|
XrML, ODRL, XMCL. Rights expression languages promise standards-based DRM. But can we really all just get along?
January 22, 2002
URL:http://www.drdobbs.com/web-development/freedom-of-expression/184413231
It goes without saying that digitally distributed media has broad consumer appeal. Napster proved that. What began as a one-man hobby in a college dorm room soon erupted into an Internet phenomenon rivalling even Netscape's early hype. Following on the heels of Napster's MP3 file sharing success, competing P2P applications like Gnutella, Audio Galaxy, and Morpheus emerged. These applications opened up the possibility of digital distribution for the full range of media formats, including audio, video, still images, e-books, and more.
Then came the hard part. For many in the media industry, the legal morass Napster found itself in proved one thing: before online digital distribution could become a viable business model, the delivery method needed to be tied to technologies that programmatically enforce the rights of content creators and publishers. The race to provide those technologies reached its peak in the year 2000, by which time more than 100 companies had announced products and services for the burgeoning digital rights management (DRM) market.
Unfortunately, these multiple simultaneous efforts led the industry to an impasse. None of the various competing and conflicting DRM schemes is a clear winner, and the inevitability of consolidation looms overhead. Given this volatile environment, potential customers have been hesitant to hitch their fortunes to any one approach. Thus, the adoption rate of DRM technologies has remained slow. For digital distribution to succeed, there must be broad interoperability between the different protection, commerce, and media management systems.
Monolithic DRM platforms offer one kind of interoperability. If your media delivery and rights management system is an end-to-end solution provided by a single vendor, you can be reasonably certain that all of the components will work well together. Perhaps the best example of this approach is Microsoft Windows Media Rights Manager (WMRM). With this suite, Microsoft provides tools for the entire life cycle of rights-protected media, from packaging and distribution to license management, all the way to playback on the client side.
Predictably, OS platform independence is the first casualty of Microsoft's design. A Microsoft product brochure describes WMRM as "A Win/Win for Everyone." This description is accurate, if ironic. While the initial release supported the Mac OS and several third-party media player applications, with version 7.1 Microsoft scaled back WMRM's scope to make it a truly "Win/Win" solution: Windows on the client, and Windows on the server. Other than Microsoft's own Windows Media Player, only the Windows version of MusicMatch Jukebox is cited as supporting the new WMRM playback architecture.
This tactic is in keeping with Microsoft's Windows platform strategy as a whole, but it also offers significant technical advantages. Most notably, it lets Microsoft build a foundation for its DRM capabilities by introducing a security layer at the OS level—an idea it recently patented (see the "Rights: Patent Pending" sidebar for more information). This foundation was at least partially delivered with Windows XP. However, even Microsoft is hesitant to claim it as a bulletproof measure. "We don't believe any possible DRM system is actually invulnerable," says David Caulton, product manager for the Microsoft Windows digital media division.
Likewise, most content publishers think the idea of a single, universal OS and media player application is unrealistic, if not downright ludicrous. The Association of American Publishers was unequivocal in its position paper on DRM for e-books, which was published in November 2000. Today's reality, the association asserted, is that digital media providers must be prepared to deliver content using as many media formats and DRM architectures as necessary to reach the broadest possible audience.
The paper goes on to dismiss the touted benefits of end-to-end DRM solutions like Microsoft's as "hyperbole," insisting that development of industry standards is the only true road to DRM interoperability in the e-book arena. Perhaps it's no surprise then that Microsoft chose the e-book marketplace as the testing ground for its departure from WMRM's monolithic DRM architecture with the release of Microsoft Digital Asset Server (DAS) in late 2000. The key to this new, standards-based approach wasn't a Microsoft product, however. Rather, it was a little-known XML-based rights expression language called XrML.
XrML's origins can be traced to 1996, when Dr. Mark Stefik of Xerox's Palo Alto Research Center had an idea. The best way to enable Internet commerce, Stefik reasoned, would be to develop a device that could store data under a strictly defined set of access procedures. Unless those procedures were followed, the device would act like the digital equivalent of a bank vault, allowing data neither to enter nor leave.
Stefik named this hypothetical device a "trusted system," and
he published the idea in a seminal paper entitled "Letting Loose the Light: Igniting Commerce in Electronic Publishing." The paper proposed that trusted systems would require machine-readable languages for defining the access procedures. The first example of such a rights expression language was Stefik's Digital Property Rights Language (DPRL), which he based on Lisp.
