Getting Inventive With Software Patents
The dangers of software patents for free software have always been a hot issue. But with the news that Red Hat and Novell are being sued for alleged patent infringement by IP Innovation, the matter has moved from theory into practice. In fact, in the battle against software patents, it turns out that the open source world already has a rather powerful weapon in its armoury – even if it's one that few people know about.
The Open Invention Network (OIN) is a company set up in 2005 with investments from IBM, NEC, Novell, Philips, Red Hat and Sony. It has a war-chest of tens of millions of dollars (it won't give an exact figure) that enables it to buy software patents on the open market for the defence of the GNU/Linux ecosystem, which it does in a very novel way. OIN's growing collection of patents is available royalty-free to any company, institution or individual – provided that the latter agrees not to assert its own patents against GNU/Linux, defined very broadly.
A couple weeks ago I had the opportunity to talk to OIN's CEO, Jerry Rosenthal, who was with IBM for 37 years before setting up OIN. His comments about patent trolls – companies whose business is based purely around suing others for alleged patent infringement, rather than on making and selling things – and how to deal with them, are particularly interesting in the light of what has happened since the interview took place.
GM: Was there a particular trigger for the creation of OIN?
JR: I don't know that it was a trigger; I think it was more a series of events. We saw the problem existed; we sat down and talked about the problem and had to address the problem in the environment that exists. As you know, there is this question of whether or not there should or shouldn't be software patents. That's a problem that's just going to take time to play out.
The thing we needed to deal with [was]: how do we deal with it in today's environment in the traditional sense of [patent] cross-licences? Once we got to say, "Well, how do two companies deal?", then we got to the point of well, we've got this problem because Linux is not a traditional company. It's sort of, how can we make Linux into, in essence, a more traditional company? And that's why I use the term that we are sort of the surrogate patent portfolio company for Linux.
GM: It seems slightly strange that IBM, with a highly lucrative patent portfolio, should be supporting this scheme, which tries to reduce their impact: what's the thinking behind it?
JR: Although IBM is a huge proprietary company, remember: now they have a huge investment in Linux and open source. It is in IBM's, and Sony's and Philips' best interest to see them succeed - all these other companies embed Linux in their products. It was simple to say, therefore, how do we make sure Linux survives?
GM: How important are the patents that you own?
JR: I think they are very important. Certainly, the first portfolio we bought we know is valuable because we had to bid at an open auction to buy it. We spent $15 million: that says there were a lot of other people who recognize the value. We just were able to have enough money to outbid them. I have been involved in several auctions since then where I have spent significant amounts of money to buy the patents I buy.
GM: Will you be adding further to that portfolio?
JR: Absolutely: we continue to buy patents.
GM: When somebody takes out an OIN license, do they increase your patent pool?
JR: All they have agreed is, they will not use their own patents against Linux. In the sense of creating an ecosystem of patents that won't be used against Linux, they are increasing the pool.
GM: So there are two classes of patents: the ones the OIN has bought and owns itself, and the ones that are part of that larger pool?
JR: Right. So, for instance, the IBMs and Philips of the world have said their entire patent portfolio will not be used against other OIN licensees for Linux purposes.
GM: Can licensees withdraw?
JR: They can, at any time, agree they will no longer participate going forward. In other words, whatever patents they have got up to this point, they have agreed they will not use against Linux. Their future patents, they can do what they will with. The same thing, from our patent perspective: they can use all of our patents up to the date that they decide they want to terminate, but they don't have access to any of our future patents.
GM: Recently you announced that Google became the first end-user licensee: what does that mean?
JR: It's the same licence agreement. But if you think for a minute, most of our licensees are the distributors of either Linux or other open source software, or in the business. Google is just the user. The message I was trying to get out was, we welcome users [as licensees] as well as people who either distribute or embed Linux, or write software around Linux.
GM: You use "Linux" in a very broad sense: you don't just mean the kernel, you include many other programs – how many altogether?
JR: A thousand modules! When we started the company we tried to define Linux, and we realized you can't define Linux. And so basically it's what might be in a typical distribution that makes it usable.
GM: Why don't you extend your protection to any OSI-licensed program?
JR: We'll extend it from time to time. The problem you have is you don't know which code is corrupted; which code people really developed on their own; which, unfortunately, maybe they took from somebody that they shouldn't. So we try to be very careful, to make sure that the stuff we have has been vetted, is clean.
GM: If somebody tried to sue someone in the GNU/Linux ecosystem for alleged patent infringement, what would you do?
JR: I would take it to our board, which is made up of the six investors, and say: "Is this the right case for us to go in?" What we don't want is to get sidetracked in lots of little cases and spend millions of dollars in research because somebody is trying to pick off one of us.
And so we would look and say: Is this the case that really is the seminal case that's going to define whether or not the patents of the competitor are good? Have they brought suit with their good patents? Is the suit really about Linux? Or is the suit maybe a tangential suit, where they have an issue with the XYZ Company for reasons we don't know - it's really a copyright suit, it's not a patent suit, it's a trademark suit? We want to make sure it's the right suit.
GM: Assuming it is, what happens then?
JR: I can't defend, because as you know, in a court of law, company A sues company B, I'm company C. So I would have to bring an independent lawsuit with our patents against company A.
GM: So you would find something in your portfolio that you think company A has infringed on?
JR: And then I'd bring suit.
GM: Supposing it's Microsoft that is suing someone in the GNU/Linux ecosystem; how does the agreement with Novell change its rights?
JR: Doesn't. That has nothing to do with it. Nothing. Novell, on their website, has made myriad comments about their relationship with OIN has nothing to do with their deal with Microsoft. And Microsoft has made it clear they have nothing to do with it.
GM: Typically, patent trolls don't have any products, so they are unlikely to be infringing on any of your patents. Isn't that a problem for the OIN approach?
JR: Very clearly there's not much we can do with regard to patent trolls. On the other hand, it's my belief that patent trolls go where the money is. They're after the big dollars. That's probably why they haven't, and not for a long time will, go after the open-source movement - that's not where the money is. Could it become a problem in the future? The answer is, sure it could.
Hopefully, the laws and the courts are catching up with the trolls and what they're doing. And that problem, hopefully over the next few years, will start to take care of itself. The fact of the matter is, the trolls have relied very heavily on injunctive relief and threatening. That's pretty much been taken care of by the courts. They've also relied on patents that are not good, that everybody has said are not good; and recent rulings on obviousness issues are going to begin to do it. [There's] this new effort with the US Patent Office on peer-to-peer review, where they're going to provide databases so that the Patent Office can do a better job.
It'll never go away - problems never go 100 per cent away. But I think we are going to start to see it diminish. And certainly, when I'm looking to buy patents, if I see a patent that's available for sale, that perhaps does [affect] Linux, and we're worried that a troll might buy it, I might just take it off the table. Remember, trolls don't own any patents, they go out and buy patents on the open market. Well, I'm buying patents on the open market, so I might take some patents away.
Glyn Moody writes about open source at opendotdotdot.
- High-Availability Storage with HA-LVM
- DNSMasq, the Pint-Sized Super Dæmon!
- Localhost DNS Cache
- Days Between Dates: the Counting
- Real-Time Rogue Wireless Access Point Detection with the Raspberry Pi
- You're the Boss with UBOS
- The Usability of GNOME
- Linux for Astronomers
- Multitenant Sites
- PostgreSQL, the NoSQL Database