Your browser is not supported. Please update it.

16 November 2005

Sun IPTV juggles switches, identity, Jini, DRM and mountains!?

Sun plans to weld its work on switched fabric processors, its new Streamstar video servers, its work on identity servers and its plug and play software, Jini and a brand new set top architecture into a comprehensive IPTV strategy in 2006.

But to give you an idea of how big Sun is thinking, first we want to ask a question.

What’s the biggest mountain range in the world? OK, bow your head in shame all those that picked the Himalayas, home of the largest mountain, Everest. In fact it’s the Andes and by some huge margin.

The Andes is around 4,700 miles long, spanning seven countries, and including some of the highest mountains on Earth. Over 50 of the Andes’ peaks are over 20,000 feet high.

So when Sun Microsystems decided to initiate a project called Andes, but spelled backwards as Sedna, it barely disguised the scope of its ambition. Sedna is expected to be huge, but what exactly is it?

Well according to the SunLabs web site, where only meager descriptions of the company’s modern architectures are issued, Sedna is about architectures and technologies for future data center interconnects that scale to thousands of compute nodes and to bandwidths of tens of terabits per second.

Talking once again to Sun’s Glenn Eden Senior VP at Sun for Media, Entertainment and Telco, he has a better use for Sedna, and believes it can become the basis of an irresistible IPTV strategy that no other company has yet thought of.

To be honest Sun will have to come up with something pretty outlandish and original if it is to unseat the leaders in the IPTV space, notably arch-enemy Microsoft, starting as it does from such a long distance behind, but Eden as always is keen to remind us that Sun has managed this in the past.

‘We are very interested at SunLabs in figuring out what carriers will need in the 5 to 7 years time frame. It takes that long to take an idea through to becoming a platform and on to testing, provisioning, deployment and converting all that into new customer revenues.

‘The telco space is pretty clear. They need more revenues, they need to keep their existing customer base and they need to ‘kill the other guy’ whoever that is, and they are going to have to do that with digital content, running on IP transports, that are cost saving and flexible, which implies web services and service oriented architectures.

‘When we looked at the nature of a big complex TV head end and the digital version of the central office, we realized that they weren’t that different, the control plane and the bearer are merging, IP does that for you and creating a head end becomes mostly a switching problem.’

This work clearly came out of Sun’s efforts to offer virtual data centers using newly emerging grid and utility computing models.

‘What Telcos really want is an array of processors, assigned one task one minute and the next minute it can be assigned another. What Sun thinks they want is a big switch with a set of hierarchical storage, a combination of both disk and tape, with a bunch of digital to analog converters attached to a big fat pipe.

‘If you think about what a telco will do in the future, it will acquires one analog signal, convert it to digital, transport it via IP and then convert it back into an analog for delivery over the last mile.

‘The digital plant needs to drive the physical plant in the last mile access with whatever signal, be it DSLAM, a QAM modulator or any other radio transmitter or a GPON translator. The rest of the problem is a software problem.’

‘You have to think about how we can create real time delivery for things like voice and in the future for things like stereo telephony (Sun says it is doing a lot of work there) and for messaging and all media services.

Sun sees Sedna is a replacement for the big public metropolitan switches, with one city sharing a server which both acts as a head end and a softswitch, or even 2 or 3 cities sharing one. Sun also believes there would be neighborhood fiber nodes, in an emulation of a web cashing architecture.

‘We’re trying to design systems that would let a Telco distribute power to anywhere it needed it in the network, but we are also very concerned about the costs of such a system,’ said Edens.

‘Sun uses a proximity I/O system which lets chips talk to each other across small distances using capacitive coupling, so that wiring is not needed. This means we can have 1,000s of tiny couplings with each one running at 10 Gbps., which creates a non-blocking fabric, where the only possibility of congestion is in the output.’

‘That kind of switch fabric is a switch designer’s dream. We can get a 10 terabit non-blocking switch the size of a 2 drawer filing cabinet with our Sedna technology.’

