedded in an application that you write, so we think it could go one of two ways. We could go back to box base pricing because boxes have a predictable capacity, and say okay put as many users as you want on it.
You may allow the public at large to use it but you have a capacity that's built in by the technology of that box that has some problems as well, or you try to meter it in some fashion. That there is a metering software that understands the usage and every time the database is hit or every time the application is used. But I don't think anybody has worked this out yet, and so we've got a lot of people working on it.
Chess Krawczyk, ex Microsoft, current techno-nerd: One comment, one question. One comment is, split the government and corporate where they have 90 per cent penetration from the consumers and introduce the Intranet as a concept there.
Ray Lane: I'm sorry I didn't understand that. Split --
Chess Krawczyk: Your penetration figure of, you know, 30 per cent penetration in Australia and US, does not apply in government where there are more PCs than employees at the moment, or in corporate where they have one PC per employee and then talk about the consumers and end users. My question is , all of a sudden your network computer has a smart card PC MCIA slot which follows on from the previous user. Currently I buy a one use shrink wrap package and I use it, I play my CDs on, my cost of usage for my home PC is not 15,000, I have no expenditure once I own it and use it, now I have to put a PC MCIA smart card in, charge for use of software and everything else. Is there a difference between government which already owns the hardware and uses it and are you trying to create a new consumer market and generate more income?
Ray Lane: What a lot of questions. I agree that there is a big difference in the corporate and the consumer market place in terms of penetration and they are a very different economic model. I do believe that the NC will have a great appeal from a cost of ownership standpoint in corporations so even though everybody has a PC it doesn't make it any less complex or any less costly today. You can bring down the cost of the PC certainly as Microsoft has done by announcing now a Net PC so you can bring down the cost of a PC , but it offers great promise in corporations especially when you are on local area networks all the time to save a bunch of money.
Economics are not the greatest consideration in the home to a consumer. They are to a great percentage that haven't adopted PCs because they can't afford even that first purchase, and if you don't believe that you're spending money to maintain your PC then you ought to go back and look at your Visa card, because you are going to update your software, you are going to add things to it, you are going to download from the Network and all that stuff has cost to it. True, you don't have maybe the help desk issue that a corporation does but there is an economic question there but I don't think that will be the primary thing. It will be fact, the economic model for buying NCs might actually be like the cellular telephone model that you get them for free and if you want to take a very common basic service like E mail then you get it for free for signing up with the ISP, the Internet service provider that happens to be servicing you and for that service of $20 a month or $30 a month flat charge you get the NC for free, a basic service. And then they may want to extend options.
The smart card is only, is not a communicating device. The smart card only boots up the operating system and downloads that operating system to then put the NC online with the server. So it is basically a personal profile that you're carrying around with you. It does not allow you to add a lot of bells and whistles. The NC is a closed device. You don't open it up and add features to it. It is a closed device. It might be a closed special purpose device. It might go into automobiles, it might go into televisions, it may go into telephones.
I think we may see in the market within the next year telephony servers, servers that are switches essentially that go into corporations or governments that say well, here is a new telephone switching system. So your handset changes. Well as long as your handset is going to change and that telephony server has Internet capability and it's a server, take the computer off the desk top too and have one device, it happens to have a telephone, it happens to have a screen, served by a telephony server. So the integration of the desktop - why have the telephone and the computer in two different places? So I think the point you are making is there are different buying criteria in the consumer market and the corporate market and I totally agree with you.
Bill Trestrail, Silicon Graphics: Are Oracle moving into distributing their own NC manufacturing, what's their business model for doing so? It seems to be a very commodity driven hardware game that some of the electronics companies may be a very good distribution arm for. I just wondering if you have any comments on that aspect of it?
