|
Abstract: Intellectual
property rights (IPR), and particularly patent claims, provide special
challenges to standards developers. Following adoption of standard, members may be unwilling to share
implementation rights on licensing terms that are conducive to wide adoption of
the standard (or at all). Or, during
the development phase, they may secretly seek to ensure that a standard will
infringe their undisclosed patent claims, in order to reap a harvest of
royalties when those rights are revealed after the marketplace has already been
locked in. Externally, there will often
be patent claims that are essential to the implementation of a standards, but
whose owners have little or no incentive to make their valuable IPR available
on acceptable terms (or at all). As a result,
standard setting organizations must have IPR policies that are designed to
minimize - although they cannot totally eliminate - such problems. In this article, I briefly review: the
history of IPR issues; the changes in the standard setting landscape that are
accentuating IPR concerns today; the ways in which IPR (and particularly patent
claims) give rise to concerns for standards developers and implementers; the
documents that comprise an appropriate IPR management regime; and finally those
areas in which IPR policies are most significantly being challenged and
reworked today.
Untitled Document
This article is part of the ConsortiumInfo.org Essential Guide to Standard Setting and Standard Setting Organizations, which can be found here: at <http://www.consortiuminfo.org/essentialguide/>. Other Guide topics may be accessed using the links to the left of this abstract.
|
To:
I Introduction
The need to deal with
intellectual property rights – principally copyrights and patents – lie at the
core of standard setting. At the most
basic level, the standards, reference software, compliance test suites, white
papers, and other common deliverables that standard setting organizations
(SSOs) create are themselves copyrightable works, and can be sold by SSOs, if
desired, in order to defray some or all of the costs of their production. At the level of implementation, standards
that dictate technical design features (as compared to performance standards
that only specify output parameters, or physical interoperability standards,
which deal with otherwise arbitrary characteristics such as physical dimensions)
have the potential to require the infringement of patents owned by members and
non-members. Patent infringement has
more significant economic implications, since the owner of a patent can require
the payment of a per-product fee to implement a standard, or can withhold
permission to implement the standard at all.
Every SSO therefore needs a
set of rules that addresses IPR in order to ensure that the SSO owns its work
product upon completion, and to decrease the risk that its completed standards
will encounter IPR-based impediments to broad implementation. Those rules typically address questions such
as whether and when patent claims must be disclosed, so that participants will
know whether all members (at least) will make available any essential patents
claims on what are traditionally referred to as "reasonable and
non-discriminatory" (or RAND) terms.
Achieving agreement on what
such policies should state, and on the process rules under which they are
implemented, can be as contentious a process as creating standards with high
commercial impact. Differences of
opinion on specific tenets and terms exist not only between individual
companies, but also between industries and sometimes among nations. In recent years, the importance of patents
in some sectors, and particularly in the information and communications
technology (ICT) industries, has focused increased attention on IPR
policies. That attention has made
existing SSOs reexamine their rules, and made it all but impossible to form a
new SSO until the founders have agreed upon the IPR policy that the new
organization will adopt.
In this article, I will
survey the changes in the standard setting landscape that are accentuating IPR
concerns today; the ways in which IPR (and particularly patent claims) give
rise to concerns for standards developers and implementers; the documents that
comprise an appropriate IPR management regime; some of the SSOs that have
grappled with difficult IPR issues today, and the changes that they have made
to their IPR policies; and finally those areas in which IPR policies are most
significantly being challenged and reworked today. But first, I will briefly describe the more peaceful days and
uniform policies that prevailed before things began to change.
II The Placid Past
For the first hundred years
of standard setting, IPR concerns were relatively easily addressed, because
most participants in the standard setting process had roughly similar views on
how the rights of inventors and standards implementers should be balanced. During that time period, rules were
developed that were reasonably uniform across the world of standard setting,
minimizing (if not eliminating) the degree to which owning and sharing IPR
complicated the standard setting process.
The relative lack of
friction that prevailed during these years also arose from the fact that
individual setting organizations often constituted independent ecosystems, with
little reason to interact with each other. These organizations could develop rules of behavior under otherwise
substantially similar IPR policies that were well understood by their
members. In some SSOs, those rules
might be highly tolerant of including IPR in a standard that would result in
implementers needing to pay royalties to one owner or another, while in others,
such an outcome would be seen as very undesirable. Moreover, a single standards organization often served all of the
standards needs of an industry niche. As a result, many or all of the standards that a given group of industry
participants might need in order to build their products were developed within
a single, independent organization, and thus subject to the same set of rules
and expectations.
Until recently, standards in
most organizations were created by career professionals, who knew each other
and collaborated over long periods of time on standards, allowing many
potential issues to be identified early in the process, and for differences to
therefore be resolved more amicably. In
addition, a large percentage of the standards created in traditional industries
were unlikely, or indeed unable, to infringe anything that was patentable
(e.g., a standard that sets light bulb wattage increments of 40, 60, 75 and so
on, or a standard that dictates the gauges of electrical wiring).
Standard setting also
occurred primarily at the national level. Indeed, national standards were often created, or at least used,
expressly to make it more difficult for foreign vendors to enter the domestic
marketplace. In such a situation,
infringement of patents could be considered a favorable attribute of a
standard. But in either event, only IPR
issues of national relevance were of concern.
For all of these reasons,
addressing IPR was a comparatively easy challenge, and did not consume a great deal
of time on the part of standard setting organizations, virtually all of which
were accredited standards development organizations (SDOs) that were, in
addition, required to comply with a single high-level set of IPR rules in order
to obtain and maintain accredited status. To the extent that rules varied from organization to organization within
the boundaries permitted by such overall rules, these distinctions were not
likely to cause problems, due to the level of isolation (noted above) that
tended to exist between industry niches, and therefore between those that might
be affected by any differences.
