IBM’s Linux Initiative

It's funny what associations particular times of the year can bring to mind. For me, one of the things the time around Thanksgiving has come to mean is the formation of the Internet Division ten years ago -- a time I recently wrote about.  And the December holidays and New Year period, among all their other meanings, remind me of IBM's embrace of Linux.  It was over this time that we organized this groundbreaking effort six years ago.

In retrospect, it's remarkable how rapidly this all coalesced. Personally, I first started hearing seriously about Linux from people in the supercomputing community early in 1999, although I was aware that some Linux activities were already underway in our labs in IBM.  Universities and research institutions were using clusters of Intel processors running Linux as a way of building relatively inexpensive and increasingly powerful supercomputers.  So were fields like seismic processing and life sciences, whose applications were well suited to parallel Linux clusters.  At the same time, Linux-based systems were becoming popular in all kinds of Internet infrastructure applications, such as web servers, file and print servers, and network firewalls. 

The word about Linux was getting around. I specifically remember a mid-1999 meeting of PITAC, the President's Information Technology Advisory Committee, for which I was serving as co-chair.  A number of the members, highly respected leaders from industry and universities, felt that we should initiate a study on the importance of Linux and open source software in general to the advancement of high-performance computing.  We launched the study later in the year and formally published our report in September 2000.

By the summer of 1999 Linux was picking up steam in the marketplace. Around that time, we launched two major studies in IBM, under the auspices of our corporate technology committee: one focused on the use of Linux in supercomputing, the other on Linux as a high-volume platform for Internet applications and application development.  Both studies strongly recommended that IBM embrace Linux across its product lines, that we work closely with the open Linux community as a partner in the development of Linux, and that we establish an IBM-wide organization to coordinate Linux activities across the company.  Ultimately the recommendations went to Sam Palmisano, current IBM Chairman and CEO, who was then head of IBM's Enterprise Systems Group.

The Monday before Christmas 1999, I was in Boston attending a meeting when I got a call from John M. Thompson.  At the time I was general manager of the Internet Division reporting to John, who led our Software Group.  John told me that Sam was ready to implement the recommendations to embrace Linux across IBM and wanted me to lead this new Linux initiative reporting to him.  The following day I talked to Sam from New York City, where I was meeting with customers.  Sam wanted to make the announcements as soon as possible.  Since the Internet was by then so integrated into all the activities of the company, we agreed that we no longer needed a formal Internet Division, as we had for the last four years. I would continue to oversee our “Next Generation Internet" programs from my new position, as well as a number of advanced systems technology programs.

This was actually a rather remarkable turning point.  On the one hand, it showed how thoroughly IBM had embraced the Internet -- the Internet and everything around it was just standard operating procedure for IBM now.  But perhaps even more striking was the way this former icon of proprietary ownership  -- the company that still led the world in patent creation, as we continue to do to this day -- had rapidly recognized the business potential of the open, collaborative  approach to technology development and innovation.  By late 1999, there really wasn't a lot of debate on the matter within IBM. It was obvious that this was important, that it could prove extremely valuable to our clients and to the world at large, and that we should take a leadership position in helping it succeed.

We had to put together the new organization, as well as make plans for announcing our new Linux strategy to the world over the next couple of weeks, an especially challenging task since many people were out on vacation the last two weeks in December.  We reached people at home and started putting things together, and then continued the work when we got back to the office in January.  On Monday, January 10, 2000 we announced IBM's embrace of Linux, as well as the formation of the new Linux organization.

Our announcement got a mixed reception.  Many welcomed our strong support of Linux and open source communities.  But in January 2000 Linux was still not all that well known in the commercial marketplace.  A number of people viewed Linux as just another operating system developed by a bunch of "hackers" with nothing better to do, and were totally perplexed that IBM was so aggressively supporting an initiative that, in their opinion, was so removed from the IT mainstream. 

Over the next year, I spent quite a bit of time explaining why we were supporting Linux.  On February 3, 2000, I gave a keynote presentation at the LinuxWorld Conference in New York, saying that we did not view Linux as just another operating system any more than we viewed the Internet as just another network when we announced our Internet initiative four years earlier.  We viewed Linux very much as part of the evolution toward open standards to help integrate systems, applications and information over the Internet.  Linux was then and continues to be the only popular operating system that runs on every single platform regardless of vendor or architecture, a property it shares with just about all major software associated with the Internet.

Also like the Internet four years earlier, we were convinced that Linux was going to be a success in the marketplace.  That was the message we were hearing from the world's technical community, as well as from our own technical people in IBM who had thoroughly studied Linux and some of whom were themselves part of that community.  Even though the software ran only on relatively smaller systems at the time, Linux was already attracting a very strong following of the best and brightest programmers and computer scientists around the world, and we could tell that it was only a matter of time before it got significantly better.  I believe that when community efforts like Linux are successful and keep attracting top talent to work on them, the marketplace is sending a strong signal that companies should not ignore.  If Linux or other such initiatives are going to be a marketplace success with or without you, you are much better off being part of the success than fighting it or ignoring it.