Ray Lane: No, we're not, we're a software company, we're not a hardware company so what we've done is created a subsidiary called NCI, Network Computer Inc, and they are building the operating system and the personal productivity software necessary to make the NC work. So certainly with the NC it's not good enough just to have. You could certainly run Microsoft Office from the server but if you want to run it independently, say run a word processor or a presentation graphics package you need something that is much smaller than Microsoft Office. So we've written a package called a Hatrick , it's our lab name called Hatrick, and it is basically 80 per cent of the functionality for 20 per cent of the code, you can't underline misspelled words six or eight different ways. You get one option. The Network computer like stands for no choice. You do it one way, and so it's a word processor, with an Office front end - it looks like Word, but you don't get all of these options which make Microsoft Office an 8 or 12 megabyte package. The operating system for the NC is 350 K. A lot of you look old enough to remember what K means. And Hatrick is a megabyte. In fact, we just introduced an Intel NC. It is not as fast or efficient as a Strongarm - we demonstrated a Strongarm in Brisbane yesterday. Strongarm 200 megahertz NC basically takes the power of two penlight batteries - very very cool operation, low voltage and very very powerful. You could see it in the application, you could see it when it was running, you could see the speed of this thing. Go over to the Intel and I think Intel will catch up in this game, but you get the first Intel NC, you get this big fan running in there and it's about a 130 - 150 megahertz but we're now in the game. You've got to have Intel in the game. Intel has got to be making these things and so they're in the game.
Ian Excell, Microsoft Canberra: You see the dedication with the name. You mentioned band width as the impediment to the growth of this concept. Where do you see that coming from and who's going to be the driving force to enhance the band width that we already have, especially in the environment that we face in Australia?
Ray Lane: Environment in Australia meaning the -
Ian Excell: The telephony services and the providers of that.
Ray Lane: Well, it will be the Telcos cable companies. It has to be them. They have the greatest opportunity for providing these services to consumers. In those countries where you have choice for a telephone provider or telephone service provider then you have a lot of competition that will go on for ISP Services and most every Aurbach and most every large PTT around the world has started an Internet service provider capability. There of course are a lot of Internet service providers that exist today that will not exist a year from today. So there will be a lot of consolidation in this when the Telcos get into it for real and so I think the Telcos are basically it. They obviously have to continue to invest in the infrastructure and that's going to be an economic game over the next five years as to whether they're willing to invest ahead of the revenue. But most of them that I talked to are aggressively moving into this and looking at hyperfibre coax technology, EDSL technology, ATM technology for local area networks.
Laurie Wilson: I'd just like to take the opportunity of asking the last question if I could, Ray. When I look around the room here today -
Serge Isragreysovet, Network Computers: We were all told that client servers would significantly reduce costs of ownership. They didn't. Isn't there a possibility that we are again setting ourselves up for a fall?
Ray Lane: Yes, I think so. I think like any technology it can be overhyped. Although quickly with client server, that was the early hype and quickly the industry backed off of it because it wasn't true. It wasn't a cost benefit, it was more of a flexibility benefit. And providing a lot of capability on the desktop, but it wasn't a cost benefit and it was not a simplicity benefit. So, yes, I think so but to me companies and institutions are getting much smarter at evaluating these things and they not going to simply say well fine NCs are everywhere now but I do think we'll see in 1997 a number of a leader companies that will say that rather than deploying 5,000 PCs we will deploy 3,000 PCs and 2,000 NCs for this application and start trying it or if they are using the right development tools they will deploy on PCs and then install NCs and the application will run just the same. With our development tools, called designer and developer 2,000, you can develop an application basically design a data model and then automatically generate that application code for a Windows client and then a UNIX back in, the client server application. From that same data model you can generate HTML for the Web and a thin client. All the prepackaged applications we sell for client server in February were automatically run on the thin client because we've introduced a thin client veneer and that's Gatebrowser that allows you to run it on any client and then run the application on the server not the client, where it traditionally runs. So now you get better manageability, better skill ability and lower cost and I think that will be a major major benefit for most companies.
