Over the last year, many CIOs have moved from the sidelines to the playing field in the search for a successor to IBM MVS, AS/400 O/S, Sun Solaris, HP/UX, and Microsoft Windows NT/2000 in the data centre. Based on recent announcements and rollouts, that successor might just turn out to be Linux—the one OS that will run on all today’s hardware.
Let’s look at the recent uptake in Linux adoption and examine how it could affect decisions to use this upstart OS in your enterprise.
The hits just keep on coming
Linux is a natural choice for enterprises, such as retail systems and cable networks, with potentially thousands of limited function devices. In May, Sherwin Williams, one of the largest paint manufacturers and distributors in North America, announced that IBM had assisted it in creating a new retail sales system for its 2,500 stores. It will be one of the largest rollouts of a Linux-based system to date.
In addition to handling its point-of-sale operations, the system will run the company’s colour-match software and let employees communicate using Linux-based e-mail and browser software. Sherwin Williams will save millions of dollars in software licensing fees—funds it can funnel toward systems development and customer service. Running the incredibly efficient Linux on less powerful—and less expensive—user terminals at the endpoints of its network will save additional dollars that drop directly to the bottom line.
Linux is also gaining momentum in financial services, a traditional UNIX stronghold. Both Credit Suisse First Boston and Merrill Lynch have converted major systems to Linux. And don’t think that Microsoft is the only one affected by the Linux revolution. In the financial services arena, Sun has a big bulls-eye on its back. As more software companies provide UNIX product versions, financial services companies are discovering that they can provide the same services on commodity Intel PCs using Linux instead of the Sun SparcStations and Sun operating systems they’ve been buying for years.
Winners and losers
As more markets warm to Linux, there will be repercussions across the tech industry. If Linux adoption were to accelerate rapidly, some companies are in a position to lose big, while some others could become the big winners.
The losers: Sun, then Microsoft
Most Linux lovers are Microsoft bashers. But the reality is that Linux hurts Sun much more in the short term.
Linux adoption is a no-brainer for companies who can run the same software on inexpensive Intel machines with Linux as they can on expensive Sun machines running Solaris. Microsoft doesn’t have a significant presence in the retail market anyway, which has been dominated by fixed-function machines from companies like NCR and IBM. These are the devices that have been targeted by manufacturers using Linux as an embedded OS.
But in the long run, Linux may dramatically lower Microsoft’s chances of taking over the data centre with Windows 2000, the .NET Server, and its .NET Framework. That’s because there are so many companies that would benefit from having a data-centre OS standard that nobody owns and requires no licensing fees.
The winners: IBM, Dell, and software makers
Who would benefit from such an open source standard? The big winner would be IBM, which has made a billion-dollar investment in Linux, which is just beginning to pay off. IBM's hardware, consulting services, and software groups all benefit from the adoption of Linux in the data centre.
In the server and workstation (not desktop PC) market, companies like Dell and HP/Compaq have the chance to supplant thousands of existing servers and workstations with commodity machines running Linux. Of course, software manufacturers have to make products available on Linux/Intel platforms, and that appears to be happening.
Today’s CIOs are more likely to consider using Linux as part of their infrastructure. I attribute this, in great measure, to the investment that hardware and software manufacturers have made to get their products ready to run with Linux.
An ironic fate for Linux?
Adopters of Linux-based data centres will also have to deal with a key strategic issue. The novelty of working on an OS as a public service will ultimately wear off. When the open source fever dies down, one of two things is likely to happen: First, the companies benefiting most from its existence–-processor manufacturers, hardware OEMs, IBM, set top manufacturers, and software makers—will have to fund a new company with a profit motive for enhancing the operating system, if Linux is to grow.
The second likely scenario is that these same companies will create their own enhanced, incompatible Linux versions in order to differentiate themselves. Linux could become like UNIX in the 80s. And the need for a unified platform on which to build large volume applications will drive people toward their only commercially viable option—Microsoft Windows.