Question:
On our March 6th call, I was asked to provide more information on two points:
- Describe (to the extent I feel comfortable) Intel's criteria for deciding between RF and (F)RAND regimes.
- Provide further clarity on my statements for when Ex-Ante disclosure of licensing terms is appropriate and useful.
Background
During the call (and previously) I had noted that Intel and by implication, many other businesses, looked at standards licensing regimes on a case-by case basis. This is obviously different from an approach that all software or all software interoperability standards should be handled one way or another and the group asked for more explanation.
Point #1: It's NOT about royalty free versus "royalty bearing"
"As a technology leader, manufacturer and supplier to a variety of markets, Intel maintains a significant patent portfolio for mostly defensive purposes. We use our patents to protect our: Innovations (especially around our core products); customers and channels of distribution; ability to design and manufacture product. We also may use our patents to enable or expand markets."
Collecting royalties is simply not a priority for us in most cases. What we are looking for is "freedom of action" and we see this as a two way street - So long as you are not copying my products, I would prefer we not get into a patent infringement dispute. But in order to achieve that détente, we need to preserve the value of our patents and we need to be cognizant of the environment we are participating in.
For patents related to standards, a (F)RAND promise provides for a minimal erosion of patent value. There is some but hopefully the standard enables market growth in support of our business interests. We obviously would not commit to license even (F)RAND if enabling clone products or unnecessarily advantaging a competitor at little or no business value to us. If you do decide to claim patent infringement against us, our customers or our channels of distribution, you should expect us to use every mechanism at our disposal (including patents in standards) to defend ourselves.
Assuming healthy market forces (I know, I know - just bear with me for a second) a (F)RAND promise between implementers usually provides for the widest range of possible licensing negotiations and often does not necessarily even trigger a negotiation much less a royalty payment outcome. This isn't to say (F)RAND is perfect or that market forces can't be skewed. Just that (F)RAND is efficient and effective in most standards (note that this didn't expressly say "software standards").
A Royalty Free regime impacts the value of the patent even in a defensive sense and requires careful internal scrutiny before we join. We ask ourselves questions like: what is the potential effect on market growth? Is it in our business interest? Do we unnecessarily disadvantage ourselves in a future anticipated dispute/negotiation? What are our competitors doing? Are there key patent holders holding out for royalties? How would our decision impact our ability to deal with threats and/or strategies of other companies?
Don't forget, patent holders can claim infringement on many different aspects of your products (potentially essential patents from multiple standards, implementation decisions unrelated to standards, etc). At the end of the day, the question is "what is the best course on action in support of our products/business?"
The seemingly easy RF decisions:
So, would HTTP ever have succeeded if patent holders constantly sought royalties or triggered disputes? I think not. The market growth potential for HTTP far outweighed any potential royalties or the defensive value in that area. In another example, Intel clearly supported the W3C work on Voice Browsers because of the value to the market. We also can agree and even champion RF in some hardware standards (USB, SATA, etc). Our approach is pragmatic and quite often favors RF.
Should all Software Standards be RF?
I cannot say that unequivocally. Each standard must be evaluated in its own context. We certainly would not want to enable use of software supporting a proprietary product just because someone else wanted to call it a standard. SMEs would also find it extremely distressing if they could not protect their core business. As already stated in the commission report, OSS is often used in a hybrid form in support of proprietary goods or services.
There are multiple Open Source Licenses, often for a good reason. For example, we do see benefit to the GPL in some areas but the Apache License in others and potentially others as well. These are typically product decisions and one size does not fit all situations. Choice of a license is more of a product decision but I do see how standards play a role. In point of fact today, no international standard from the W3C, OASIS, the IETF can guarantee that patents will not be implicated.
There are ways for OSS solutions to implement standards from these groups.
What about Software Interoperability Standards?
I appreciate the desire to grow an ecosystem (especially in Europe) of SME software providers who may be able to leverage OSS as part of their business models. If given a particular context (e.g., the Linux Kernel) I can be very supportive but again, I find the blanket statement too vague to draw a universal conclusion. Note that each of the three words would need definition and context. For example, sometimes it is difficult to draw the distinction between software and hardware and often implementations can be made either way. Does this matter? Maybe yes, may no. It depends on the context. "Interoperability"
In whose eyes? Just because your competitor wants to interoperate with your proprietary product, should you loose your statutory rights without recourse or question? What is meant by standard? I just think that I can only intelligently discuss the merits of one licensing model versus another in a particular context.
