From the behemoth that once made and sold only proprietary systems, to one of the worlds keenest advocates of the open source movement, the IBM supertanker executed one of the biggest U-turns in the industry. Five years on from IBMs conversion to the cause of Linux, and 10 years after the advent of OS/2 as a popular (among techies) but failed operating system, we nabbed IBMs Worldwide Linux Software Strategy Manager Adam Jollans at Londons Linux User and Developer Expo, and asked him how Big Blue now sees the future of Linux and IBMs role in it.
Q: Whats your main role?
A: I have a worldwide role as Linux strategy manager. My job is to look across all our middleware - application servers and so on - and ask where customers want to use Linux, where are they likely to use it, and what we can do to encourage the ecosystem.
What weve seen is a combination of two main drivers of Linux, one business, one technology, operating simultaneously. The business driver is cost. As a result of the economic situation, enterprises have flat IT budgets and they want to save money. But IT managers are also interested in improving quality of service, reliability and security.
The biggest trend is choice. In the past, you went with one vendor but now, with Linux, you make a decision on a standard and choose who your hardware and software vendors are.
Q: Linux is free - how do you make a business out of it?
A: We dont sell Linux. IBM decided five years ago we werent going to develop an IBM version of Linux but instead decided to work with main vendors such as SuSE and Red Hat, selling hardware and services alongside Linux, just as Red Hat sells services and support. Enterprise customers want the same level of 24/7 support as with other systems, so the revenue is in the services area and has been a good business.
And its profitable. For example, 17% of mainframe MIPS are running Linux now - in fact its revived the mainframe, which is no longer seen as old technology. We invest much of that money back into open source development lab.
Q: Are there lessons to be drawn from the OS/2 debacle 10 years ago?
A: It was a great product but not the right time. The other main lesson was that its about the applications. There werent any. The issues with Linux are the same, which is why we want to encourage developers to write applications. We provide development tools for ISVs, help business partners with marketing and so on.
Thats going very well. Big ISVs such as PeopleSoft and SAP are all porting to Linux. Sage accounting in the middle space is doing so, and even smaller ISVs are porting to Linux because it provides a route for them to get to lots of people because its cross-platform. This means they dont have to invest in porting to lots of different systems.
Q: What about the technology? OS/2 was hard to configure, doesnt Linux suffer from the same problem?
A: Linux has been adopted on servers first rather then the desktop, and its been an easy migration from the Unix base in terms of both applications and systems administrators, since a lot of Linux administrators have a Unix background.
On the desktop, Linux is starting to happen. Device support is absolutely key and thats down to the dynamics of market share. OS/2 was an pretty much an IBM-only initiative, Linux is not just one or two companies but a whole community and IBM now understands that no one company can define standards.
Q: Does IBM see itself as driving the development if Linux?
A: Were very careful to ensure that we arent seen as driving the development of Linux. We want to help Linux develop and become more enterprise-ready because we think thats of benefit to our customers. Theres a group called the Linux development centre - 500 engineers and programmers round the world working on open source projects. We provide them with PCs, desks and a salary and they work on improving scalability and internationalisation. This means theyre acknowledged as being part of the open source community, which is important. If it were me, it wouldnt be IBM but Adam Jollans. This is great from their point of view because what drives them is peer recognition not money. Like universities, you make a contribution and get it recognised by your peers.
Q: So getting your name in the kernel code is the best thing you could do
A: Yes, Its very much a meritocracy. When we started looking at Linux back in 98 we asked how this crazy, anarchic development methodology ever be better than rows and rows of IBM programmers with rows and rows of IBM managers telling them what to do. We found it works very well because its not anarchy: theres a hierarchy since, to get to be a kernel hacker, you have to be good a it. If your code is poor, its on the Internet and others can see that. If its good, you get to write kernel code. Its unlike conventional methods where the only time you find out if a device driver is poor is when it keeps crashing your machine.
Q: Isnt control an issue for IBM?
A: No - in fact the reverse, we have Linux running on all our platforms and it works very well. Just like TCP/IP 15 years ago, I would have told you SNA was brilliant and better than TCP/IP because of this and that. Whats happened? The open standard won. So were keen to encourage that.
Q: Is Linux going to be the one OS for IBM?
A: There will never be a time when you have only one OS -- its always going to be a heterogeneous environment. Linux is the one thats available across all hardware but there are plenty of places where different OSes are in places and wont be easily displaced.
Q: Much has been made about the cost of supporting Linux compared to Windows. How will that be resolved?
A: Universities use Linux a lot and students can take the OS apart, find out how it works and its free. They then come out of university and have skills immediately applicable to the real world. Its just like Unix 20-30 years ago. There wont be a skills shortage soon.
Q: Is SCO helping the Linux movement?
A: Im limited by legal constraints but weve not seen the SCO copyright case affect anything.
Q: Are you indemnifying customers?
A: No, we dont think thats the right way forward?
Q: Youve invested $50 million in Novell recently. Was that a wise investment and has it borne fruit?
A: We see it as one of many investments in Novell and in the Linux community. We want to help Novell customers migrate from NT to Linux and this helps them do that. Its all part of our strategy.
Q: And the future?
A: Linux continues to surprise us. Each year we analyse where its going and it continues to disrupt the industry. This year, the new 2.6 kernel adds scalability up to eight and 16-way systems, so Linux will move into the sweet spot of Unix.