Because of this same
insularity, cross licenses were often easy to arrange among the traditional
leaders of a given industry. And so on,
in a process common to matured industries that have an interest in sorting
things out, and which have evolved common strategies and practices that to
create orderly and predictable market conditions within which to compete. In many SSOs, IPR disputes were therefore
unknown, and many members were only dimly aware of the specifics of the IPR
policies of the SSOs in which they participated – assuming those policies had
specifics at all. Many such policies
were in fact only one or two pages of high-level statements of principle, or
simple references to, or adoption of, the policies mandated by the accrediting
authority to which they were subject.
In short, standard setting
could progress apace with relatively little concern over IPR, either curing or
after the adoption of a standard.
III The Turbulent Present
3.1 A changing landscape. This placid
picture began to deteriorate as the ICT industry evolved with increasing
rapidity in the 1980s. During these
years and after, ICT came to play an ever-larger role in the world economy, and
new SSOs (mostly unaccredited organizations referred to generally as
"consortia") began to proliferate to create the standards (and in
particular, interoperability standards) demanded by these new technologies. The standards that these new SSOs created
applied to technical areas that were often the subject of intense patent
activity, although not uniformly so. Software was not yet patentable in the United States in the early years
of this progression, and therefore SSOs setting standards for microprocessors
and hardware had concerns not shared by the many consortia that sprang up to
create software standards.
The industries that filed
these patents were also used to asserting them aggressively, in order to defend
their right to recoup their often significant research and development costs,
and to reap the profits they expected to earn on such investments. Many also had internal cultures that granted
high status, and sometimes large bonuses as well, upon those that were
successful in acquiring the most patents.
The convergence of
technologies in single devices and business situations caused further problems,
as standards broke out of their sectoral silos. Now, standards that required license fees to implement might be
needed in situations where such costs were regarded as intolerable. To make matters worse, a new business model
arose that involved the creation of patentable inventions for the sole purpose
of generating license fees, rather than to protect the designs of products that
the patent owners themselves manufactured. Companies adopting this model profit only from fees paid by third
parties, and those license fees can increase dramatically if a patented
invention, by accident or design, finds its way into a widely adopted standard. Moreover, since the patent owner (or
"troll," as such a business is often unaffectionately called) does
not build anything that can infringe someone else's patent, there is no
potential to resolve disputes through negotiating a cross license.
As a result, private parties
began to sue companies they believed to have abused the standard setting
process by failing to disclose their patent claims, and to complain to
government regulators as well. When
this occurred, companies with large patent portfolios were at a loss to know
whether to be more reassured or fearful, since the only way to be sure that
they disclosed a relevant patent claim would be to conduct a search of their
patent portfolios. Since some companies
participate in hundreds of SSOs, institutionalizing such a practice would be
prohibitively expensive. Not
surprisingly, these companies strongly resisted any IPR policy change that
might suggest that such a search should become a practical necessity. At the same time, they were concerned that they
might inadvertently fail to disclose a patent, and either lose the ability to
profit from it, or be fined if they tried to do so.
Small companies, or
companies with few patents, had different concerns. Not only could they easily determine whether or not they had an
essential patent, but they also would be unlikely to negotiate a cost-free
cross license with a large competitor that owned a patent infringed by a
standard, although its other large competitors could. Accordingly, companies with few patents focused instead on the
potential for a patent owner to hide its patent until a standard was widely
adopted, and then extort high license fees after the industry had become
"locked in," and unable to redesign the standard in order to avoid a
commercial tax upon the sale of compliant implementations.
These
issues first attracted wide concern in 1996, when Dell Computer entered into a
consent decree with the Federal Trade Commission in the United States. That investigation was brought in response
to private party complaints that Dell had participated in a standard
setting process hosted by the Video Electronics Standards Association (VESA),
and had failed to disclose that it owned a patent that it believed would be
infringed by any implementation of the standard under consideration. Only after the adoption of that standard and
its initial commercialization did Dell identify its patent and assert a right
to require royalties. The FTC concluded that such behavior was a violation of
the antitrust laws, and as part of the consent decree, Dell agreed that it
would grant a royalty-free license to any implementer of the standard; it was also
required to subject itself to oversight in its standards-related activities for
a period of ten years. Had Dell disclosed its patent during the course of the adoption
process, it could have stated its intention to require a royalty from
implementers, and the working group would have had the opportunity to decide
whether to rewrite the standard in such a way as to avoid infringement, or to
knowingly adopt the standard subject to the royalty requirement.
The
impact of the Dell consent decree was substantial, as news of the facts and the
significant penalty became widely known. This was less because the integrity of the standards development process
had been upheld, but because Dell had
contended that it's representative in VESA did not know of the patent in
question, and that its failure to disclose was inadvertent.
Regardless of the truth or
falsity of Dell's defense, the possibility of making an innocent mistake seemed
all too real to large patent holders. From the date the consent decree was
announced until the present, scrutiny of IPR policies has steadily
increased. In recent years, litigation
and investigations have again been launched, based upon similar patters of
alleged conduct. The most prominent
example involves a semiconductor design company called Rambus, Inc. Rambus has been mired in litigation with
multiple semiconductor manufacturers for most of the current decade, and also
been convicted of creating an illegal monopoly by the FTC. More recently, telecommunications giants
QUALCOMM Incorporated and Broadcom
Corporation have been locked in combat, in part over Broadcom's accusation that
Qualcomm has engaged in conduct reminiscent of Rambus.
