IETF Patent Policy: A Quantum Approach

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.

The following defined terms are used: a set of principles and practices intended to promote access to the design and production of goods and knowledge through full disclosure of all relevant design materials. a software or hardware implementation that is licensed on terms compatible with Open Source principles. An Open Source implementation that is provided at no charge to any party and which may be extended or adapted without restriction. a set of exclusive rights granted by a state to a patentee for a fixed period of time in exchange for a disclosure of an invention. terms that allows a party to practice the invention covered by a patent without risk of a lawsuit. a contract granting a Patent Access. Reasonable and Non Discriminatory terms Reasonable, Non Discriminatory and Zero cost terms Reasonable and Non Discriminatory terms in which a royalty fee is required for some or all uses.. rights granted by one party to another on the condition that the recipient grant terms that are equally favorable to the granting party. Certain definitions are adapted from text found in Wikipedia.

A Patent Rights Holder may seek revenue in the form of license fees. The terms offered for these rights may or may not be RAND. In many instances a Patent Rights Holder seeking to obtain license revenue will offer RANDZ terms for certain types of use (e.g. non-commercial, educational) or certain uses of the patent (e.g. allow free use of rights to decode content but charge fees for encoders). Such terms are properly regarded as a business model rather than Zero Cost terms.

A Patent Rights Holder may seek to use patent rights to establish or extend a dominant or otherwised advantageous market position. As with a revenue exploitation model the Patent Rights Holder may offer liberal terms in some areas while protecting others.

As issue of Patent grants has increased Implementors and Users, particularly those with deep pockets have become increasingly concerned about their exposure to litigation and damages risks. Defensive patent applications provide a degree of mitigation for this risk. A defensive patent application establishes prior art for the invention(s) claimed. This MAY pre-empt subsequent grant of a patent for the same claims by another party or allow for a successful defense should a subequent patent be issued. This is not certain however. A defensive patent applicatiopn MAY be used to obtain reciprocal rights from other patent rights holders.

A Patent Rights Holder may be subeject to limitations of their rights or lose them entirely for certain specific types of inequiable conduct. This has occurred in the US courts for failure to disclose the existence of rights to an invention when the Patent Rights Holder was under a specific obligation to do so.

A Patent Rights Holder that has invested substantial resources in developing an invention may desire credit for doing so.

The value of a communication protocol is dependent on adoption. As Nicholas Negroponte observed, the utility of the first fax machine was nil until the second one was produce and has climbed steadily as the number of faxes has increased. The need for widespread adoption is the reason that standards play an unusually critical role in the development of communication technologies. Without standardization of features such as thread pitch, nuts and bolts would be considerably more expensive. Without standardization a communications protocol is no use whatsoever. Without adoption, the value of a typical communications invention is zero. It is thus in the interests of Patent Rights Holders to seek adoption of the inventions they have rights to by promoting them in standards process.

The cost of defending a patent infringement lawsuit can be enormous. Litigation risk is a particular concern in the US where costs cannot be recovered from a losing plaintif and cases routinely cost defendants with 'deep pockets' $2 million or more.

Damages in patent lawsuits can be very high. Several recent awards have exceeded a hundred million dollars. US law allows for damages to be trebled in cases of willful infringement. As a result some companies have policies which prohibit engineers fro engaging in any patent searches or other activities that might bring patents to their notice and thus provide grounds for a treble damages claim.

The cost of patent licenses is usually paid by implementors. Although these costs will subsequently be passed onto users certain types of implementors may be unable to meet these costs. In addition to providers of Zero Cost Implementations, smaller implementors may be unable to afford minimum license fees or meet other licensing requirements.

Once a patent license is accepted, the licensee is bound by the terms of the license. These terms may continue even after the original patent term expires or the patent is found to be invalid. A royalty-free license may still be regarded as an encumberance by an implementor. For example a royalty-free license may require the licensee to use a particular technology (e.g. programming language) or support particular features (e.g. direct access to a particular online store) that the implementor would not otherwise accept. In particular 'royalty-free' license terms that require the implementor to surrender patent rights to unrelated technologies or to provide the implementation at zero cost are generally considered an encumberance by the affected parties. Sublicensing terms have proved problematic in this respect. Commercial providers of middleware products have objected to license terms which would require their customers to obtain a royalty-free license from a Patent Rights Holder as requiring them to reveal their customer list to a competitor. Sublicensing is also a concern with certain open source licenses as a patent license obtained by the developer may not transfer to another developer that modifies the code.

Adoption is also a concern for Implementors. Implementors will not adopt a technology unless they beleive that they can access all necessary and enforceable patents on acceptable terms.