Laurie Wilson: I will just conclude, and I guess ask a layman's question. I was fascinated by some of the figures we were talking about privately before and you might want to share those with the audience here. When I look around the room the number of people we've got here is fairly modest by comparison, for instance with the numbers you had in Brisbane yesterday, but it obviously represents a fair amount of clout in terms of this town, but I was particularly interested in your perception of how quickly we are moving towards NC and the level of interest and I guess what sparked my interest was the sort of figures you were talking about in terms of the turnouts you had in Japan, particularly.
Ray Lane: Yes, it's interesting what's happened in one year. Because one year ago , we had been working on the concept in the age of interactive television and developing set top boxes and it kind of gave us the thought about putting a very lightweight operating system in a thin client, use an Internet browser to go to three tier architecture rather than two tier architecture the client server represents and so we introduced that a year ago in Paris when Larry Ellerson made a speech that introduced the concept of the network computer, and I do give Scott credit in that they have been developing a JAVA based thin client for some time but this basically changed the game when Larry introduced it as a utilitarian device, a ubiquitous device that is low cost and simple to use and in that time frame it's moved from scepticism - I mean at one time Microsoft actually said (I don't know when you left Microsoft) but at that time Microsoft, Nathan Merebald said "They're lying. " Actually said it. It's in print. They're lying, it will not work it will never work. And one year later Microsoft introduces theirs to the market. So it's come that far. Clients have gone from trying to understand it, what does it mean, how does it change the client server game to now, "How do we deploy it" because they believe it will save money.
You know it's not a big change. All of the applications if you use the right tools that run in the client server will run in the Web. So it's a major change in the distribution of information but not in the fundamental technology. So this is the wonderful thing about it, is that you can implement it today. Okay, as soon as the NC is available but until it is use an 8 megabyte PC without a disk drive. See what Del says when you order one of those. But you can run it very effectively, it runs very very well. Don't want an operating system, I want an inter-based 8 megabyte no disk drive PC and it runs fine as an NC - we showed it yesterday. That's exactly what we were running yesterday.
So that now we've got to put it into volume shipments, so we've got to see NCs into the market place now in volume shipments and until we do we'll see lightweight PCs taking their place and so now I think it's coming, it's here, we believe in it , we will start with special purpose applications like customer service and just dedicate that application to an NC based, a network computing based application and we'll gain experience.
So 1997 you'll see real applications, you will see clients getting experience and then we've talked some of the Japanese partners we have, have estimated about 100 million units shipping in the year 2000. So this would be a tremendous growth in volumes. I am not an expert in how fast consumer adapted behaviour will take place and how fast we see these NCs go on the market place but I think we will see mass shipments of these in the next few years.
Certainly the excitement is unparalleled. We've never seen anything like it. We had about 1200 people in Brisbane yesterday come for the conference and see the NC operate. In San Francisco two weeks ago we introduced the network computing architecture with NCs. About 15,000 people showed up at the conference and in Japan 125,000 people showed up for the conference. It's a unique society. They get on bullet trains and they go with their bags and they come to see the new technology and when they decide they're going to adopt something, it works.
Our share of mind in Japan in six years has gone from zero to now 60 per cent. We have 60 per cent market share in Japan today and a very large subsidiary. It is the second largest databased business we have and it didn't exist six years ago. So second to the United States, just past Australia.
Laurie Wilson: I thought that Japanese figure might even stir the interest of the hardened professionals.
Ray, thank you very much. I'm not too sure if I should admit this, but my WC to PC index is about to go one one and I've got four toilets at home , so I don't know if that's a good or a bad thing but certainly in our house we've got plenty.
I'd like to thank you very much for coming along today. I can assure you that Michael Jackson, Tiger Woods and Bill Clinton won't be getting one of these. It's a special little gift from the Press Club and we treasure your appearance here and we hope you treasure this as well. Thank you very much Ray Lane.
See also:
Provided on the Web as a community service by the Australian Computer Society. Design by Tom Worthington.