Today, however, some owners
of large patent portfolios are shifting their focus away from the possible
consequences of an inadvertent failure to disclose an essential patent to a
different concern: whether the products they hope to build will be taxed with
such high levels of patent fees that they cease to be commercially viable. This is in response to a variety of current
trends, including convergence, the increasing number of patent filings in some
areas, the possibility that the troll business model will proliferate, and the
rising importance of open source software. Several of these trends bear mention in greater detail:
3.1.1 Patent
thickets. The number and density of
"patent thickets" (i.e., areas of intense patent activity that focus
on commercially important technologies) is increasing. When standards are created in areas that
have thickets, it is almost impossible not to infringe upon one or more
patents. One consequence is that such a
standard may provide an economic advantage to companies that can cross license
their essential patents at no cost among themselves, while those that do not
hold relevant patents must pay royalties. Where the patent owner is a troll, the playing field is level as among
all competitors, because all must pay, but the cost in license fees may impede
adoption of the standard.
3.1.2 Convergence. A single device
or service may now include many more standardized technologies than in the
past. A mobile device that incorporates
a camera, video, PDA functions, a Web browser, text messaging and wireless, for
example, will implement hundreds of standards. If even a small percentage requires payment of royalties, the device may
become uneconomical to produce. Where
no royalties are involved, a manufacturer may still be forced to negotiate and
enter into many license agreements, each with its own terms.
3.1.3 Open source software. The phrase
"open source software" implies that the actual source code of the
software will be available under one of a number of different licenses that
have been accepted by the Open Source Institute (OSI) as meeting its
requirements for being designated as such. Those terms demand, at minimum, that the source code (i.e., code that
can be easily understood by a developer, as compared to object code, which is
readily "understood only by a computer) must be made available to the
licensee. Many licenses, such as the
GNU General Public License, or GPL, have additional requirements, such as
requiring that someone that modifies and sells software received under that
license must make those modifications available under the GPL as well. The GPL also prohibits including any license
terms deemed to be inconsistent with its own.
When open source software needs
to implement standards, parts of the resulting code may necessarily infringe
upon patent claims that are essential to the standard. If the standard so implemented permits the
patent owner to require license terms that are prohibited under the open source
software license, a conflict occurs. As
a result, open source community advocates, as well as those commercial entities
that are increasingly basing their business models around open source software,
do not wish standards that are important to open source implementations to be
created under IPR policies that would permit license terms to be imposed that
are incompatible with the requirements of popular open source software
licenses. The only way to lessen the likelihood
of that outcome is by adjusting the IPR policy of the SSO creating the standard
to so provide.
For all of these reasons and
more, the IPR rules and policies under which SSOs operate today are matters of
line-by-line concern to many standard setting participants.
IV IPR Policy Coverage and Implementation
While the members of
different SSOs will take different paths in creating and deploying an IPR
regime, the subject areas that each organization must face are common to
all.
4.1 Coverage. Each IPR policy that provides
comprehensive IPR protection needs to address the following subject areas:
4.1.1 Copyright. Copyright issues under IPR policies are
comparatively simple. The
near-universal practice is to provide that if a member contributes any material
for inclusion in a standard, it retains ownership in that contribution, but
irrevocably (a) licenses the SSO to make the contribution available to other
members for purposes of considering its inclusion in a standard, (b) licenses
the SSO to distribute the eventual standard with the contribution included, in
whole or in part, and (c) agrees that
the SSO will own the copyright in the final standard into which the
contribution is incorporated.
The main controversy involving copyright involves not
the terms to be used, but the ability of an SSOs right to enforce its copyright
when a standard has been referenced by, or incorporated into law. The issue is of significance to SDOs, and
not consortia, because the great majority of the former derive significant, and
in some cases the majority of their operating revenues from sales of their
standards, while the latter rarely charge for their standards at all. In general, there is no question that those
SSOs that wish to charge for access to their standards may do so. But in 2003, in a case called Southern
Building Code Congress International v. Veeck, the U.S. Court of
Appeals for the Fifth Circuit handed down a decision that caused a great deal
of consternation among SDOs. In what
has come to be simply referred to as "the Veeck case", the matter at
issue was a Texas building code that referenced a standard, and the matter
arose when someone posted a copy of the referenced standard on a public Web
site. The owner of the standard
thereupon sued, alleging that it was unjust to require the purchase of a
standard that a builder would be legally obligated to implement.
The
holding captured the immediate attention of the SDO community nationwide,
despite the fact that it was binding only in the Fifth Circuit, and even though
it related only to standards referenced, or incorporated into statute. The reason was the fear that the ruling
might be extended to all standards. Following the decision, the SDO sought the intercession of the Supreme
Court, which declined to hear the case, despite the fact that at least two
other Circuits had issued opinions that reached different conclusions. The holding of the lower court therefore
remains binding in the Fifth Circuit - as well as a precedent that courts in
other circuits may decide to follow, even if they are not legally bound to do
so. As a result of the Supreme Court deciding not to hear the appeal,
uncertainty will continue regarding whether other courts will follow (or not)
the Fifth Circuit's reasoning, and if so, whether they will do so in a more
limited or comprehensive fashion.
4.1.2 Trademark. Trademark rules under IPR policies are also
non-controversial. All policies agree
that members retain ownership of their trademarks, and SSOs retain ownership of
theirs. Trademarks are particularly
important to SSOs that conduct, or authorize the operation of, certification
testing, because a license to a well maintained trademark must be obtained by a
product or service provider as a precondition to holding their goods or
services out as being in compliance with the trademarked standard.
4.1.3 Patents. Simplicity ends with patents,
and as a result 90% of the language, and almost 100% of the controversy,
surrounds the rules that relate to so-called "essential claims." These issues will be dealt with in detail in
Section IV below.
