IBM has invested heavily in the R&D required to bring Linux to the iSeries/i5 and has made Linux the only common operating system thread between all of its eServer offerings. But how important is Linux to the i5? Is its implementation on the i5 critical to your future on the platform? And what about the importance of Linux to your management's plans? Does the outcome of the SCO lawsuit against IBM really matter to the future technology directions of your company? Or is Linux just one more piece of marketing collateral for the computing platform? Finally, what is IBM's "end game" with Linux on the i5? Will it end up being just another piece of technology that IBM uses to hook customers to its proprietary operating system? Or does it really matter?
These are the questions that all of us are asking, and the depth of feelings within the AS/400, iSeries, and i5 community are complex, subtle, and filled with anxiety.
Let's be serious for a moment and look beyond the hype. How important is Linux to the iSeries and i5 community at this time? Let's look at what we're doing with the iSeries today. The answer is very sobering.
- RPG: Most AS/400 and iSeries boxes in use today run a mixture of legacy RPG code and ILE RPG, with a smattering of Java, COBOL, and C applications that were ported from other boxes.
- Packages: Most of these boxes started with some form of package software that has been modified or customized over time.
- Financial/ERP: A large number of boxes are running J.D. Edwards ERP or Lawson Financials, BPCS, or other accounting software--often significantly customized or modified.
- Manufacturing and Distribution: The largest segment of the iSeries install base is in manufacturing and distribution, where new applications are often dictated by the largest members of an industry supply chain.
- Management: The average CEO doesn't give a whit about the operating system deployed within the box.
- Business Importance: The decision to keep or jettison the iSeries architecture has nothing to do with its worth as a development platform or its ease of use for IT management. That particular decision has more to do with the productivity of the people using the box, the requirements of the business, and the cost (and/or pain) of sustaining the entire iSeries environment.
- Heterogeneous: Most companies running iSeries boxes are supporting at least one--but possibly more than one--other operating system environment. Those other environments were deployed for their own reasons--good or bad--and probably include some product from the Microsoft Windows Server family (NT, 2000, XP).
- Integration: Relatively few iSeries boxes currently deployed are taking advantage of the full iSeries integration capabilities to support UNIX, Linux, or Windows. The reasons are complex, but they usually represent a combination of cost, division of IT labor resources, legacy management decisions, and--most importantly--management and IT exposure to the iSeries/i5 capabilities.
So how important is Linux to these organizations where the iSeries is deployed today? The answer is simple: Not very!
Status Quo Vs. Future
But that doesn't necessarily mean this status quo will persist over the next five years. And trying to guess where the future will lead technology--and where business requirements will force technology to go--is a key ingredient to measuring the impact of Linux on the i5 platform.
Today, from the user's perspective, the value of an operating system (whether it be OS/400, i5/OS, Windows Server, UNIX, or Linux) is like the value of a single automobile engine in the corporate fleet of cars. After all, operating systems are the engines of corporate productivity: Just as with car engines, very few people worry about how these machines actually work. They're more concerned about getting to their destinations. They only worry about them when they malfunction or when the price of operating them goes up. So, in a way, using a "price per gallon" analogy is how we might also see the issue of an operating system's future. If the cost of supplying it with software goes up relative to other platforms, the ROI goes down.
When IBM has looked into its crystal ball for the iSeries platform for the last five years--as it has for all its eServer models--one thing has become very clear: The cost of all software for every proprietary operating system has continued to steadily climb, and this trend shows little inclination to abate.
The Rising Cost of Software
The cost of developing any proprietary application for any operating system will always proportionally expand to consume as many resources and as much money as are available to purchase it. If there's only $500 available, that's what it will cost! If there's $5 million available, it will cost that too! There are exceptions, of course, but they are few and far between.
Even Microsoft products--which once seemed like veritable bargains--have continued to rise in price to meet demand. One seldom ever sees a price cut with the introduction of a new software product release. It doesn't make sense, and it's a secret cardinal rule of the industry that no one but an idealist or a lunatic would invest hundreds of thousands of dollars to build an application for which there is no market. Why would they? What would be the reason? Who would be so foolish? Isn't this the case?
Indeed, the problem with the iSeries platform has been exactly this problem: Fewer new applications are being built for the RPG iSeries market, their costs are proportionately growing, and consequently fewer new customers are opting onto the platform. Yes, it's true that IBM's past eServer strategy has been to build more functionality into the hardware. But CEOs don't select computing platforms; they buy solutions. Fewer solutions mean fewer sales of boxes. Fewer boxes in the marketplace mean fewer developers writing code.
Is this an irreversible trend? Well, yes! And no!
Enter the Free Software Movement and Linux!
The Free Software Movement has been questioning exactly this proposition since the 1980s, and in today's marketplace organizations like GNU (which stands for "GNU's Not UNIX") and The Open Group are seriously impacting how the software industry views this trend. Of most importance to this discussion is the GNU General Purpose License (GPL), under which the Linux operating system is distributed.
The GNU GPL is a license that allows authors of software to issue a license to their code for legal protection. This license restricts the liability of the author from lawsuits resulting from errors or flaws in the software. However, the most important aspect of the GPL is its requirement that both the resulting object software and the source code that created it be freely distributed to encourage others to modify the code as required. These modifications--called "derivative works"--must be freely distributed as well.
