May 1 ’09
Linux on System z: “Let the Sunshine In …”
Contemplating the old tune “Aquarius/Let the Sunshine In” from The Fifth Dimension somehow adds to the surreal nature of the idea of Sun as a potential part of IBM. If the deal proposed a few days before this column was written actually proceeds to a happy conclusion, it has two interesting points that may affect how the Linux world can proceed in the enterprise space. The tune proposes that universal enlightenment will arrive through love and understanding, but the quintessential message is that we need to accept what is, not what the surface appears to show us.
The love and understanding bit is clearly something we could gain from the merger of these two technology giants. Sun has a number of interesting technologies that render the acquisition as an obvious advantage to IBM. Two powerful examples are the formidable dynamic tracing tool and language DTrace and the sophisticated storage management capabilities of ZFS, a mélange of file system, storage manager, and physical device management for all types of online and offline storage media. Sun has been a leader in enterprise software technology for many years, and the ubiquity (if arguably not so valuable religious wars caused by it) of Java as a method of expressing CPU architecture-free implementation of program code and programming environments is a clear example of the kind of intelligence and foresight possible when you try to think differently about how programming should work, and invest the kind of money and resources into the necessary development and evangelism to make it happen. We also (for better or worse) are the result of several generations of engineering and science graduates who know nothing but the Solaris environment and Sun hardware as the way to perform computing. These two items—mind share and technology— could be arguments in themselves for IBM to save Sun. It will be up to U.S. anti-trust regulators to determine whether the deal can continue, given the market power such a deal could give IBM.
Where sharing the love intersects with Linux is how IBM could treat that intellectual capital. Many of these technical advances have been locked up with obscure licensing agreements that were incompatible with the GNU Public License that governs the Linux code base; you couldn’t use them, or ideas from them, in GPL code. In recent years, IBM has provided more of its ideas and intellectual property in GPL-compatible forms. IBM (in a PR move worthy of the Aquarian ideal) could choose to release those basic technologies in that way—a staggering advance over more primitive volume and storage management techniques such as Logical Virtual Manager (LVM). I’d like very much to see how IBM, or others, could leverage that.
That said, there may be a shadow under all this cheery metaphor. In addition to their technical work, Sun also has been a fairly consistent contributor to the open source community. The merger of the two companies may cause that contribution by the commercial sector of Linux and other open source projects to be drastically reduced—two givers to one, and one less inclined to share due to the substantial price required to redeem Sun from the consequences of the dot.com bubble and its inability to reposition and re-invent itself in the hardware space.
Those of you who have attended my OpenSolaris for System z presentations have heard me ask the question of how Sun can maintain itself in the enterprise space without a substantial improvement in its hardware to match its capabilities in the software engineering and design space. You’ve also heard me say that I feel IBM has always excelled in the hardware space and failed in the software space. This merger offers an opportunity to have the best of both.
P.S. Check out the latest release of OpenSolaris for System z!