Some Implementors are also Patent Rights Holders. Their interests in one role may conflict with those in the other. In some cases this conflict may lead to proloinged indecision and thus inaction which in itself impedes the standards process. A particular concern raised by some Implementors/Patent Rights Holders is that the reciprocal rights clause of a Patent License may not be enforceable against a party that obtains a patent grant indirectly through a sublicense. While it has been argued that such reciprocal rights are likely to be found enforceable by a court, in the absence of binding precedent this argument does little to convince a party concerned about litigation risk.

A lawsuit for patent infringement may in certain circumstances be brought against the user of an infringing article. Users with deep pockets may thus be exposed to the same litigation and damages risks as developers and all users can expect that licensing costs paid by implementors will be passed onto them. In some recent cases a plaintif who has been successful in a patent infringement lawsuit against an implementor has subsequently brought suit against users, effectivly claiming double damages for the same infringement.

Users may seek an indemnity against these risks from a third party, typically, but not necessarily, the implementor.

Current IETF practice is that access to all essential claims necessary to implement a specification MUST be available to all parties on RAND terms. The IETF does not require Zero Cost access to be available. While it is clear that there is no IETF consensus to change this policy, there is an even stronger preference for choosing technology available on RANDZ terms unless an encumbered technology overs exceptionally strong advantages.

The most frequent complaint made of the IETF patent policy is the amount of time spent dicussing it. Extended discussion of patent access issues is particularly problematic when it occurs during Working Group or IETF last call at a point where the Working Group has invested considerable time and effort in developing the specification. At present each IETF Working Group is responsible for deciding its own IPR policy within the constraint of meeting the IETF RAND requirments. While this allows WGs great flexibility in chosing their IPR criteria it leads to a great deal of unnecessarily repetative discussion at both the Working Group and IETF level. A particular problem here is the fact that most IETF participants are engineers. Almost none are qualified lawyers. As a result discussions tend to revolve around the twin poles of ideological commitments and the concerns of those who have in the past become involuntary participants in patent litigation.

Allowing Working Groups to negotiate IPR terms with Patent Rights Holders directly is unsatisfactory because they are placed in a weak negotiating position.

Negotiating with a Working Group as a whole is entirely unsatisfactory for the Patent Rights Holder as no party in the Working Group has decision making power, either for the Working Group itself or the IETF as a whole. Any concession made by one member of the Working Group can be repudiated by the group as a whole and even if the Working Group comes to agreement with the Patent Rights Holder this agreement may be repudiated at the IETF level The result is a negotiating posture similar to Axelrod's Prisoner's Dilema in which the rules of the game cause the parties to choose an outcome that is suboptimal for all concerned.

Any concession made by a Patent Rights Holder to one Working Group becomes an unequal precedent. Having revealled their bargaining position, the Patent Rights Holder knows that in the future it will be expected to provide terms that are at least as favorable to any other Working Group. The precedent established is only one way however, the Patent Rights Holder has no assurance that its competitors will be required to provide equally favorable terms in the same circumstance. Unequal precedence encourages Patents Rights Holders to offer the least favorable terms that are likely to be acceptable to the Working Group and is thus counter to the spirit of mutuality on which the IETF is based.

The mere assertion of a patent claim does not by itself mean that the claim is valid or enforceable. Patent applications are particularly problematic in this respect as nobody knows if the application will be granted. Bad faith assertions of patent claims are far from unknown. The most famous example being perhaps the notorious Selden 'automobile patent' which prevented production of low cost volume automobiles until the patent was disqualified in 1911 at the conclusion of a case brought against Henry Ford. A bad faith patent claim might be raised in a standards process in the hope of receiving unjust royalty payments or to influence the process itself. A rule against inclusion of any technology against which a patent claim has been asserted would require the IETF to either examine the validity of the patent claim or allow the Patent Rights Holder to exercise a veto.

Uncertainty imposes costs on all legitimate stakeholders. A Patent Rights Holder seeking to fairly obtain royalties for use of their technology has little incentive to invest in engagement in a standards process if it is uncertain that Royalty RAND terms will be acceptable. Equally an implementor is likely to object if having engaged in a standards process under the expectation that the royalty terms will be RANDZ and find that this is not the case at a late stage. The disadvantage created by uncertainty might be used to advantage by a bad faith Patent Rights Holder, encouraging development of the specification to proceed on the expectation that RANDZ terms will be available without the intention of ever doing so. This is not as great a concern in standards bodies where implementation and deployment follow agreement of the specification. In the IETF and other Internet standards bodies deployment and use is considered an essential criterion for recognition of a standard.