In fact, all derivative works that result from the GPL licensed work must also follow the requirements of the GPL. In other words, the GPL license is a anti-copyright copyright: It makes code available for modification and free distribution, but it retains ownership of the intellectual property for the author. It doesn't prevent a distributor from making money through the distribution of the code, if that is the organization's business model, but it does legally remove the code from the legal and economic classification of proprietary ownership.
The Impact of Linux Under the GPL
When Linus Torvalds released his Linux operating system kernel under the GNU GPL, he gave hundreds of thousands of developers and users free access to his work. Some organizations, like Red Hat and others (including SCO), began redistributing the operating system--along with the source code--with their own utilities and applications. They, in turn, started making money on the bundling of their non-derivative works that used the Linux kernel. Then, a couple of years ago, IBM joined the movement and began modifying the Linux kernel to work with its entire line of eServer computers, including the iSeries eServer.
Each company that participates in the redistribution of Linux has a business plan that has an implicit "value add" statement. In IBM's case, the value add is IBM's hardware platforms and IBM's middleware products (DB2, Domino, WebSphere, Tivoli, and Rational).
So, how does this odd, totally nontraditional method of market distribution of an operating system like Linux lower the cost of software? In four ways:
- It has allowed Linux to quickly become an industry standard.
- It has enabled more developers to contribute to the development, evolution, and maintenance of the operating system platform.
- It has unleashed the power of independent developers to create new applications for the platform at the lowest possible cost.
- It has enabled customers of the platform to have the broadest possible access to the fastest-growing base of applications.
In a word, then, it allows the creation of a new software market--often called a bazaar--for developers to hawk their wares and customers to get products. And this is exactly what the GNU GLP release of the Linux operating system has done. It has bootstrapped the Linux kernel from being the hobby of a student into an engine of commerce.
The Impact to the iSeries/i5 Platform
So how does this, in all practicality, impact the IBM i5 platform? After all, if only a small portion of the iSeries/i5 community really cares about other platform issues, why would including Linux on the platform make any difference?
The problem that Linux solves for the iSeries/i5 is simple: It opens the platform to new applications that--in all probability--would otherwise never be ported to the i5. More applications in the Linux bazaar will, potentially, lower the costs of obtaining those applications for the i5. This in turn should attract management's attention to the i5 as a more cost-effective platform upon which to consolidate its applications. More consolidation means more cost-savings for the owners. So Linux offers the IBM i5 the opportunity to break that economic cardinal rule of software investment.
So how realistic is this theoretical marketing model in today's iSeries/i5 world? That's a question that IBM and its eServer Group management team are trying to orchestrate! It's also the same tune that IBM's Software Group is whistling. IBM's support of the Linux dynamic is an effort to recreate the software/hardware market segment by providing a tremendous value add statement in an expanding bazaar of offerings.
Why the SCO Lawsuit Matters
But what if somebody owns the very territory upon which the Linux bazaar is being held? That's the crux of SCO's $5 billion lawsuit against IBM. SCO claims that, through a series of acquisitions, it has the ownership of the copyright of the UNIX operating system and that IBM has misappropriated software code from its separate license of UNIX and released that software code under the GNU GPL as a derivative work of its Linux R&D. SCO has threatened to sue companies that use Linux that do not also enter into an agreement with SCO. In essence, SCO is saying, "If you use Linux, you are actually using UNIX! So you owe us money because we own the bazaar!"
Regardless of the merits of the SCO lawsuit against IBM--and legal experts are divided over those merits--SCO's claim of ownership represents the single largest threat to IBM's i5 and its Linux strategy. If the courts find in SCO's favor, it represents merely the nose of the camel under the flap of the tent. Despite SCO's press releases to the contrary, many legal analysts believe that if SCO wins against IBM, it will move against the validity of the GNU GPL, claiming that Linux is a de facto derivative work of the UNIX operating system.
Linux Matters to the i5
So, regardless of where your personal stand is on the value of Linux to the iSeries/i5 community--its technical value to the platform, its potential value to your particular organization, or the merits of the SCO lawsuit against IBM--Linux does matter to you and your company!
It matters to the future of the iSeries/i5, to the cost of software applications that run on it, and to the cost of software for the industry as a whole. The age of proprietary operating systems will probably never end, but it's clear that IBM plans to mutate those that it owns to meet the demands of commerce. IBM believes that the company with the best strategy for transforming itself to meet those demands will succeed in the race for dominance in the marketplace.
IBM is helping to build a new market segment in which software is more standardized, more open for development, and cheaper to purchase and maintain. Linux is the operating system platform that is IBM's current venue where it hopes to achieve this goal. With it, IBM hopes to transform the software market. Without it, IBM will undoubtedly find another set of tent poles and try again to build its bazaar somewhere else.
In the meantime, we on the IBM iSeries/i5 platform are sitting with ringside seats. It will be one of the most interesting industry shows of all time.
Thomas M. Stockwell is Editor in Chief of MC Press Online, LP.