Monday, March 14, 2011

Red Hat feeds the patent trolls and fools the FOSS community

I have dug up a court filing that reveals one of the best-kept secrets in the open source industry: Red Hat coughed up $4.2 million in June 2008 to settle a patent infringement suit that had been brought by FireStar Software.

In a world in which hundreds of thousands of software-related inventions have been patented, chances are that Red Hat has concluded more transactions of the FireStar kind. Red Hat settled another patent lawsuit (with Acacia subsidiary Software Tree LLC) in October 2010 and most likely made a payment (which might have been higher than in the FireStar case). The vast majority of all license deals are negotiated privately without litigation or announcements, so Red Hat may have quietly signed patent license agreements with a number of other right holders and never even talked about it. For example, just in connection with the acquisition of JBoss, Red Hat set aside $43 million for the settlement of claims that included (but were not limited to) the FireStar lawsuit.

While patent license agreements are just a routine deal for traditional IT companies, the significance of this finding lies in how it reflects on Red Hat's credibility:

  • In political and regulatory contexts, the market-leading Linux distributor insists on the false claim that open source licenses (especially the GPL) are absolutely incompatible with the payment of patent royalties (while this simply depends on the specific terms and conditions of a deal).

  • The GPL also serves as Red Hat's pretext for refusing to offer its customers implementations of patent-encumbered standards such as the popular AVC/H.264 video codec.

  • Another important credibility issue is that the world's largest open source company actually fails to be open about its patent-related dealings and positions, a fact that others have criticized before. It hides behind self-imposed confidentiality obligations and employs legalese terminology that in the eyes of many laypeople will appear far more favorable to Red Hat than a straightforward representation of the facts would.

I take the revelation of Red Hat's patent royalty payment as a confirmation that the open source industry actually accepts software patents as a fact of life and opposes them only as lip service to curry favor with the FOSS community. As much as I regret that, at this stage I just try to describe what's actually going on in the industry and to see through smokescreens.

Red Hat is a publicly traded company trying to make quarterly and annual numbers, not an open source project, let alone an activist group. Red Hat's shareholders support the community's causes only to the extent they consider such efforts economically advantageous from a very short-term perspective. If they can achieve the desired goodwill with make-belief at a lower cost, then that's what they'll do. Don't let them fool you. To them, this is just business.

They also put plain business interests above openness in connection with their hidden Linux kernel patches.

The payment was kept confidential but surfaced in a different lawsuit

In 2006, FireStar Software filed a lawsuit with the US District Court for the Eastern District of Texas, alleging the infringement of a patent related to linked databases (US Patent No. 6,101,502) by Hibernate, a JBoss product. When FireStar first approached JBoss, the company was in the process of being acquired by Red Hat, a fact that FireStar may have regarded as a financial opportunity.

The patent-in-suit was later assigned to a Texas company named Datatern. An investment group named Amphion Innovations US, which owns a stake in FireStar and set up its wholly-owned subsidiary DataTern as a patent licensing company (i.e., a non-practicing entity), also became a party to the suit.

The settlement related to any or all patents held by those companies that might be infringed by Red Hat's products. Red Hat published a redacted version of the settlement agreement and a "reader's guide" to it. the only part that had been redacted was Section 3, "PAYMENT". There were different references to that payment in the publicly available parts of the agreement (and the parties agreed to pay their own legal fees), but the amount was not stated anywhere else. Section 7 stipulates a confidentiality obligation concerning the section on payment.

Later the plaintiffs sued the law firm that had represented them against Red Hat for alleged malpractice. In connection with that case (case no. 2:09-cv-00038, Eastern District of Texas), the parties initially didn't mention the amount paid by Red Hat, but then an additional group of companies came in: IP Navigation Group and affiliated entities. Those companies belong to Erich Spangenberg. who according to runs one of the "largest, and most litigious, patent-holding companies" and recommends a "sue first, ask questions later" approach. The article quotes patent defense company PatentFreedom, according to which "entities connected to Spangenberg have sued more than 500 [by now even more than 600] companies for patent infringement since 2005". This aggressive patent monetizer became involved with the FireStar/Red Hat dispute when the right holders were unhappy about the state of affairs and brought him in as additional firepower to pressure Red Hat into a settlement.

On October 1, 2009, Foley & Lardner (the firm that represented FireStar), amended its "counterclaims, third-party complaint, and cross-claims", and paragraph 11 of that document made the following revelation:

"Of the $4.2 million settlement that the Plutus Parties [Spangenberg] obtained on behalf of Plaintiffs [Amphion/Datatern], the Plutus Parties took $3.4 million."

Statements made in such a court filing must have evidentiary support, so we can rely on the accuracy of that information.