The advantage that Sun has in this market is that 80% of the world’s softswitches sit on its solaris operating system already, and it has OEM deals with Siemens, Nortel and Cisco in this area. But the fact that this is still only at prototype stage and without a full understanding of where this would sit in, for instance, the RBOCs existing IPTV strategy, which will be mid-rolled out before this research hits the market in 18 months, still makes it difficult to gauge the potential impact of Sun in the media market.

‘When you add that kind of switching power to our set of applications for messaging (Verizon uses Sun for its messaging apps) and our strength in identity management, then you begin to see the building blocks that Sun will build its IPTV strategy on,’ said Eden.

Quite clearly Sun is building up to a series of announcements during 2006, when this switch fabric and other technologies come together to form the rudiments of an IPTV and Telco strategy.

But Eden also talks about messaging and billing platforms and there seems to be something of an overlap between the types of additional software based services that Eden is suggesting and IMS, the IP Multimedia Subsystem.

‘IMS is a set of protocols for marrying the old world with the new IP world. If we were further along the transition for IP, and if we had web services worked out more, IMS would never have happened, and in 5 to 10 years from now it will be just a collection of web services. Soap, SIP and IMS will all become one and the same thing,’ he said.

‘And the telco world will be a battlefield between Microsoft’s .Net and J2EE.’

Another prototype that Edens is happy to talk about is what he refers to as Streamstar.

This is based on Galaxy servers with the next generation of multi-threading Sparc processors in them. ‘We plan to integrate with a number of different software vendors and begin to deliver video,’ he said.

This technology is understood to be based on technology Sun gained when it bought Kealia last year, a company run by Andy Bechtolsheim, another of the original founders of Sun Microsystems.

Sun gave demonstrations of the product with Irish interactive television provider Digisoft, at Supercomm in Chicago earlier this year. Co-incidentally all of the Digisoft interactive TV products are based on Java.

At the time Sun said that this prototype would soon find its way into tests being carried out by Telcos in the US, Asia and Europe. This sounds more like a conventional video server and will find itself up against the usual suspects in IBM and Hewlett-Packard but also the more established video specialists, Seachange International, C-Cor and Concurrent, as well as BitBand and of course any company that offers the Windows 2003 servers which drive the Microsoft IPTV strategy.

Eden’s view of the digital home also suggests that Sun’s has the potential for a role there. ‘In the home there are a different set of problems. There are a lot of content companies, consumer electronics businesses and carriers focused here and we believe Sun’s role is to build a bridge between the home and the mobile phone and other portable devices such as those that will play in automobiles.

‘Sun can take on the DRM issues, and it can help bring self install, self provisioning and device discovery using Jini, and we can also help with user interface issues.’ Sun recently talked us through how it plans to use the Java SIM card in a mobile phone as the basis a federated identity scheme for DRM.

And Jini, which grew from early work in Java on distributed computing, is designed so that computerized appliances can automatically join a network of other devices and services. So within Sun’s home device strategy Eden said that Sun will offer machine to machine technology so that devices can sense each other and says they are working on exciting prototypes in this area.

‘We also have a set top box prototype built around Broadcom’s system on a chip, We’ve not decided yet on whether to make these ourselves or to license the architecture to other players. The device runs OCAP (OpenCable Applications Platform which is middleware for interactive TV services) and MHP (DVB equivalent of OCAP) and DVB.

‘We can make this extremely low cost for standard definition TV and we have a HDTV version in prototype,’ said Eden.

Sun is also recognized at the codec level having recently joined the cluster of businesses that hold essential patents within the MPEG-4 standard and it is a member of the Internet Streaming Media Alliance (ISMA) pushing video streaming technology.

Add all that together and Sun is justified in saying that it has all the building blocks. But can it build an edifice worthy of the Sedna name? We won’t be sure of that until at least the end of 2007 when all these products will be on general release.