4.1.4 Confidentiality. IPR policies usually deal
with confidentiality in one of two ways: by either establishing rules for
describing what is entitled to be maintained in confidence, or by declaring
that nothing is to considered confidential. The latter approach is more common, and is indicative of an organization
that maintains "best practices" in creating what are understood to be
"open standards."
That
said, there is often a distinction between whether information will be public,
and if so, when it will be made so available. In part, this is a reflection of economic concerns, since early access
to the text of an evolving standard can have commercial value, and thus provide
an incentive to become a member. In addition,
those that can view the text and direction of a standard under development
could surreptitiously seek to file patent applications on inventions inspired
by the evolving standard. If the
evolving standard is visible only to members that are bound to disclose any
relevant patents or patent applications, then such a result can be avoided (at
least to the extent that members abide by the rules).
But
if the emerging text is available to all, then a non-member could file, or
amend, a patent application to place it in the way of implementers without
violating any rules at all. For this
reason, many SSOs limit access to an evolving specification to members alone,
and sometimes only to those that have enrolled in the working group that is
producing it, until a later date. For
some SSOs, that later date is after final adoption, while for most it is when
the draft standard is in near-final form. At that point, many SSOs post the draft for public, as well as member,
comment. Those comments will usually be
technical, but may also include a warning by a patent owner that implementation
might result in infringement of its patent, or an alert by someone else that it
is aware of a patent of possible concern.
4.2 Implementation. IPR policies are most often deployed in one of two ways: either
by a Board (and sometimes member) adopted policy, or by inclusion in the
application that each member must sign to become a member. In the former case, a member signs a much
shorter application, but that application includes language that makes all
Board-adopted policies binding upon each member.
IPR
policies, however deployed, tend to be high-level documents. As a result, a more detailed set of policies
and procedures is also required, in order to provide the many procedural rules
necessary to guide multiple standards working groups in creating and
reconciling the standards being created, in setting rules of quorum and voting,
and so on.
As
court cases have consistently indicated, it is essential that an IPR policy, once
adopted, be consistently described and enforced. Similarly, in order for commitments that
members make to be of reliable use in the future, they must be clearly stated. The documentation that instantiates an IPR
regime must therefore be carefully drafted and maintained.
The
final suite of documents to both institute and implement an IPR regime, and
their functions, are commonly as follows:
Document |
Function |
Membership
Application |
A
binding contract between the SSO and the member, that requires the member to
obey the IPR Policy and all related rules and procedures |
Bylaw
language |
A
section of this document gives the Board the authority to adopt, amend and
administer the IPR Policy and all related rules and procedures |
IPR
Policy |
A
set of high-level rules that address all crucial patent, copyright and
trademark issues |
Assertion
Forms |
A
set of short forms used to disclose IPR and to commit to licensing
choices. One form must be completed
in connection with, and must accompany, any submission made by a member for
possible inclusion in a standard, and a similar form must be completed by
every member of a working group at the time that it votes to recommend a
draft standard for adoption (some SSOs require non-working group members that
vote to adopt a standard to complete a form as well). The forms contain language that tracks the
IPR policy, and a member may only select which licensing statement it chooses
to make, and not change the language of the commitment itself. Exhibits are provided for the disclosure
of essential patent claims, and the portion of a draft standard to which they
relate |
Rules
of Procedure |
The
"bylaws" of the technical process |
Trademark
Guidelines |
Rules
of usage for SSO trademarks |
Trademark
License |
As
needed, if a compliance or certification mark program is put into place |
VI Patent
Issues
By
far and away the most difficult IPR issues that SSOs must address involve
patent claims. Unlike other types of
IPR, patent claims can often only be designed around with great difficulty. Sometimes, patent claims can represent
complete roadblocks to a solution, either because no alternative approach is
technically possible, or economically feasible, to implement.
As
a result, SSOs go to great lengths to avoid knowingly adopting specifications
that infringe upon the patent claims of a member or third party that is not
willing to cooperate. At worst, such a
patent owner may be unwilling to provide a license to anyone that wishes to
implement the standard, or may only be willing to provide a license to some,
but not all would-be implementers. Almost as seriously, the patent owner may wish to charge a fee or impose
other terms that would impede or preclude wide adoption.
Unfortunately,
there are two reasons why it is impossible to totally avoid the potential for
inadvertent patent infringement: first,
conducting patent searches for every standard on a worldwide basis would be
both prohibitively expensive, as well as ineffective in avoiding allegations of
infringement, because patent owners will often have different opinions about
the coverage of their patents than will third parties. Second, only a small percentage of the
holders of affected patents will typically be members of a given SSO, or even
implementers of its standards. Consequently, simply asking those members that are willing to act in
good faith whether they have any patents of concern can reveal at best only
some of the patents that may be infringed by any given standard. Even then, a response from an individual to
the best of her knowledge regarding the reach of the patent portfolio of their
multinational employer provides only minor comfort at best.
Nor
can SSO members be required to conduct internal patent searches, since the
companies that participate in the largest numbers of SSOs own thousands, and
even tens of thousands, of patents. As
already noted, no SSO would find itself politically able to impose such a
requirement unless its membership included only those with few, or not, patents. Indeed, almost all IPR policies for this
reason include a very specific term stating that no member shall ever be
required to conduct a formal patent search by reason of its participation in
the SSO.
Even
if all of these problems could be resolved as to members under IPR policies and
rules, non-member patent owners would still be free to act as they wished, and
could therefore at best only be cajoled, rather than compelled, into providing
license rights to all would-be implementers on RAND terms for any patent claims
that might be infringed by a standard.
