One of CCIA’s smaller members, TurboHercules SAS, has received a lot of attention recently surrounding its antitrust dispute with IBM, especially because IBM has chosen to rattle its many patent sabers at this much smaller competitor by providing a list of 173 potentially infringed patents. What makes this story a little different from a classic patent shakedown is that TurboHercules is an open source company, and IBM is a champion of open source software. In early April of this year it was reported that IBM accused TurboHercules of “potentially” violating two of the patents in IBM’s “legally binding” Open Source Patent Pledge from 2005. In its pledge, IBM promised the world that it would not assert 500 selected patents (out of IBM’s portfolio of over 40,000) against open source projects. IBM then issued a series of conflicting statements regarding the two pledged patents (see here and here), including one that suggested that the pledge was limited to “qualified” companies that did not compete with IBM. Needless to say, this left many in the Open Source community scratching their heads. The story also highlighted many of the problems endemic to the current patent system.
Why TurboHercules? Isn’t IBM famous as a friend of open source? Didn’t IBM play a hand in setting up the Open Invention Network as a portfolio to defend Linux and related open source projects?
The problem is that TurboHercules is a software emulator that allows standard computers with common operating systems (Linux, Windows, Solaris or MacOS) to run software designed for IBM mainframes and IBM operating systems, old and new. In other words, TurboHercules promises to inject some competition into the stagnant mainframe world (especially on the hardware side). The mainframe business accounts for 40% of IBM’s annual profits
, thanks to IBM’s near 100% share of the mainframe market, while the expense and uncertainty of even the simplest migration efforts enables IBM to keep old customers locked in
Ironically, IBM is doing exactly what the Open Invention Network was set up to guard against: using patents to suppress competition. Successful intimidation does not require filing an infringement suit. The uncertainty, complexity and expense of actually defending oneself from a patent attack from a major corporation – especially when you are a small startup – is simply unaffordable. The knowledge that it might happen can be paralyzing, not only for the startup but for its investors, business partners, and customers as well.
There has been much discussion about the threat patent trolls pose to established companies
. But this matter illustrates another problem our broken patent system foists upon our economy. Large corporations and monopolists can – under the banner of “protecting their intellectual property” – thwart disruptively innovative startups that threaten their profit streams. Because these small companies have little or no patent portfolios of their own, they have little to counter the huge portfolios of incumbents.
This is not the first time that IBM has thrown its weight around in the mainframe arena. IBM first accused Platform Solutions, Inc. (PSI) of infringing
its patents before eventually purchasing them
– a threat that may well have helped IBM drive down the purchase price (and avoid automatic antitrust review because the deal fell below the Hart-Scott-Rodino
threshold). Perhaps not so surprisingly, IBM’s patent attack came when PSI was allegedly close to a deal with HP, IBM’s main competitor. The patent threat killed the deal – a deal that would have put this disruptive technology in the hands of a competitor who had the ability (and the patent portfolio) to compete aggressively with IBM.
Other companies and their customers have also faced threats from IBM’s patents, including T3 Technologies
. These companies have offered IBM’s locked-in customers more options, lower costs, and interoperability between IBM’s operating systems and other hardware. If this tale tells us anything, it is that even “good guys” can’t be trusted when business interests encourage them to exploit patent thickets to block competition and innovation.
Certainly a legislative fix is needed to better align the patent system with its original purpose of promoting innovation, not protecting legacy profit streams. Although Congress has debated patent reform legislation for the past five years, proposed legislation offers little to address the underlying problem: we have two different patent systems. One system serves pharmaceuticals and chemicals reasonably well, where patents protect products and offer useful information about what is the state of the art. However, in the other system, Information Technology products comprise thousands or tens of thousands of patentable components and functions. It is very costly to determine what functions are in fact patented, who owns the patents, the validity of those patents, and whether a particular implementation infringes the patent. This creates a paradox: patents are normally less valuable in IT because the market demands fully functional packages, not individual processes. To make a useful product, all the patents must be assembled, but there are many more patents in IT than in pharmaceuticals (where they are most valuable). And just like printing more dollars doesn’t make everyone richer, the proliferation of patents in a given product doesn’t make that product any more valuable. On the contrary, the individual patents become less valuable. In spite of this, however, patents remain highly valuable for holding competitors’ products hostage.
This misalignment of incentives and the subsequent patent gridlock it creates hits small innovators the hardest – that is the ones that actually bring their inventions to the market (like TurboHercules). They get exposed to patent lawsuits because they have a product (or products) in the market, but they can’t rely on the mutually assured destruction defense because they don’t have large portfolios of their own. This dynamic threatens the very heart of the tech sector where the concept of small, disruptive companies challenging the status quo is almost axiomatic.
Despite the risks, entrenched companies with very large portfolios, such as IBM, may feel that a vast enough collection of patent assets (even those of dubious validity) outweighs the potential liabilities. Much depends on the company’s ability to manage both sides of the equation strategically, for example:
- Economies of scale in legal resources. Small companies generally have to pay top dollar to defend themselves by securing legal services from outside law firms at retail rates. Companies with substantial resources and a reputation for scorched-earth defenses can deter trolls that are looking for quick and easy settlements.
- Whether the company is actively innovating for the market. Innovation for the market creates exposure in two senses: Creating a new product brings the risk that someone will have thought of a particular element and will patent it first, or already has. Once on the market, its functionality will be exposed for all to see.
The unhealthy combination of large numbers of patents, astronomical costs, high degrees of uncertainty and risk, and massive potential liability is bound to have an adverse economic effect. It induces small companies to stay upstream – or to sell themselves to large companies who are much better positioned to manage risk. Even for large companies, it creates incentives to withdraw from, or at least not aggressively innovate in, some markets. This not only reduces the risk of liability in absolute terms, it allows the company to assert patents aggressively with minimum exposure to counterclaims. Indeed, IBM no longer makes PCs and has largely withdrawn from the mass market for software.
IBM’s inclusion of two of its pledged patents should perhaps be viewed as a blunder by its famously aggressive legal department. However, IBM dug itself in even deeper by suggesting that the pledge it made was limited to non-competitors.
The real problem is not the two patents but the 171 that were not subject to the pledge. In effect, IBM claims that there can be no competition in the mainframe space because IBM holds a thicket of patents that preclude interoperation with its legacy-based software and hardware without IBM’s permission. From the patent holder’s perspective, the nice thing about portfolios and thickets is 1) they make legal challenges prohibitively expensive and 2) they can be “evergreened” so that as long as the portfolio owner keeps making minor changes and adding new patents, the portfolio can exclude competitors forever – instead of the 20-year term of individual patents. These portfolio practices in IT effectively turn the basic idea of the patent system on its head. Instead of the individual patent serving as protection for the little guy, portfolios serve the big guys by allowing them to maintain market share by creating massive legal roadblocks, justified in the name of “intellectual property.”
An unfortunate consequence of the IBM – TurboHercules situation is the fact that one of Open Source Software’s biggest defenders, IBM, has lost the moral high ground in the rhetorical war and likely has gone a long way to validate more recent bad behavior, including Microsoft’s unfortunate threats againstAndroid and its prior threats against Linux and Linux users (a situation that bears an uncanny resemblance to the current IBM-TurboHercules kerfuffle).
Prior to the IBM/Hercules case, not only had IBM been an outspoken and effective supporter of open source at many levels, but its employees had contributed to the underlying Hercules software that TurboHercules uses. The fact that IBM contributed and supported this effort in the past, only to turn on it later as a commercial version emerged, raises substantial questions within the open source community about whether IBM is in fact a trustworthy and committed partner.