Finally, I also believe that our Linux initiative and our relationship with the Linux community have had a major impact on the IBM culture.  It has made us a more open company, more ready to collaborate closely with others in open communities and to share our intellectual property with them when appropriate.  I am convinced that this kind of collaborative culture is required for businesses that want to be innovation leaders in the 21st century.  When I look back, I feel very good about those times six years ago when we launched our Linux initiative.
   

Comments

[Vinnie Mirchandani]

Irving, you say this has influenced the whole IBM company. But your global services group (over half your revenues now)is definitely in the "cathedral" school of contolled tech development and deployement v/s the "bazaar" approach of Linux and Open source. Would your systems inetgrator group use the open source community for testing for example? would they pass the lowered economics on to clients?

The tech services world needs radical transfomration as I have written in Frederick Taylor and Technology services (see link below) ...learning from the Linux community would be great for them...

http://dealarchitect.typepad.com/deal_architect/2006/01/frederick_taylo.html

[Nicholas Beale]

I agree your embrace of Linux is stratgically very good and important. In our small strategy consultancy we've been using Linux for years but it's still frustratingly difficult to get decent support. My desktop has been Linux-only for several years but my Laptop dual-boots, mostly to Windows 'cos there are still apps that don't work properly under Linux. As soon as this playing field is levelled there will be a massive improvement in productive power/performance.

[Tony Smith]

My bIG problem...

I like linux. SuSE, Debian (not so much RedHat) I use them a lot.
I want to bring Linux into my company. I could create an "ActiveDirectory like" setup all integrated with file sharing, remote assistance and everything I need to make life good and easy.
I have about 600 users in my company and at least 550 of them should find everything they need in Linux.

But how can I switch operating systems when my users don't know how to use Linux. I can't imagine the frustration to take Windows away from them.
And assuming we get over this part and everything is working... a new employee comes aboard... again training, frustration... this would just go on...

My thought is that until Linux breaks into every home it will be difficult to bring it into the business environment. A few years back the biggest problem was compatibility between Linux and Windows. Now you can integrate the two but getting users to actually use it.... is next to impossible regardless of how much management support you have.

Regards,
Tony

[Kiruthik]

"Now you can integrate the two but getting users to actually use it.... is next to impossible regardless of how much management support you have"
--I agree Tony's Word

Its just people, who has to understand and think.

I struggled to bring people from MS Office to Star Office once I was IT Mgr. I failed at last. Just I succedded in implementing it in internet sharing PC for its native Firewall support (IP Tables).

[Robert Pogson]

Linux has nowhere to go but up and IBM, having jumped in early, will get a big share of the business. This is smart and visionary. I am absolutely amazed that there are still organizations that do not have this vision. Last year, I worked for an outfit that, in response to a dozen reasons why they should switch to Linux on the desktop, could only respond "we are used to Windows" as a reason to stay with Windows. They did absolutely nothing that Linux could not have done better with much lower costs. The previous year, I worked for an outfit that upgraded to XP for five times what it would have cost to upgrade to Linux. In both cases, the ongoing operating costs would have been slashed. IMHO, the choice of Linux is one of the great separators of businesses. Depending only on what fraction of budgets goes to IT, the Linux choice will be a major indicator of an efficient operation. The agile upstarts will have an advantage over more established firms that have more invested in the old ways.

[Chris Ward]

One way to learn is to make some 'Linux-for-Windows' screensavers, like I did for developerWorks here
http://www-128.ibm.com/developerworks/linux/library/l-scrnsave/?ca=dgr-lnxw09ScreenSaver

Another way is to get one of the 'Live Linux' CDs or DVDs (or USB Storage Keys); find an old PC that is about to be thrown in the trash; and try using it. Or a new PC; the 'Microsoft Windows' hologram says that you have the right to use Microsoft Windows on it; that doesn't mean you have the obligation. Novell/SuSE, an IBM partner, make one available here http://en.opensuse.org/Welcome_to_openSUSE.org .
I've had success with that one, and with others; but of course that's no guarantee that anyone else will.

If you want commercial support, you may contract with IBM Global Services; or with anyone else you see fit. If you're a school or charity, you could see if there are any IBMers willing to help out in their spare time.

Linux and applications are actually very good at being standards-compliant. They have to be, for there to be any hope of others in the wider community understanding what they do, determining whether a behaviour is a 'bug' or not, and being able to maintain and build on them. So 'Bazaar' it may be, but it's usually fairly easy to get a 'Cathedral' application to run on it.

Im my opinion, Sony's Playstation 3 running Linux will be a most interesting possibility for a 'desktop workstation' in the not too distant future. http://www-128.ibm.com/developerworks/power/cell/index.html
For 'Playstation 3', IBM is making its top-of-the-line C++ compiler (xlC) available at no charge. Of course, if you want a guarantee that it will support your business, then signing a contract with IBM Global Services would be a very good idea; but if you just want a compiler, it's there.