MPEG LA Releases DRM Reference Model 2.0
August 4, 2004
By Bill Rosenblatt

The MPEG Licensing Authority (MPEG LA) released Version 2.0 of its DRM Reference Model last week.  MPEG LA's DRM Reference Model is part of its attempt to create a pool of licensable patents on DRM-related technology, so that would-be DRM implementors can license relevant intellectual property with the convenience of one-stop shopping and the confidence that IP holders are unlikely to come out of the woodwork after the fact and demand royalties. 

The Reference Model Version 2.0 is designed specifically to cover DRM implementations that conform to the Open Mobile Alliance's OMA DRM 1.0 specification.  The focus on OMA DRM 1.0 may raise a few eyebrows, but it does make sense given the number of vendors rushing to create OMA-compliant DRM implementations these days. 

We have said in the past that MPEG LA has a hard job before it; DRM is a far more complex field than that of content compression technologies, which MPEG LA has successfully tackled in the past.  This version of the Reference Model, which is the first one that MPEG LA has made at all public, is an excellent step.  It uses data flow diagrams, decision tables, and relatively simple process specifications (PSPECs) to describe a series of processes that captures most of what goes on in a typical encryption-based DRM technology.  For the most part, it is a superset of, and goes into more detail than, the Generic DRM Architecture defined in my 2001 book, Digital Rights Management: Business and Technology. 

The most important consideration that MPEG LA has had to address is where to draw the boundaries of DRM for the purposes of the Reference Model.  It appears that MPEG LA has chosen to do so based on the boundaries of IP that already exists.  For example, the Reference Model does not include the details of encryption schemes and security operations; it stops short at their basic functions, such as entity credentialing, tamper proofing, data encryption, digest generation, and so on.  The Reference Model also does not get into the details of how content consumers' identities are established (e.g., those of devices and/or human users) or how they establish their credentials to exercise rights to content (e.g., by making payments).  However, it does encompass such things as license issuance and the use of rights expressions, both of which are part of the IP landscape of DRM.

Yet we did find one boundary that MPEG LA seems to have drawn that we feel should have not. Although the Reference Model does take into account the specification of rights, both the set of all rights that a content owner wishes to allow on a given piece of content, which is established during the content packaging process, and the set of particular rights granted to a user in a license, the Model says nothing about how the latter rights specifications are communicated to an application for exercise.  In fact, the Model seems to assumpte that the only thing that an application will do with content is "consume" it.  This may be adequate for conformance with OMA DRM 1.0, which is geared towards mobile devices with simple content-related capabilities, but it is not adequate for more general DRM implementations.

Rights expressions can get very granular about consumption, such as copy/paste, view, print (i.e. render to fixed physical media), and so on; they can also be used to specify rights that do not fall under the rubric of "consumption," such as derivative work rights (e.g, excerpting, language translation).  There is and/or will be intellectual property that covers the communication of those types of rights to applications that exercise them.  The current Reference Model simply suggests that those applications are "certified" and "credentialed," i.e., trusted to do exactly what they are told to do, but that's as far as it goes. 

Once MPEG LA finalizes the Reference Model, it can be used as a map for both patents and the specs of actual implementations.  A DRM technology expert will need to map claims in patents to steps in the Reference Model's PSPECs.  What keeps this from being a straightforward process is the lack of consistency in terminology used in patents that date back from over ten years ago to the present (and future).  This ambiguity was a major factor in the Microsoft/InterTrust patent litigation that settled in April of this year, and it should be a source of controversy that will inevitably slow down MPEG LA's process of getting its DRM patent pool accepted by the industry.

Of course, it is in the industry's best interest that the process be as objective as possible and that everyone cooperate in it.  MPEG LA is trying to do the right thing.

1:30:15 PM    

The FCC last Wednesday approved a number of technologies for sharing of digital video content on home networks in advance of its Broadcast Flag regulation, which is to go into effect next year.  The approvals were made amid some controversies over so-called proximity controls and about the effect on innovation of IP licensing terms in approved technologies.

A total of 14 technologies were approved.  Those from individual vendors included Microsoft's Windows Media DRM, RealNetworks's Helix DRM Trusted Recorder, TiVo's TiVoGuard, Sony's Open MagicGate Type R, HP and Philips Electronics's Vidi, Thomson's SmartRight, and JVC's D-VHS.  In addition, the licensable standard 5C Digital Transmission Content Production (DTCP) protocol and the 4C Entity's Content Protection Recordable Media for Video Content were approved.

All of these technologies provide some flavor of restriction amid ability to transmit digital video around a home entertainment network or to copy it to fixed physical media such as recordable DVDs.  Although none of the FCC commissioners dissented from the approval, one of them (Kevin Martin, a Republican) raised two objections. 

One of these had to do with proximity controls, which relate to the definition of "home" in "home networking."  One of the most problematic aspects of DRM for home networking is where one draws the boundaries of "home" -- whether "home" includes the car in the driveway, the portable device in the backyard, and so on.  The more one thinks about this issue, the more difficult it seems to pin down; for example, basing proximity on physical distance from a set-top box could either penalize rural homeowners or provide insufficient protection from the activities of urban apartment dwellers.  Commissioner Martin objected specifically to the lack of proximity controls in TiVo's technology, which brought on complaints from various content owners.  This is an argument that we suspect will drag on for some time to come.

The other concern raised by Commissioner Martin is not about the technology itself but about intellectual property issues.  Our view of the entire Broadcast Flag proceedings is that they have been, in effect, conducted as an exclusive club of large entities such as Microsoft and members of the 4C and 5C consortia.  Technology definition and approval processes take so long and involve so much bureaucracy that innovative startup companies could be starved out of the process before any decisions are reached. 

The FCC has taken no position on whether the approved technologies must be licensable under any specific regime, such as RAND (reasonable and non-discriminatory) or royalty-free terms; most global standards bodies take affirmative positions on such matters.  Commissioner Martin expressed the concern that the DTCP standard requires licensees to give up intellectual property rights to any related technologies they may have.  If (as expected) DTCP becomes the predominant standard in the home digital video networking market, then innovators will lose their incentives to participate; they could only garner IP licensing revenue if they do not take a license to DTCP.  This stipulation in the DTCP license is clearly favorable to major consumer electronics vendors at the expense of small innovators.  This situation could result in lack of momentum around the DTCP standard, lack of innovation, or both. 

We would prefer the market to determine standards for this type of technology.  By passing the Broadcast Flag regulation last year, the FCC has already put that outcome in jeopardy.  In that context, we are glad to see the FCC moving quickly to approve technologies in time for the 2004 holiday consumer electronics shopping season instead of holding back until next year, but we hope that the FCC does not cause any more bottlenecks to innovation which would hinder or slow down market-based progress in home entertainment networks.

1:29:08 PM    

Playback¡GJPEG/Motion-JPEG/MPEG-1/MPEG-2/MP3
Support¡GCF/IBM Microdrive/SM/SD/MMC/MS/MS-PRO

10:18:28 AM