Despite
this discouraging landscape, SSOs nevertheless do what they can to craft IPR
policies intended to clear the field of problems involving at least
member-owned patents, and in particular to prevent any game playing by members
that might otherwise seek to "plant" patent claims in a standard in
order to reap future economic rewards.
Achieving
consensus on the rules underlying such policies can be a daunting challenge,
however, due to the great value that most companies associate with their
patents, and the variety of opinions that they hold on how best to protect that
value in the context of standard setting. As a result, and absent special concerns, it is usually sensible for an
SSO to adopt as mainstream a policy as possible, since the middle ground
represented by such a policy will likely be the only feasible meeting point for
conflicting viewpoints. Where an SSO
strays from this middle ground, it sometimes finds that its mission may be hampered,
or even defeated, due to its inability to attract a broad membership.
It
is therefore essential that those who are tasked with drafting an IPR policy
understand the positions that individual companies are likely to take on
specific IPR policy terms, and the legal and practical concerns that underlie
these positions. Sadly, after stating
that all patent claims must be available on at least RAND terms, it becomes
difficult to achieve consensus on almost everything else.
The
issues that lead to this diversity of viewpoint may be usefully sorted and
analyzed under the familiar headings of "Who", "What",
"When", "Where" and "Why."
6.1 Who? Some SSO participants believe that it is sufficient to
collect licensing commitments only from those members that directly participate
in creating a standard, since these are the patent owners that are able to
actively "game" the system, either passively, by failing to disclose
their patent claims, or actively, by pushing the process towards adopting a
standard which would entitle them to levy a royalty. This requirement to make a binding commitment to license, or to
disclose the possibility that a RAND license might not be forthcoming, makes
obvious sense from a practical perspective. However, other SSO members may believe that
a licensing commitment should be required of every member of the SSO, whether
or not it chooses to participate directly in a given process. The practical effect of such a requirement
is that the circle of safety is expanded simply by requiring the commitments of
more potential patent owners. More
significantly, if non-participating members have access to specification drafts
as they mature, those non-participating members could file patent applications
that track the evolving standard. Unless required to disclose such patent applications, those members that
were acting in good faith could later be taken by surprise.
Other
SSO members believe that the rules should reach still farther, and insist that
any implementer of a standard - whether a member or not - should grant all
other implementers a cross license of its own patent claims to the extent
necessary to avoid infringement. The
impact of this position may impede implementation of a standard, however, since
most standards need promotion before they become pervasively adopted, and many
implementers may have very low motivations to adopt a standard at all. Where a cross license is required, such
non-member companies may choose not to implement the standard at all.
Of
course, such a mandatory cross license is not likely to capture someone that
has truly bad intentions in any event, as such a patent owner may intend to
reap a far higher return from taxing all implementers than through implementing
the standard itself. As a result, a
standard may be unlikely to receive much benefit - but can be severely impaired
- by a broad cross license requirement. In the case of standards that are relevant to software, such a mandatory
cross license term would violate the terms of many open source licenses as
well, and thus any standard that was created under an IPR policy that permitted
patent owners to require such a cross license would be unacceptable to the
developers of the open source software.
For
all of these reasons, the most common method of addressing this concern under
IPR policies is to permit what has come to be called a "defensive
suspension" term in licenses of essential patent claims. Under this licensing term, a patent claim
owner that has agreed to provide a license to implementers of a standard on
RAND terms is entitled to revoke the license of any implementer that asserts
its own patent under non-RAND terms. This levels the playing field for further negotiation, and is considered
to be consistent with a RAND licensing commitment.
6.2 Who (II)? A related issue briefly alluded to above is whether an assertion
by a member representative that she has no patent claims to disclose should be
limited to the knowledge of the individual alone, or should extend to the
deemed knowledge of her corporate employer (the actual member) as well. For a large multinational company with
perhaps thousands of engineers participating in hundreds of standards working
groups, this prospect represents an IPR manager's nightmare, unless a decision
has already been made by the corporate member that it is willing to make all
patents available on RAND terms – assuming that RAND terms are all that the IPR
policy requires (and not royalty-free licensing). On the other hand, an assertion to the personal knowledge of a
single employee that she is unaware of any potential for infringement is useful
only to the extent that it helps preclude conscious misbehavior. While that is an important result, it is
less useful than an undertaking that a member will never assert a patent at all
on other than on RAND terms.
As
a result, many IPR policies acknowledge that disclosures are made only
to the knowledge of the individual participant, but may still require the
member to commit to a given course of action, if it later becomes aware
of a patent claim that might be infringed. Regardless of which approach is taken with respect to eventual
disclosure, a well thought through IPR policy (or the procedural document
supporting that policy) will contain a term stating that the member will become
irrevocably bound to make a disclosure and licensing statement at the end of
the process once it has participated in a working group for some period of time
(e.g., 60 days). Otherwise, a member
could attend a working group in public, and file patent applications in
private, until just before the requirement to put its IPR cards on the table
matured, and then drop out, only to reveal its "submarine patents"
after the standard had become widely adopted.
6.3 What? There is also a difference of opinion over what rights a
patent holder must grant. Some believe
that, in at least some situations, royalty free licenses should be required
from every company that participates in the adoption process. The most fervent champions of free licensing
would require every member of an SSO to agree to grant a license to whatever
standards may be developed while they are a member (although most would permit
a member to resign to avoid this result in a given situation). Not surprisingly, such a comprehensive rule
would lead many technology companies to refuse to participate in an SSO with a
broad development mission, unless participating in the work of that SSO was
extremely important to their business.