Xerox waffled on plans to deploy DPRL commercially, but the language succeeded in capturing the imagination of the digital publishing industry—although not without some criticism. "In my view, DPRL is not quite ready for prime time as a standard," John Erickson, V.P. of technology strategy for Yankee Book Peddler, wrote in 1998. "Both DPRL and the electronic publishing community would be far better served if this language were XML-based."
A rights expression language built with XML would be extensible, Erickson reasoned, permitting it to adapt to the evolving needs of rights management, and giving it a much greater chance of interoperability with other emerging data standards for e-commerce. What's more, XML is human-readable as well as machine-readable. An XML-based rights expression language would make it easier
for humans to comprehend and create elaborate rights policies.
As it turned out, Erickson's comments presaged another of
the defining moments in DRM. That same year, Stefik and Xerox unveiled the specification for DPRL 2.0, the first XML-derived rights expression language. This iteration showed enough promise that it earned first its own business unit at Xerox, then its own company. Xerox and Microsoft launched ContentGuard in April 2000, with the aim of further refining and commercializing DPRL. The result was XrML.
"The DRM infrastructure business hasn't gotten traction for a number of reasons," says Rajan Samtani, ContentGuard's director of sales and marketing. "Obviously XrML is not, in and of itself, going to be the be-all end-all. But we do believe this is one important foundational piece."
Samtani is being modest; in fact, ContentGuard is so convinced
of the essential role that rights expression languages will play in
the future of DRM that it abandoned its DRM software business
in August 2001. It now concentrates exclusively on developing the XrML standard.
Central to ContentGuard's strategy is the idea that the market is likely to increasingly favor best-of-breed DRM solutions from multiple vendors. "This notion of end-to-end DRM solutions is not going to
fly over time," says Samtani. Instead, he explains, XrML lets content publishers build heterogeneous networks of trusted systems, where all of the systems can communicate with one another.
How diverse can XrML's trusted systems be? In November 2001, ContentGuard published the specification for XrML 2.0, the first significant revision of the language since its evolution from DPRL. This new iteration broadened the number of supported business models beyond the Napster approach to include both content-driven and service-driven approaches. Instead of merely assigning rights to individual pieces of content, XrML 2.0 lets you define access procedures for libraries, collections, databases, or even Web services.
The most frequent criticism leveled at XrML is that the language is overly complex, because it comprises more than 100 distinct elements. However, its comprehensive feature set has earned it an impressive roster of endorsements from digital media and publishing heavyweights including Adobe, Audible, Barnes and Noble, Hewlett-Packard, McGraw-Hill, St. Martin's Press, Wiley, and Time Warner Trade Publishing.
Even given these companies' collective blessing, Microsoft remains ContentGuard's most important partner. The development of Microsoft Digital Asset Server indicated that XrML had moved beyond white papers and specifications to become a real-world solution with active deployment. Thus, XrML remains unique, even though it must now share the market with some more recent challengers.
Not to be outdone, RealNetworks, the leader in streaming media
and Microsoft's top competitor in the network-delivered video on demand market, began work on its own DRM language in 1999.
It called the new XML-derived dialect eXtensible Media Commerce Language (XMCL), and quickly began drumming up industry support for what it hoped would become a new standard in
rights expression.
The goal of XMCL is essentially the same as that of XrML: to enable interoperability between trusted systems.
Yet the fact that RealNetworks chose to call its standard a commerce language is significant. Instead of developing a broad grammar for defining access parameters for a variety of networked content and services as ContentGuard has done, XMCL is more narrowly focused on describing the business rules that define existing e-commerce business models.
By June 2001, the XMCL Initiative boasted the endorsements of Adobe, Hewlett-Packard, IBM, Intel, Napster, and Sun among others. Conspicuously absent from the attendance roster, however, were representatives from Microsoft—and for good reason. None of them were invited. As one unnamed RealNetworks executive told Wired magazine, "We were afraid they'd wreck the party."