A word of caution about facilitating business
Note that we are dealing in a global market and everyone needs to compete globally. Anything you do to advantage or disadvantage any particular business model can be use for and against you by others in the global market. I am not suggesting that the patent system cannot be improved or that real concerns don't exist but I just caution that any EC policy will also effect companies outside of the EU and/or other countries could adopt the same approaches. I would think the last thing we want to do is enable European SMEs to stay SMEs forever. That doesn't help Intel either. We want innovators to grow existing markets and make new markets that we can participate in. We do see great potential in OSS solutions.
My point on Ex-Ante Disclosure of Licensing Terms
Intel favors the prudent use of voluntary ex-ante disclosure of licensing terms as a mechanism to provide general information that may help in determining what technology to include in a standard.
It is a data point of limited value
Disclosure of licensing terms may provide some additional information but keep in mind that much is unknown during the drafting of a standard and even if you get a disclosure, it does not provide the "whole licensing picture" associated with implementing a product. You also might note that concepts like "what max royalty will you charge?" could be impractical early in the standards development process; divisive if your competitors game you to come out first and lower than they do; and problematic in a defensive strategy. Where this makes the most sense is when comparing to like technologies for inclusion in the standard or when faced with potentially including technology from a typically high priced source. Remember this is NOT your entire product licensing exposure.
You also need to avoid antitrust concerns. Intel strongly encouraged IEEE and ETSI to formalize their IPR policy around voluntary disclosure of Ex-Ante licensing terms. We also spent considerable effort in helping with the AT guidelines for these are other groups.
Note to be prudent in use
Generation and the analysis of such disclosures can consume considerable specialized resources, such requests should be carefully considered and applied under reasonable rules when appropriate. It takes time and energy to both generate the disclosure and to analyze the data. An SDO like IEEE and ETSI cannot afford the disruption or overhead of mandatory disclosures from every possible contributor every time the draft advances. Be prudent.
Statement of 11 Mar 2009
To be read in context with the suggestion (from the moderator) to have a communication platform that allows to coordinate strategies between actors and to think about a patent clearinghouse.
Regards platforms to discuss licensing terms ex-ante:
This actually has been tried formally at least once with mixed results (NGMN). The problems are how to have a meaningful conversation without triggering an antitrust problem and how to deal with the potential for gaming. These are legitimate concerns from any company's perspective (small, medium or large), and from a competition authority perspective. That said and as you point out, there may be merit to some high level discussion but we will need to be careful here. Any multilateral engagement on licensing terms is still relatively uncharted territory. As the industry gains more experience, we will be able to document some best known methods and increase our understanding of where these techniques could be useful.
Occasionally some discussions have taken place at the board level of a few organizations but usually with specialized legal counsel present and under strict rules. This has occasionally come up when competing technical proposals are accompanied by unique licensing proposals. I think you were actually making a broader point that even before an organization or work group forms, they might want to favor a particular licensing model or seek ways to minimize negative impact from IPR. This sometimes does happen when deciding between RF-RAND and (F)RAND regimes but I'll point you to the above paragraph for discussions about (F)RAND licenses.
Regards the clearing house idea
It has been proposed by the Chinese and by WIPO. The idea is that you could setup an arbitration body that could be both neutral and expert in patent valuation. There are some known problems with this approach. First, when people negotiate for a patent license, they are really negotiating a license for a product or group of products (not licenses to individual essential patent claims). In the real world, you will almost never find a single patent valuation in a vacuum. What the parties negotiate will depend on quantity; ramp; market size; geographical and field of use limitations; reciprocal product licensing needs from the other side; non-monetary exchange of value; cooperative research and marketing campaigns; existence of foundries; indemnification policies, infringement claims based on multiple standards, and non-standards related infringement claims, etc. It's rare to find a practical reason to ask an arbitrating body to rule on one small element. And don't forget the overriding effect of a court of law on validation or patent misuse or AT violations, etc. There are some SDOs that have offered arbitration resources to members in case of dispute over what is RAND. However, I don't believe they ever have been triggered. That may be a testament to their usefulness as a deterrent - neither one of you is going to like the answer you get if you have to come to me... So, the idea may have merit in particular situations but usually would be subject to a bigger question that would be decided in a court.