While
not uncommon, such a rule is found mostly in organizations with small
memberships, comprising only companies that have much to gain by creating the
standard. Unless the member companies
control a large portion of the marketplace, however, the rule can be
self-defeating, since other stakeholders that are needed to support the
standard, but who have less to gain from its adoption, are often unwilling to
agree to so strict a term.
6.4 When? While most SSOs require that licensing commitments be made towards the
end of the standards development process, others require commitments to be
given earlier in order to clear the way for smooth final approval of the
standard under development. This is a sensitive issue, because many companies
are not willing to commit to grant a license until they have had the
opportunity to conduct some degree of internal investigation to discover what,
if any, valuable technology rights may be involved. Also, they may wish to assess whether the resulting standard will
be highly favorable to their business, and therefore whether any lost licensing
opportunity will be outweighed by the benefits anticipated from broad
adoption. If a commitment to license,
and especially a commitment to license on a royalty free basis, must be made at
the time of joining a working group (or soon after), the right to evaluate
probable benefits thus becomes limited. Finally, companies often fear that if they make a licensing commitment
in advance, they may be trapped by competitors that conspire to include
valuable IPR owned by the first company in the resulting standard.
For
companies with substantial patent portfolios and the desire to participate in
many standard setting efforts, early decision making is therefore an issue of
significance. Conversely, other
companies (and particularly companies with small, easily searched patent
portfolios) may be unwilling to spend months on helping develop a draft
standard, only to learn at the time that a vote to adopt is taken that a
participant has a blocking patent, and is unwilling to make rights under that
patent available on acceptable terms. Companies that endorse this viewpoint have concluded that the standards
that are developed are more important to them than maximizing the commercial
return on their patent portfolios, so long as they can reserve the right to
charge a royalty on any of their IPR that may eventually be found to be covered
by a finally adopted standard.
6.5 Where? Once a standard has been adopted that requires a vendor to obtain
one or more licenses as a precondition to implementation, the question arises
where and how an implementer can obtain those rights. In an ideal world, all of those rights could be obtained at the
Web site of the SSO that developed the standard, along with the SSO's
permission relating to the standard itself. In fact, patent owners rarely permit an SSO to sublicense patent claims,
and the SSO therefore may only grant rights in the copyright that it owns in
the standard itself. This is typically
accomplished using a short and simple "clickwrap" license, which
exists principally to exclude any warranties of any type (e.g., as to
ownership, non-infringement, and so on).
Where
any member or third party has asserted the right to require a royalty and/or to
precondition implementation upon other license terms (e.g., it may want a
defensive suspension right), a would-be implementer most go direct to the
patent owner to obtain the necessary rights. Typically, the standard setting
body does not become directly involved with the terms of such licenses, and
never takes a position as to whether an asserted patent claim is indeed
essential, it will usually provide a list of those patent owners that claim
that a license is required, together with contact information for such owners.
At
times, there are so many patents that are asserted against a single standard
(as can occur when a standard overlies one of the "patent thickets"
referred to above), the owners of those patents will sometimes form a patent
"pool." In such a case, an
administrator is retained to manage the licensing and economic terms on behalf
of all of the patent owners. This
allows an implementer to pay a single fee, and sign a single license, in order
to obtain all required rights relating to the underlying patents. The fees are then divided among the pool
participants, according to a mutually agreed upon formula.
Regardless
of the licensing arrangements that relate to patents underlying a standard, a
license that forbids the implementer to sublicense can have an adverse impact
under many open source software licenses. Such a restriction will not be problematic for an implementer that
wishes to build an application for internal use only. But if it is a vendor that wishes to sell that application, then
the terms of many open source license agreements would be violated if its
customers were obligated to return to the patent owner for their own license
(and even if that license could be obtained free of charge).
6.6 How? The great majority of IPR policies in existence today still rely upon
RAND licensing commitments. A
reasonable and important question, therefore, is what exactly does
"RAND" mean, and what terms would be considered to be acceptable
under a license intended to meet such a commitment? Perhaps surprisingly, SSOs almost never try to define exactly
what either "reasonable" or "non-discriminatory" is
intended to mean. Even more
surprisingly, until recently there have been relatively few disputes litigated
between patent owners and implementers over these terms. Most SSOs have shrunk from becoming involved
in such disputes, since they lack the resources or the will to take an active
part.
Despite
this overall reluctance for any party involved to tackle the RAND definition
issue, the popularity of the simple RAND rubric continues, giving rise not only
to obvious questions (e.g., how high can a royalty be before it becomes
"unreasonable?"), but to more subtle ones as well, such as whether it
is discriminatory to charge a cross license partner nothing to implement a
standard, while charging a royalty to the competitor of the partner with whom the
patent owner does not have a patent license. If such a position does not represent discrimination, then the second
company will be at a price disadvantage to the first.
6.7 Why? Another interesting question
relates to why given companies take particular positions on certain
issues. A simple explanation is that it
is difficult for a company to step outside the realities of its familiar
proprietary world and assume the mind-set necessary to give away something (i.e.,
valuable IPR underlying a standard) in order to gain something of greater
commercial value. One often
under-appreciated benefit is the luxury of making a safe strategic decision
(e.g., for a manufacturer, knowing in advance that it is committing to what
will prove to be the "VHS" rather than the "Betamax"
standard). Thus enabled, it can compete
with other SSO members in making better and more appealing products based on
the adopted standard, and address a more swiftly and surely developing market
for those products. Even long-term
participants in the standards process can sometimes catch themselves taking a
position that is inconsistent with consortium goals, simply out of habit. The most enlightened and successful
participants in standard setting (in this author's opinion) are therefore those
that most thoroughly "get" the fact that they have far more to gain
from the success of a standard than they could expect to gain in patent
royalties on any underlying patents.
