Folks,

Steve Jobs' statement on music and DRM came out yesterday, neatly dovetailing with a piece I wrote last week. I intended to send it out next week on my regular schedule, but since it's very timely I thought I would just send it out early.

<http://www.apple.com/hotnews/thoughtsonmusic/>

I read an interesting article on the International Herald Tribune web site via Slashdot:

<http://www.iht.com/articles/2007/01/21/yourmoney/music.php/>

Are the record companies finally giving up? I don't know, but I can tell you that in the long run Digital Rights Management software cannot succeed. Why? It has to do with a combination of economics and technology.

Let's start with the economics of DRM. DRM-protected content doesn't just get played back on computers, it also gets played back on portable consumer electronic devices like DVD players and iPods. Thus, the DRM has to be lightweight enough to be implemented in a chip that can be sold as a part of a consumer electronics device. For instance, the DRM for a DVD player cannot use a full-blown modern CPU, it has to function on a chip that can be sold for no more than $10 or so and that can function (in the case of a portable device) on a tiny fraction of the power budget. Otherwise, the DVD player manufacturer will not be able to sell the DVD player at a profit in today's marketplace. The DRM had to do the same work on a chip of not-too-much-higher cost (maybe $50) *years* ago. How much processing power does such a chip have? Not a lot. Making the protection algorithm any stronger would require a more costly, power-hungry CPU and make the DRM impractical. Furthermore, media formats for consumer electronics devices need to last at least a decade. Otherwise, consumers will perceive the format as a ripoff and not buy the media -- look at what happened to the original DivX format.

What happens when the wimpy DRM that is feasible on a portable consumer electronics device runs into a general purpose computer five or six years after the DRM format comes into use? Well, to start with, a general purpose CPU that was a contemporary of the original DRM chip probably had a couple of orders of magnitude more processing power. Then, the current general purpose CPU will benefit from three or four generations of Moore's Law, so add on several more orders of magnitude of processing power. The general purpose CPU will be able to crack the DRM hundreds of thousands of times faster than the little built-in chip. If that weren't enough, the DRM can be attacked by a network of multi-core CPU's operating in parallel, a la the SETI@Home project. Remember, the DRM has to be cracked just once and the content is accessible to everyone.

As time marches on, the economics of the problem just get worse. Consumers expect devices to get cheaper and thus less money is available to implement a strong DRM scheme.

The attitude within Apple towards DRM seems to be somewhere between thinly veiled contempt and sourness at a necessary but very disliked evil. No one will say it from Apple for the record, but if the record companies would agree to drop DRM requirements, the iTunes Store would switch to DRM-less AAC's in a heartbeat. DRM just raises costs for Apple, and provides little benefit. Apple makes its money on iPod sales, not on song sales. The cost of maintaining the FairPlay DRM code for song sales on the server and in iTunes/QuickTime is a drain that gives no benefit to Apple.

OK, I wrote that last paragraph a week before Steve Jobs himself posted his thoughts on Apple's website. Mr. Jobs himself has stated for the record that DRM does nothing for Apple, is not a part of a lock-in strategy, and would be history in a second if the record companies would agree to let Apple drop it.


--Paul