There’s quite a bit in the press these days about companies (surprisingly, some very large ones) aggressively investing to expand their IP portfolios by purchasing patents or filing for patents on anything that can be imagined – often without stopping to consider whether the “innovation” has utility and is truly novel and non-obvious. Sun is often approached by companies looking to purchase patents (a reflection of the value of our IP) and at times we do sell patents under the right terms, conditions and circumstances.
To some degree, this topic has a very Cold War feel to it with companies growing patent stockpiles to use if attacked or as a form of “mutual deterrence”. But, at some point, a company needs to ask how many patents it really needs. And, that’s exactly what we did about three years ago. Up to that time Sun was filing well over 1,000 patent applications per year. But, in 2005, we made the decision to reduce our patent filings to the point that we had about 700 patents issued last year. And this number may decline in the future. While this is still a sizable number for most companies, it is a significant decline for Sun and occurs during a period in which we have more innovation than at any point in Sun’s history.
Why the change? Part of the reason is financial. On average, it costs more than $20,000 to obtain a U.S. patent and this figure grows significantly when you file around the world. Also, this amount does not include annual annuities required to keep a patent in effect. Being selective in what you patent can result in significant savings. However, the bigger reason for the change is that our focus has shifted from quantity to quality. To this end, we have completely re-architected the manner in which we determine the innovations we will patent. As part of this process, inventions are reviewed by a panel of the chief technology officers from across our different lines of businesses with input from distinguished engineers and other experienced innovators. We apply a significant amount of scrutiny to determine whether something is truly innovative before we submit it to the PTO. For us, it doesn’t make sense to patent everything. Rather, our focus is on patents that represent significant technological innovation. (In this regard, we were happy to see that the Federal Circuit will reconsider the patentability of business methods. )
Aside from our focus on patent quality, there is another reason we are filing fewer patents. It has to do with our business model. Unlike some companies, we don’t have a corporate goal for revenue derived from patents (and patent litigation). Instead, we invest in patents to support our customers and the communities in which we participate. This support can be in the form of a defensive response to an attack on a community or in the form of the assurance provided by the patent licensing provisions of the CDDL or GPLv3. In the end, it’s about delivering innovation to our customers and communities.
If they succeed, we succeed.
7 responses to “The Patent Arms Race”
Hi Mike, Thanks for sharing your views on the patent arms race and the links you provided. I am intrigued by stuff under the Ocean Tomo patents/IP auction link. I might have misread this but taking one of the old patent (issued in 1991) in Lot 19a for example together with the provided "Sample Forward Citation Analysis" (including companies like HP, IBM, Canon), it almost reads like an invitation to buy the patent to avoid being sued or buy it to sue some of these companies. Am I really off in this view?
At the same time, for a company of your size the cost of a patent is minimal compared to an engineer’s salary. Even if you filed one patent per engineer per year, that would be a huge number of patents but would be sound protection given that the engineer probably did at least one thing novel and worth protecting over the course of a year of R&D. Actual rates of filing are much lower than that. If you look at a company like IBM that is filing probably 5,000 patents per year, this is only $100 million total which is such a small part of their annual revenue.
I agree though that quantity should take a back seat to quality. If a company is filing just to get their numbers up and not with a specific protective angle in mind for each application, then it is wasted money.
I’m glad to hear that Sun is focussed on quality and value. There are too many large corporations that look only to the number of issued patents as a metric for value.
May I suggest that a good metric for a reasonable return on IP might be the benchmark set by IBM licensing over a half-decade period from the late 1990s to the early 2000s? Over this long period of time, IBM saw about 20% returns from its investments in R&D through technology licensing revenue (i.e., both patent and know-how). This figure does not include the contribution that R&D had to increases in IBM’s equity value, which I believe more than accounted for the remaining 80%. What it reflects is the value that IBM developed R&D had to OTHER companies that were entering markets that IBM was not interested in pursuing.
IBM has scale advantages that many small businesses do not. Nonetheless, that 20% return seems like a fair one for a patent owner to expect from third-party users of a technology that the patent owner was first to disclose to the public.
Can you provide an update on the NetApp ZFS lawsuit, and your countersuits against NetApp? I really hope Sun prevails because a loss of ZFS would seriously harm Sun’s open storage ambitions. For example, what are the court dates? How long with the re-exam of NetApp’s patents take the USPTO? Is there a chance of out-of-court settlement?
The patent system has been designed by the patent industry and then applied to the software industry. Everybody knows in our industry that 20 years of monopoly around algorithms implemented in the form of software running on a computer is silly as it is not related to the rate of innovation.
It is time for the software industry to fight back and get rid of this parasite not designed for the need of fast and incremental innovation.
Mason Boswell says "…for a company of your size the cost of a patent is minimal compared to an engineer’s salary. Even if you filed one patent per engineer per year, that would be a huge number of patents…"
True, but this ignores the engineers’ time working on the patents. In my experience, even with a patent lawyer writing the actual words, the time required for the engineer to describe the invention to a review board, describe it again to the lawyer and check and correct the patent application (and I find it hard to understand my own inventions when they are in patentese, let alone anyone else’s) can be enormous. Once the patent is in the system, engineer time is required at intervals to answer the question of whether the patent should be renewed.
I assume that Sun has both hardware and software patents since Sun produces both hardware and software. In the article Mike Dillon makes no distinction between software patents and hardware patents except in the last paragraph where here refers to the patent terms in some software licenses.
Does Sun make different policy decisions with hardware patents and software patents? There is a lot of discussion about software patents generally being of poor quality as compared to other types of patents. When Sun decided to file fewer, higher quality patent applications did this change reduce the number of software patent applications more than it reduced the number of hardware patent applications?