Another
cause of confusion and insistence on unnecessary and counterproductive
positions is the superficial similarity between commercial joint ventures and
SSOs. In the former, a small number of
companies forms an alliance under a joint development agreement to create a
product or other deliverable that the participants can then sell, or otherwise
exploit. In this type of activity, it
is typical for all participants to cross license all patent claims to each
other that would be infringed by a joint created specification, and to permit
each other to sublicense implementation rights to third parties as well. But the core cross-license rights are
usually restricted to the small number of "by invitation only"
participants, that wish to keep control of the design, as well as the most
lucrative commercial benefits, to themselves.
In
an SSO, however, the goal is to create a standard that is adopted and
implemented by as many companies as possible. As a result, every implementer is given equal rights to every other
implementer, as an incentive to participate in the ever widening pool of
adopters.
While both types of efforts involve multiple competitors gathering to agree on
technology solutions, there are several significant differences. First, the participants in a commercial
joint venture are highly motivated to achieve a carefully defined and limited
common goal, and are therefore willing to share all IPR needed to achieve that
goal. Similarly, their partners are
highly motivated to gain access to the same rights, and are therefore willing
to enter into sublicenses and agree to payment terms. The legal vehicle employed by the joint venture participants - a
contract - is appropriate, since few or no new members are expected to join,
and the founding members are not expected to leave until the goal has been
achieved. Finally, there is no need to
create a pretense of "openness."
In
sharp contrast, an SSO needs to allow members to join and leave, and needs to
make it as easy and attractive as possible for non-members of many stripes to
adopt and implement its standards. A
key component in achieving this goal is to be structured and operate in as
"open" a way as possible, to negate any appearance that one or more
companies can unduly influence the eventual nature or availability of its
standards, thus giving them a commercial advantage over other implementers.
Since most individuals who
represent companies in SSO activities - and even those who are tasked with
forming new consortia - have limited knowledge about the theory and practice of
SSO formation and operation, it is easy for them to assume that whatever
previous organization they have participated in represents the gold standard of
structure and governance. One
unfortunate outcome of this reality is the surprisingly large number of consortia
that have been formed on the more restrictive commercial joint development
model, even where the tight controls and high demands of that model were
unnecessary. Where this has occurred,
the result has usually been to hamper, rather than lead to, the success of the
organization.
VII Current Issues and Trends
Thus
far, this article has described the general shape of IPR policies, positions
and solutions that have prevailed over the past decade. Today, however, some participants in the
standard setting community are calling for significant changes in IPR policies
and practice. Topics currently under
discussion include the following:
7.1 Ex ante disclosure. The basic concept behind ex ante disclosure is "earlier is better, and more is better." In practice, what that means is permitting or
requiring those that are involved in creating a standard to state more than
just whether they have patents, and whether they are willing to license them on
RAND terms, but what those terms are in greater detail.
The
very topic of ex ante disclosure causes alarm for many, due in part to
antitrust concerns. For decades, those
that participated in standard setting activities have been instructed by their
lawyers to never make any mention of prices in a room populated with
competitors. While certainly it is true
that the actual negotiation of prices could have dire results, representatives
of both the Federal Trade Commission as well as the Department of Justice have
made statements in the last two years indicating that ex ante disclosure, properly undertaken, could be looked at favorably by the
regulators, and in fact could have procompetitive, rather than anticompetitive,
effects.
Recently,
two ANSI accredited SSOs, the IEEE and VITA, have drafted amendments to their
IPR policies that would permit (in the case of IEEE) and require (in the case
of VITA) the disclosure of the terms upon which essential patent claims would
be licensed under a draft standard, if finally adopted. Both organizations submitted requests to
the FTC for "business review letters," under which the FTC would
evaluate the SSOs proposed ex ante procedures. In each case, the FTC responded favorably. The guidance provided
in those letters, as well as in any letters that may issued to other SSOs in
the future, will be very useful to those SSOs that wish to consider similar
amendments.
The
IEEE and VITA policies will also be reviewed in the context of the rules of the
American National Standards Institute (ANSI), the accreditor of each SSO. That review will be conducted by the ANSI
Executive Standards Council (ExSC), the appointed body for the task under the
ANSI Essential Requirements: Due Process Requirements for American National
Standards. As with the FTC business review letters, the
results of these reviews will be instructive for other ANSI accredited SSOs
that may wish to amend their own policies and procedures.
The
overall topic of ex ante disclosure, however, is a difficult one, and
there is a wide variety of opinion regarding the advisability of wide-scale
adoption of changes to permit such disclosures. Some very large companies, particularly in the information
technology sector, are strongly in favor of such changes, while others in the
same market space are opposed. The
reasons for such opposition range from concerns over antitrust exposure, to the
belief that such disclosures will not prove to be useful, to a desire to retain
as much freedom of movement as possible in order to maximize licensing returns.
7.2 Royalty free licensing. There has long been a preference in many sectors
for adopting standards that would successfully avoid infringing patents that
would have price tags attached. With
the advent of the Web, this objective became more determined in some
organizations, and in particular in the World Wide Web Consortium (W3C), an
organization with a broad technical program, a very large membership (including
most of the largest IT companies in the world), and an important mission. Those who participated in the revision
committee at the W3C endured a marathon, three-year process that eventually
resulted (in May of 2003) in an amended IPR policy intended to make it nearly
impossible to finally approve a standard that knowingly would result in the
requirement of paying a royalty or other fee to a patent owner.
