Technology News Technology News

Another Perspective


The legal dispute between IBM and Platform Solutions, which offers IBM mainframe-compatible servers, moved to a new stage on March 8.  The battle began on November 29, when IBM sought to block the use of its software on its machines, largely on the basis of patent infringement and breach of contract claims.  Platform responded on January 29, asking the court to compel IBM to license its software for the platform.  Platform also alleged IBM's restrictive behavior was a breach of antitrust laws.  Now IBM has not only responded to Platform's countersuit, but also significantly broadened its legal case.

IBM's filing in the United States Court for the Southern District of New York, in White Plains responds, most often by simple denial, to Platform's defenses and refutes Platform's antitrust claims.  IBM also makes some new points, expanding the scope of its case and, should it succeed, setting out what amounts to new policies and practices governing the use of its software and patented technologies on any hardware made by others.

Basically, IBM says it does not have to license its mainframe software to makers of compatible mainframes and that from now on it will not do so.  It wants to get its code off the Platform Solutions systems that it previously licensed to use its software, in some cases under terms it says have lapsed or are no longer applicable, in other cases due to what it characterizes as mistakes made by IBM personnel.

With its most recent court filing, IBM is attempting to draw the curtain not only on Platform systems but also on X86-based Flex-ES mainframe emulators for which it formerly licensed its systems software and other products.  Additionally, if IBM prevails, there is no chance for legal use of the Hercules emulator with current or recent IBM software, nor for any other IBM-compatible system anyone might offer in the future.

IBM's legal position seems to extend beyond mainframes.  If ratified by the court, the posture IBM has taken in its lawsuit could be extended to other platforms, preventing competitors from offering Unix platforms or midrange machines that are compatible with IBM's two lines of Power-based systems.

This may seem farfetched, but it is a real issue.  When Apple moved from Power family chips to Intel, it used technology from Transitive that enabled some of its legacy code to cross the architecture gap.  Similar technology could extent enable X86 and X64 circuits to run code originally written for System p and System i.  Transitive vigorously promotes its migration products for moving applications from Sparc to X86 and Sparc to Itanium alongside a PowerPC-to-X86 offering.

IBM's Power server chips have moved beyond PowerPC architecture.  Transitive, which IBM has employed to help it support Linux on X86 application migration to System p isn't promoting technology to move applications the other way.  Transitive does not appear interested in migrating applications of its capital from the bank to the courtroom.

On the other hand, if Platform prevails, it will open doors for others, and not just doors into the IBM-compatible mainframe world.  It might provide the inspiration and the legal framework that gives birth to companies that compete with IBM's Power technology, or with all Sun's servers, much the way the advent of IBM-compatible mainframes played a role in the birth of the IBM PC clone industry.

This possibility, that the technologies epitomized by Transitive's offerings catch on and that competition in markets defined by processor designs might cross-architectural boundaries, is not lost on IBM or its lawyers.  PC clones didn't merely compete with the originals.  They drove IBM completely out of that market segment.  That's a disquieting notion for IBM, which relies on the exclusivity of its Power processor technology and for Sun, which depends on the Sparc family.  It's might even become an issue at Intel and HP, if it turned out that applications written for Itanium could be moved to X86.

The IBM reply, like its original complaint and Platform's response, doesn't begin with the big conceptual issues.  It follows the order of its original complaint.  The points that may reshape competition in the computer business pop up here and there in the latest filing, like raisins in a bun.

Thus, the initial sections of IBM's reply deal with points raised by Platform in its defense against IBM's patent infringement claims.  Platform argued that it did not infringe IBM patents, or, alternatively, if, by accident, it did, the infringements were of no meaningful consequence.  IBM insists this is not the case.  Platform also said that IBM's patents don't hold water; IBM reiterates that its patents are indeed valid.

In the matter of patents, Platform says IBM had in the past proclaimed that its policy was to license patents under reasonable terms.  Platform points out that IBM formerly had a page on its web site that spelled this out.  IBM agrees it did have that page on the web, but pulled it down on February 21, 2006.  IBM notes that its web has always said that IBM licensing practices were subject to change.