Conventional consideration of patent policy has centered on patent license terms, this despite the fact that few parties have in fact obtained licenses when offered. In most cases the interest of the implementor is not to obtain a patent license, it is to obtain an assurance that a lawsuit will not be brought for infringement. Nor is it likely to be in the interests of a Patent Rights Holder to litigate for infringement of a RANDZ patent license except as a countersuit to an infringement case brought by the alleged infringer. The objectives of the parties is therefore to grant and obtain Patent Access as opposed to a Patent License. In the Microsoft Open Specification Promise this takes the form of an irrevocable promise not to assert any Microsoft Necessary Claims in connection with an implementation that conforms to a covered specification provided that no similar claims are made against Microsoft. This approach meets Microsoft's goal of ensuring the enforcability of the Reciprocal Rights clause in the patent license while avoiding the need for any party to actually obtain a license.

Any changes made to the IETF patent policy must bear in mind the fact that it is only binding on IETF participants. If the policy is changed in ways that Patent Rights Holders who are also Implementors consider to be too onerous, they may cease participating. Patent claims by non-participants have in practice proved far more problematic than those brought by participants. A Patent Rights Holder who is also an Implementor has to consider both the

Some standards bodies begin all proceedings, including face to face meetings and conference calls with a formal recitation of the IPR (and frequently anti-trust) requirements.

All parties should have equal understanding of the consequences of a decision. Effective disclosure of IPR claims is thus essential.

When a proposal is made the parties should be able to predict the likely response.

No party should be able to benefit from unfair practices such as concealling the existence of IPR or bad-faith offers of Patent Access on deliberately ambiguous terms.

Display of NOTE WELL advice should be more prominent than is currently the case. In particular ackinowledgement of NOTE WELL notification should become a required element of subscription to all IETF mailing lists, including lists discussing formation of BOFs.

The charter of a Working Group SHOULD specify the IPR terms which proposals MUST meet in order to be accepted as Working Group work items. In cases where the IPR situation is unclear the chater SHOULD specify clarification of the IPR terms as a work item. While a Working Group does not always meet the objectives set out in its charter, the objectives should always be understood at the outset. It is exceptionally unlikely that a Working Group that is founded with the purpose of developing a RANDZ protocol is going to subsequently choose a Royalty RAND alternative. In some cases a Working Group will be chartered to create a protocol for a technology that is subsequently discovered to be subject to credible Patent claims. While this situation is unavoidable it does not have to necessarily lead to the collapse of the Working Group as has happened in the past. Stating the criteria for acceptable IPR terms at the outset puts the Patent Rights Holder on notice of the criteria they must meet for acceptance and provides them with an assurance that the terms will be accepted if offered. The purpose of a Working Group charter is to avoid unnecessary and repetative discussions that are unproductive. Working Groups are routinely chartered to rule out of scope discussions that would duplicate the work of other Working Groups or have been found to be unproductive in the past. These include definitions of SPAM, development of PKIs in groups not charterted for that express purpose, development of new cryptographic algorithms and the like. The authors of Working Group Charters should be encouraged to add discussion of IPR terms to this list.

Working Groups should be encouraged to require compliance with IPR policy that has already been developed rather than develop new IPR terms within IETF process. In particular requiring compliance with the existing W3C policy requirements is recommended. Reuse of existing IPR terms reduces the work that Implementors and must perform to decide if the terms are acceptable and for Patent Rights Holders to arrive at acceptable terms. The less decision time required by the lawyers, the greater the chance of a satisfactory result. Standardized terms transform a legal question to be decided by the legal department into a business question to be decided by the business unit. Standardized terms have the further benefit of being more predictable if the parties do end up in a potential litigation situation. Standardized terms make it more likely that an applicable precedent can be found. While precedent is not binding in all legal traditions it is still informative in jurisdictions where it is not binding.

Development of IPR policy terms is not core to the mission of the IETF or any othe Internet Standards body. If extended discussion of such issues is required this should be persued in a venue that has the appropriate domain expertise. In addition to developing the policy requirements themselves such a forum might assist in the development of mutually acceptable support materials such as standard Patent Access terms, standard Patent License contracts, NOTE WELL and disclosure agreements for pre-standards activity and the like. Standardization of the legal layer of the Internet becomes increasingly critical as the Internet becomes a social infrastructure.

This draft does not require any action by IANA.

IPR management rasies critical security issues for all the stakeholders involved. IPR is an asset which stakeholders must carefully consider in deciding to grant access