Hewlett-Packard has the task of migrating its Unix customers running HP-UX on PA-RISC, as well as recently acquired Tru64 Alpha customers from Compaq, over to Intel's Itanium architecture. Making sure this transition happens smoothly in Australia is Steve Williamson who transferred from Compaq and now is business development manager in HP's business critical systems division.
Q: How big is the HP-UX install base in Australia?We have 4000 to 5000 customers in Australia and they range from big data centre organisations like AMP and BHP right through to small end users.
Q: Give us an idea of the spread of applications running on HP-UX on PA-RISC in Australia.You can look at it a couple of ways. One of the things is ISVs take HP9000 (PA-RISC) as Tier 1, so basically every ISV develops for HP-UX which is from my perspective quite refreshing having come from a Tru64 background where that wasn't quite the case.
So yes, we have Oracle, SAP, PeopleSoft - nothing that you wouldn't expect on an NT platform or Sun platform. Where the differences are, I guess, is where we're progressing, where we are spending our money. There is going to be a lot of consolidation in the industry, so a lot of organisations are going to get together and some [others] will disappear.
We are putting our resources behind organisations like BEA, Oracle and SAP. We are not going to put them behind other database vendors. If you look at the overall database market it is SQL, Oracle and DB2. They are the ones we want to put our resources behind.
I don't think HP has ever had a problem with applications. HP-UX has always been one that has had lots of them.
Q: What would the split be between packaged applications and in-house or outsourcing-developed applications?If you look at the overall base, about 80 per cent [of customers] buy packages and modify them. Very few write their own and that is generally consistent with both our Alpha and PA-RISC platform. It is almost not the done thing any more to develop your own.
Q: Give us an outline of what an organisation would have to do to transition from HP-UX on PA-RISC to HP-UX on Itanium?Nothing. If you develop a 64 bit Linux application and take those binaries and copy them across to HP-UX, they will run. If you have a PA-RISC application today, copy it across to Itanium/HP-UX it will run. There is no difference. It just copies. It 100 per cent runs!
What we are trying to do with Tru64 (Alpha) for example, is the same thing. It is not much of an event.
On Itanium we have five operating systems, but we believe everybody should develop on Linux and then they can transport across to different platforms. Linux is the easiest, most portable platform to develop on. We are not proposing to develop on HP-UX. I mean you can develop on HP-UX and put it on Linux, but it is a cheaper proposition for a customer to buy a cheap Linux platform, do their development and then deploy it on whatever platform you want - HP-UX or whatever.
What we are saying is develop on one and you can deploy on any platform really. Linux is a key strategy for that. It basically opens the window for everybody from Oracle all the way down to the local ISV to be able to develop that code cheaper than on a proprietary platform. It is a no brainer really.
Q: What happens if your customers decide to stay with HP-UX on PA-RISC in terms of support and availability of hardware?We'll support them through to 2011 for PA-RISC or Alpha. They will be able to buy new products till 2006 and there will be a marketing organisation that will sell them refurbished equipment. We are not saying to customers "go out and change tomorrow". We are saying '"we are giving you a 10-year window at the very minimum" which allows them to decide when they want to transition and move across. We are not saying "right, next week it is all Itanium and that's it".
The value proposition is going to be so great that people will say "I am going to switch over."
Q: So why do you think there has been a slow take-up of Itanium?I don't think it hasn't taken off. Now is the time for ISVs to start investing again. But they are not doing that. But on the hardware side, Dell, IBM and HP have all committed to it. It is a product line that is going to go forward. The ISVs have the challenge. They are now being told Itanium is the go. We have Microsoft on side, PeopleSoft, BEA, Oracle and really that is where we want to invest our money, because as we invest our money in those guys the rest will come along anyway.
I don't think there is a problem. The last list I saw [had] about 700 applications that have been ported across. Bear in mind we have three streams. There is the .Net stream, the Linux stream and the HP-UX stream, so you have layers upon layers of different organisations.
It is not that there has not been take-up, it is just that the job has been so big. Once it is complete it will be so much easier for customers. They will only have one platform to worry about.
Q: Has it been difficult educating HP customers of Itanium?Yes, because we have the flat earth society out there. If you go to the people who thought the earth was flat they believe there is a paradigm and you have got customers who believe this is true. "This can't possibly be true. You have this Itanium stuff and I have heard it all before and you can't demonstrate it and it will never happen," they say. So they believe the earth was flat and you drop off the edge. This is the same with our competitors too. They want the flat earth society to keep existing, because they don't want people to wake up and smell the roses. People are just not taking it up as well as we would like them too. But in the long term it is a winner.
Q: What will be the role of RISC in the future?Niche markets. If you look at HP it realised in the mid-1990s that it couldn't afford to develop the chips. Dell has not even tried to. Intel is the ultimate winner. It has 90 per cent of the market share.