The
arduous process followed by the W3C involved the business and legal
representatives of many of the largest technology companies in the world, and
marked something of a watershed. Prior
to the approval of the new policy, such an outcome would be been difficult to imagine,
but afterwards, other SSOs could more seriously consider adopting policies that
either mandated the same result, or permitted individual working groups to
operate under a similar rule set. The
Organization for the Advancement of Structured Information Standards (OASIS)
took this approach in an amended policy (which itself took two and a half years
to complete) that was adopted on April 15, 2005.
7.3 Open source licensing. Many of the licenses under which
"open source" software, and in particular "free and open source
software" (FOSS), prohibit certain types of conduct on the part of anyone
that wishes to commercially redistribute a software product that includes the
open source software, whether in whole or in part. As has already been mentioned, some of those prohibitions relate
to terms that would be otherwise acceptable in licenses that apply to patents
underlying the implementation of open standards.
The
result has been not only a periodic clash between advocates of open source
licensing and developers of open standards, but further complications for the
already difficult process of achieving consensus over IPR policy rules. When OASIS adopted the new IPR policy noted
above, it added a track under which a given working group could decide to
create a standard under rules conducive to open source implementations (it also
included tracks that allow only RAND-free implementations, and finally a third
track, that would allow RAND-royalty assertions).
7.4 Harmonization. Given the difficulty of achieving consensus
on IPR policy terms and the great amount of effort that negotiations over such
terms can consume, several recent initiatives have been directed at
harmonizing, or at least better understanding, IPR terms.
One
such effort has been underway in a subcommittee of the American Bar Association
for over two and a half years. During
this period, attorneys have been meeting regularly by telephone and in person
to create an illustrative IPR policy, with annotations explaining the rationale
behind common IPR policy terms. One
important goal of this project is to permit new SSOs, as well as existing SSOs
that wish to upgrade their IPR policies, to more easily achieve consensus over
what their own IPR policy should state.
Another
effort relates to making it less confusing and burdensome to participate in
multiple SSOs. On March 19, 2007, the
three "Big I's" of global standardization, the International
Organization for Standardization (ISO), the International Electrotechnical
Commission (IEC) and the International Telecommunication Union (ITU), jointly
announced that they had "harmonized" and "aligned" their
respective patent policies.
7.5 RAND specificity. Although no specific actions are known to
this author to be underway, the vagueness of a RAND declaration is becoming the
subject of increasing criticism, as well as litigation, as earlier noted. While mandatory ex ante disclosure
requirements could moot the need for RAND declarations entirely in some SSOs,
it is possible that other SSOs may someday decide to more precisely define
exactly what RAND should mean instead, in order to narrow the scope for
potential downstream surprises.
VIII Summary
It
would be an extreme understatement to suggest that the technology industry has
reached a state of clear consensus on what constitutes the ideal IPR
Policy. A single example from the
author's experience will demonstrate how elusive a goal such a state of
universal enlightenment could prove to be. Not long ago, I spoke separately with two in-house attorneys from the
same company, on the same day, relating to the IPR policies of two different
consortia. In each case, the lawyer to
whom I spoke was adamant in his contention that his company could not, and
would not, join any SSO with an IPR policy that included the language in
question. The problem was, of course,
that the language under discussion in each case was essentially the same, while
the positions taken by the two attorneys were completely opposite.
The
good news is that awareness of the issues at stake in IPR policies has risen
remarkably, in part due to the fact that several high-profile cases involving
standard setting abuse have been recently litigated. While this increased attention has brought more stakeholders into
the discussion, and therefore has slowed the process of achieving consensus,
that same dialogue has served to make those participating in these discussions
better versed regarding common IPR policy concerns, as well as better
acquainted with the compromise terms upon which these concerns are most likely
to be resolved.
Notwithstanding
this relative progress, it will likely remain a challenge for each SSO to craft
an IPR policy that all of its members can live with, and which can still
facilitate the achievement of standard setting goals. While the process of promulgating a robust policy can hardly be
described as recreational, it is a task that cannot be avoided by any new
consortium that wishes to recruit members. Similarly, every existing SSO that has not reviewed its IPR policy in
recent years would be well advised to subject its process and rules to a
critical analysis, with the goal of purging it of ambiguities and to bring it
up to date with current member expectations.
For an existing consortium,
tackling an overhaul of current policies will inevitably be a challenge. But for the members of a new consortium,
there is an advantage for tackling the process while their numbers are few, and
consensus is therefore more easily achieved. Thus, the virtue of facing up to a rather thankless chore early on will
be rewarded by an easier path to a successful conclusion. Those who shrink from that chore will
inevitably face a much more difficult task down the road.
NOTE: If you participate in the standards development
process or are a member of the management of a standard setting organization,
please feel free to contact Andrew Updegrove if you have questions about this topic.
Copyright 2007 Andrew Updegrove
When this is later realized, such an
organization often restructures itself, and adopts a more open model. I have assisted in several such latter day
conversions, all of which were followed by increased success for the SSOs
involved.
Section 2 of the W3C Patent Policy now provides as follows:
In order to promote the widest adoption of Web
standards, W3C seeks to issue Recommendations that can be implemented on a Royalty-Free (RF) basis.
Subject to the conditions of this policy, W3C will not approve a Recommendation
if it is aware that Essential
Claims exist which are not available on Royalty-Free terms.
The
world’s leading international standards organizations have adopted a harmonized
approach to address the inclusion of patented technology in standards. IEC
(International Electrotechnical Commission), ISO (International Organization
for Standardization) and ITU, under the banner of the World Standards Cooperation
(WSC), have aligned their policies which allow for commercial entities to
contribute the fruits of their research and development (R&D) activity safe
in the knowledge that their intellectual property rights are respected.
Untitled Document