Your browser is not supported. Please update it.

5 October 2005

Is the DRM cavalry riding out of the Sun?

We opened up both barrels when we heard in August that Sun Microsystems was introducing DReaM, an open source DRM system, firing off the headline ‘Sun’s oxymoron, an edOpen Source DRM, and opining that it was too little too late.

Sun gave it a big shot this week at convincing us that we were wrong, saying it’s not too late, that the most robust security technology is robust because it’s open, and Sun claims a big patent hoard in DRM, a pedigree that goes back to the turn of the century and it says it will have a DRM for mobile phones in 12 to 18 months.

Sun also raised the thought that eventually the Java SIM card in a cellphone will be the basis of all future DRM identity management and has already had one major DRM win with part of Sony’s PSP rollout.

First off, Sun’s Glenn Edens, Senior VP at Sun for Media, Entertainment and Telco walked us through the debacle that has been OMA 2.0 DRM.

‘The OMA is one of the main instigators for us doing the DReaM project,’ said Edens (it stands for DRM everywhere available).

‘When the fuss over the MPEG LA license terms blew up, Sun was approached by one of the large carriers and asked if we could help. We have a lot of patents in DRM, hell we were writing Rights Expression Languages before anyone had a name for them. They asked ‘Can you give us an alternative?’ So we prepared a paper called ‘Mother May I?’

This is a child’s game in the US and elsewhere and the name relates to the fact that mobile phones are always attached to a network, so there can always be a dialogue, backwards and forwards for authentication and permission.

‘If you take a look at a lot of these patents that are being pushed as being essential for OMA, there’s a lot of bullying going on. We had to ask if some of them could be overturned with prior art which had not been discovered at the patent filing, and whether any of that might be ours. We also had to ask if those claims really represent what the carriers are being asked to license. I mean often, with a little bit of re-engineering, you can work around a patent, and use someone else’s technology,’ said Edens.

These counterproposals led Sun to jump into the DRM world, and while it won’t have an alternative cellphone DRM for as long as 18 months, Edens is abundantly clear that Sun will have a product here. ‘If a cellco needs to make a decision now, then we don’t have a product, but if they are thinking of using DRM in 12 to 18 months, then we will be playing.’

But back to that issue about being too late. Surely content is exploding onto mobile phones now and content owners won’t release content unless it is securely protected? Well one strategy may be to continue to use OMA, and continue to argue with the patent holders. Patent licenses have often taken two years to settle, MPEG 2 took about that long, and if at the end of it two sides don’t agree, then the Sun solution will be ready.

Edens said, ‘between you and me that’s one strategy we’ve heard,’ and chuckled.

Sun kicked off into DRM with a European Eurescom project started in 2001 and reported on in 2003, funnily enough called OPERA, where it worked with DMD Secure, Exavio, SDC AG and T-Systems and some European operators. The inappropriate name (Opera is a browser company) came from InterOPERAbility of DRM technologies.

SDC and Sun built a system that was based on the Java SIM card found inside a mobile phone, and had Windows Media DRM authenticated with RealNetworks, and RealNetworks authenticated with SDC’s client and all of it using the Java based SIM.

From there to DReaM is not too much of a step, except that Sun’s done this before, using open source to make its systems more robust.

‘We are the same place in DRM as we were at when Sun founded the Liberty Alliance. Everyone then said that Microsoft’s Passport had already won the market for single sign on and federated identity. DRM is now reaching enough consumers that they are beginning to get frustrated with the systems they are using,’ said Edens, implying that therefore they were ready for someone to build a system with consumes in mind.

‘DReaM has now gone out to an Open Media Commons under the OSI-approved Common Development and Distribution License (CDDL), which is effectively a way to get discussion on what we’ve built so far, and we’ve contributed DReaM, and Sun streaming Assembly code.

‘We’ve made sure that a lot of smart people get to look at what we’ve done. The secrecy in a DRM is all in the keys and the authentication process, not in how the system works. By making that open and separating it from the keys, it will become more robust.’

Edens’ philosophy is to make it less complicated than other DRMs and to separate the DRM both from the hardware architecture and from the player itself. Most DRM’s are, as we have said in the past, married to a player and a hardware configuration.

‘We don’t want to authenticate a device, we want to authenticate a person,’ he said.

The key to this, thinks Edens, is to pursue the SIM card, and one outcome could be that DReaM is embedded in SIMs and becomes the basis of a person’s identity. ‘A phone tends to go with people wherever they go, the Java SIM has never been hacked, so it’s a secure identity. What we need to add is fraud management, so that if a SIM card is switched off and hacked in the future, the behavioural mechanics of how it is used after, give away the fact that it has been broken.’ So Sun is thinking up ways to record both online and offline security audit trails.

‘We have already built a prototype PKI based Conditional Access system and a Personal Rights Management prototype, and contributed that to open source, now we are working on a prototype using a phone as your identity, to see how good that would be.’

This opens up a huge new approach to DRM and it must be said that Sun is the right company to at least initiate this work. Imaging using your phone to tell a TV system or your PC, that you are who you say you are, and so you are entitled to access to certain pieces of content. It’s been mooted before, but potentially, in Sun’s hands, it has a chance of coming off.

‘Sure if you are someone like BSkyB you’re not going to throw away the Conditional Access system that you’ve got right away, but eventually over the next ten years, you are going to want to federate your customer’s identity with other companies and other services, and that’s when our approach will come into its own.’

A stolen phone can be reported and all of the content relationships can be denied, also when it is turned on, it could have its secret identity overwritten remotely. ‘A stolen phone can be spotted in a day using pattern recognition fraud determination, so even if the owner doesn’t know it is gone, systems can work out that it is being used by someone else,’ said Edens.

Sun has already had one major success in DRM, which to date has not widely published and that is its involvement in Sony’s PSP non-game content downloading service run by Sony’s So-Net ISP in Japan.

‘That uses Sun servers for content delivery and a simple conditional access scheme which relies on a secret key encoded into the hardware of a PlayStation Portable,’ said Edens. The service is mostly used to sell video onto PSPs over an internet line, and is in test mode right now.

In our first article on DReaM we also said that Sun should join the Coral Consortium on DRM interoperability. ‘DReaM will be Coral Compliant,’ aid Edens. ‘We were involved in the creation of Coral at the beginning and began to drift away from it as we realized that it was a single vendor calling all the shots (meaning Intertrust). The license was unpriced and yet we were asked to sign an NDA that had an implied patent pooling element in it. In order to see the Intertrust documents we were required to sign the NDA and what didn’t belong to Intertrust, seemed to belong to Microsoft.

Faultline can vouch for the fact that even we were asked to sign an NDA for Coral, which we declined.

‘I think we’ve been involved in getting Coral into a fit state where a standards organization can begin to look at the specification seriously. We also sit on any number of standards bodies and can help introduce it to them, the IETF, the ITU, quite a few others, whichever one Coral ends up going with.

‘The next version will be out in another couple of weeks and then we will see it go to a standards body. But IP patent pooling is still not clear and it needs to be clear before anybody can use it. However we do expect to see some Coral compliant DRMs coming out once this second specification comes out,’ said Edens.

With Sun’s reputation for open source and solid engineering, perhaps we have to revise our view that it is too late to make an impact on DRM and perhaps it really is the cavalry arriving just in time to save the day. But for now first generation DRMs are still made up of proprietary systems with expensive royalty schemes, at least for the next 12 to 18 months. Let’s hope it’s not too late.