IBM's initial claims against Platform pressed for court orders to halt Platform's use of IBM software on grounds other than the patent violation, including a clause in its IBM Customer Agreement (ICA) that purports to bar "translation " as done by Platform's firmware.  Platform said this was nonsense: IBM was trying to redefine, or perhaps translate, that term on the fly.  But IBM has continued to assert it has a valid basis for this claim, although it did not offer an applicable definition of "translation."

One point made by IBM in its initial claim seems to have been significantly expanded in IBM's most recent court filing.  IBM says "it knows of no other company today that is developing new computers that are capable, lawfully, of running IBM operating systems developed by IBM for use on IBM's zSeries servers and their predecessors." A related item in IBM's filing points out that IBM did license software for Flex-ES emulators, which included patented IBM technology, but that this game is over.  Fundamental Software, which offers Flex, "had a patent license with IBM that expired on October 31, 2006."

IBM's antitrust defenses over the years, and in particular its successful effort to escape from the grip of a Consent Decree it signed in 1956 to end an antitrust suit brought by the United States government, are largely based on its claim that the relevant market for IBM mainframes, as far as antitrust law is concerned, is much broader than the market for compatible mainframes and includes other kinds of servers.  IBM successfully took a similar position in relation to its proprietary midrange systems, called AS/400 at the time the 1956 Consent Decree was wound down.

Platform's position is that IBM has huge market power in mainframes and, even if the court wanted to enlarge the definition of mainframe beyond IBM compatibility, IBM still has 85 percent of the market in systems with the defining characteristics of a mainframe.  The mainframe market, Platform more or less says, is quite distinct from the broader server market; nobody is going to get a thousand-dollar Dell server, or even a thousand of them, to replace a million dollar mainframe.

Platform and IBM are also at odds over what Platform calls illegal tying.  Tying, in antitrust law, means compelling a customer who wants a particular product to also buy another product.  IBM says that if you want z/OS you had better get a modern IBM mainframe to run it on.  While IBM has clearly and openly tied the products, its position is that the tying is not illegal, because it competes with other combinations of hardware and software that, for example, can run Cobol applications.  Fifty years ago IBM would not have had a chance to win this argument, but today it can rely on a body of case law that has limited the application of anti-tying statutes.

Antitrust remains a dynamic area of law, and one where the outcome of litigation can be very hard to predict.  It is not just a body of law that evolves in the wake of legal and philosophical argument.  It is shaped in part by what amounts to politics.

Ten years ago, the American fear of Japan, Inc., helped IBM win a reprieve from the 1956 Consent Decree.  Today, India is a source of political concern not because it might knock out American mainframes, as it was one thought the Japanese could do, but instead because it might knock out the jobs tied to mainframes using systems that come from IBM.

Platform's case does not argue that competition in the IBM mainframe segment will stem the flow of jobs to offshore services centers.  But the bigger picture, in which trends in virtualization and emulation could stimulate invention and thereby increase uncertainty in markets that are currently defined by legacy applications on IBM mainframes, is the elephant in the courtroom.  A more fluid market in computing favors American interests.  Platform's efforts to host mainframe workloads on alternative servers is an example of the battle between invention and entrenched market power that U.S.  antitrust laws are intended to govern in the interest of the American people.

IBM is not, generally speaking, regressive.  It's simply trying to protect its interests.  It wants to stem migration away from its products where it can, and finds that the courtroom rather than the marketplace currently gives it the best odds.  Its simultaneous use of Transitive technology to open pathways for X86-to-Power migration shows the company to be capable of progressive behavior when it happens to suit the particulars of a competitive situation.

If, in the end, IBM does not succeed in court through either defeat or, as is more likely, by settlement, it will then almost certainly show its progressive side to mainframe customers.  Mainframes could become less costly, and they could end up on hardware platforms that take advantage of Power family chips with their relatively high production volumes and relatively low cost compared to unique mainframe CPU circuits.

The legal case, as it has developed, makes it clear that IBM believes Platform has the right idea, or at least a very good idea.  It's just not an idea that IBM is yet prepared to contest in the marketplace, particularly if a competitor's lawyers write the rules of engagement.

— Hesh Wiener April 2007

Copyright © 1990-2019 Technology News of America Co Inc.  All rights reserved.