Unmoved, Microsoft is still backing its own horse in the rights expression race. According to product manager Geordie Wilson, XMCL "doesn't seem to add any clear benefits to XrML." Some industry analysts agree, characterizing XMCL more as a competitive tactic than a serious effort. "This whole XMCL initiative on the surface appears to fill a real need in the audio and video space, but implicitly it's kind of a jab at Microsoft," says Jupiter Media Metrix analyst Mark Mooradian.
For that jab to become a knockout punch, XMCL must first beat out the earlier standard in volume software deployment. Judging by that benchmark, ContentGuard emerges the winner. While only one company—Microsoft—is cited as having actually implemented XrML, as of this writing XMCL can't even match that claim. RealNetworks itself has yet to incorporate the language into its RealSystem Media Commerce Suite, leaving XMCL without a single product to showcase its capabilities.
Renato Iannella of Australian DRM vendor IPR Systems isn't satisfied with choosing between XrML and XMCL. Although both ContentGuard and RealNetworks claim that their languages are open standards, in reality, each is the product of a single company's vision, and each remains under the primary control of its respective parent. The problem with that, Iannella believes, is that most DRM vendors' perspectives are far too narrow.
"Traditional DRM (even though it is still a new discipline) has predominately taken a closed approach to solving problems," Iannella wrote in an IPR Systems position paper for a W3C workshop on DRM held in January 2001. "That is, DRM has primarily focused on the content protection issues more than the rights management issues." When rights management technologies are colored by ideas pertaining to security and data protection, Iannella states, the usefulness of those technologies is often severely diminished.
Years earlier, thinking about this very issue led Iannella to develop an entirely new XML-based rights expression grammar, one he called the Open Digital Rights Language (ODRL). He envisions this language eventually becoming part of a larger Open Digital Rights Management framework, one that would encompass not just the technical aspects of DRM, but the business, social, and legal issues as well.
The ODRL language focuses on a simple, extensible rights-management model that encompasses a small set of core elements. The specification is published for use without restriction, "in the spirit of open source software," and it reached version 1.0 status in November 2001. Even so, ODRL remains arguably the least mature of the current rights expression languages. Like XMCL, it has yet to be implemented in any shipping product.
It's one thing to standardize a language, but quite another for any one language to become a standard. To help legitimize their efforts, the creators of XMCL, ODRL, and XrML have each announced their intent to submit their work to various prominent standards bodies for ongoing development. The extent to which they've followed through on those promises, however, has varied.
In the past, RealNetworks has had a good track record of support for open standards. The company's efforts contributed significantly to the development of the Realtime Streaming Protocol (RTSP) and the Structured Multimedia Integration Language (SMIL). But while RealNetworks's Jeff Albertson announced in June 2001 that the company planned to submit XMCL to the W3C standards body "within a month," so far that promise has gone unfulfilled.
Supporters of the ODRL Initiative have been somewhat more proactive. In November 2001, the ODRL 1.0 specification was submitted to the ISO/IEC MPEG standards body for consideration as the rights expression language component for the developing MPEG-21 media distribution standard. The submission was backed by companies
as diverse as Adobe, IBM, IPR Systems, Nokia, and Panasonic. Surprisingly, RealNetworks also supported it, choosing to merge the XMCL specification into ODRL rather than submit its own language independently.
But in the end, it was XrML, and not ODRL, that was chosen
as the starting point for the eventual MPEG-21 rights expression language. "We think it's a pretty big win in the marketplace," says ContentGuard's Samtani. Even so, the company doesn't plan to stop there. "This is just a first step," he explains. "XrML will be submitted to any standards body that needs a rights language." So far, ContentGuard has already submitted the language to the TV Anytime Forum.
One way or another, consolidation amongst the various rights expression languages seems inevitable. ContentGuard's track record of success, though limited, makes it the current front-runner. Yet the reputation of the MPEG group may draw more attention to its own eventual, XrML-derived language, which will be specified with the input of both RealNetworks and the ODRL Initiative members.
Whatever the outcome, a standard rights expression language is but one brick in the foundation of a viable DRM platform, albeit an important one. True, gaining the support of standards bodies is an essential step. But the long-term goal—gaining the support of consumers and business customers alike—still lies ahead.
Neil is senior technology editor for New Architect.
Terms of Service | Privacy Statement | Copyright © 2024 UBM Tech, All rights reserved.
Blazing the Trail
On Guard
Real Competition
Wide Open
Toward Standardization