GPL compliance of patent royalty payment

In an effort to appease the FOSS community, in which there is a widespread belief that the GPL doesn't allow inbound patent licensing, Red Hat emphasized in its announcement that its settlement also covers "community members" (upstream and downstream) as "third party beneficiaries". Indeed, Section 9.12 of the settlement agreement contained some provisions. In the press release announcing the deal, Red Hat's patent attorney Richard Fontana said:

"Red Hat's settlement satisfies the most stringent patent provisions in open source licenses, is consistent with the letter and spirit of all versions of the GPL and provides patent safety for developers, distributors and users of open source software"

Eben Moglen of the Software Freedom Law Center reviewed the agreement and concluded that it would also be compatible with GPLv3 (although only GPLv2 was relevant to that particular deal). He liked the fact that its terms "provide additional protection to other members of the community upstream and downstream from Red Hat" and welcomed "Red Hat's efforts on the community's behalf."

Considering a $4.2 million payment compatible with the GPL is, however, in contradiction with what Red Hat and its lobbying fronts claim in different political and regulatory contexts. For example, when the European Commission was updating its European Interoperability Framework (a set of procurement guidelines), Red Hat and its different lobby groups claimed that only royalty-free standards are compatible with the GPL and other important FOSS licenses. I debunked that false claim and instead advocated a proper distinction between patent licensing terms and conditions that work for FOSS from those that don't. That is exactly the approach the EU took in the final version of those guidelines.

I believe Red Hat should have told European policymakers the truth about its FireStar settlement and unknown number of other patent license deals in place instead of trying to gain a political advantage with incorrect claims that royalties are inherently GPL-incompatible.

Patent royalties and the debate on video codecs

In the EU debate I mentioned before, Red Hat used MPEG's royalty-bearing codecs as an example of patent-encumbered statements that it claims it cannot implement because of the GPL. However, if Red Hat can pay $4.2 million to FireStar (and who knows how much to who knows how many other patent holders), why can't it pay a royalty to MPEG LA for an AVC/H.264 license? The maximum payment in the MPEG LA structure for such a license would be $6.5 million.

I looked into those terms last year, and the maximum amount has meanwhile been increased from $5 million to $6.5 million. The current amount was stated by Ed Bott on his ZDNet blog.

Where there's a will, there's a way -- and in this case I believe Red Hat simply doesn't want a solution. Unfortunately, this limits choice among codecs for its customers, for no good reason. Red Hat simply wants to maximize its profitability at the expense of innovation and progress.

Google argues that its WebM/VP8 codec is needed in order to have a video codec standard that is implementable in open source. It's uncertain whether Google's offering is indeed patent-unencumbered and therefore won’t be subject to royalties at some time down the road.

MPEG is already working on a royalty-free codec of its own. In my opinion, the open source industry should give AVC/H.264 serious consideration despite royalties. The price to be paid is limited, and one of the benefits is that it's a multi-vendor standard while WebM/VP8 is just one company's offering, even if seemingly available on open source terms.

Red Hat isn't forthcoming

Red Hat uses the word "open" a lot, but in connection with patent licensing it isn't really transparent. A few months ago, Red Hat settled litigation with an Acacia subsidiary named Software Tree LLC. Previously, Red Hat had fended off infringement allegations by a different Acacia company, IP Innovation LLC, because it was not found by a jury to infringe any valid patents. But Software Tree LLC asserted patents that had previously survived invalidation attempts and were considered reasonably strong by a Texas-based patent lawyer I asked.

Red Hat didn't announce any terms of the deal with Software Tree LLC, and its complete failure to be forthcoming had many people concerned. You can read some good questions and well-reasoned criticism from Open Source Initiative co-founder Bruce Perens on GigaOM and on other sides including (also picked up by LinuxWeeklyNews), Australia's iTWire, and techdirt.

The most likely explanation is that the deal was worse than the FireStar settlement, especially as far as the interests of the wider open source community are concerned. If Red Hat could have made the same claim about protection of other open source developers (upstream and downstream) as in the FireStar case, I guess it would have done so.

Even in the FireStar case, Red Hat should at least have made it clearer that it ended up paying significant royalties. Not huge, but definitely significant.

Richard Stallman, the president of the Free Software Foundation, also voiced a caveat concerning the FireStar announcement:

"If we can judge from Red Hat's statement, the deal is good for the free software community. I would not want to treat that as certain; they might have chosen not to mention some negative side."

Again, that related to Firestar. But I wonder what RMS thinks of Red Hat's refusal to disclose any information at all concerning its settlement with Software Tree LLC.

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents.

Share with other professionals via LinkedIn: