When we last talked to Dilithium Networks is was a one product company, focusing on real time transcoding within cellular networks to facilitate video conferencing between different network types. But that was over two years ago and when checking in with them this week, Dilithium had a very different story to tell.
That experiment into quite complex calculations at the server level in the network gave Dilithium some experience of what it took to get things to happen quickly in a network.
Call set up times were around 10 seconds for a video call before Dilithium came along and set about speeding up the transcoding. Part of the reason for this was the time it took to transcode the video from one codec type to another, for instance from a H.263 compliant codec carried on a cellular data channel to a H.264 video carried on a pure IP backbone. But the other part of the time delay was the finding out what each of the devices preferences were in the first place. The round trip to ask about codec types and physical terminal characteristics was eating around half of that delay, and with delays like that people just weren’t going to bother with video conferencing.
But as any company named after a famous Star Trek technology will tell you (the Dilithium Crystals in the Enterprise engine), just because a task is difficult is no reason to give up, and one of the outcomes of working in real time networked video was that Dilithium came up with the idea of handling the interrogation and declaration of device preferences while the handset was still ringing.
After all, the calling device knows it wants to have a video session and the network knows the nature of the device that it is calling, so it is a matter of
transmitting preference messages, which contain information about what kind of channel to set up for video and audio. This resulted in Dilithium partnering with Packet Video to author a new ITU standard called MONA or Media Oriented Acceleration Negotiations, and any device which supports MONA can now set up in superfast time, around 1 second.
A MONA capable device, either cellular or fixed, begins establishing a session by transmitting a set of Preference Messages, which contain information about its channel establishment capabilities and preferences. As soon as the devices receive these, a simple handshake process decides which is the best set of options to proceed with, based on the best communication capable between both devices, a sort of lowest common denominator approach.
All this to send video quickly across multi-layered networks, but still the video telephony market is relatively slow, so now Dilithium is trying to think up ways of triggering video experiences from mobile handsets, which will drive revenue for operators and use this technology. In effect now that it’s made video easier to send to a handset it’s looking for a killer app because that’s not video conferencing, at least not yet.
What it’s come up with is its VT-Ring video ringback platform and a video service creation environment called VIVAS.
Mitch Lewis, Dilithium Networks the general manager of marketing, had to explain video ring tones to us three times. Why would anybody want to announce their arrival in a call by sending a video to their friend’s phone? But apparently in places like Indonesia and China, the idea is catching on fast.
A web based bulk standard Intel or Linux server sits in the network and stores as many video tones as you can beg, borrow or upload, and another application server handles SIP (Session Initiation Protocol) sessions. The MONA technology does its stuff, and then there are three components in the loop. The two handsets and an IP attached server speaking through the operator’s mobile gateway. In under 1 second the right video has to be selected, the handset being called needs a MONA interrogation, and the right codec version has to be identified, and then the video transcoded to that codec in real time and streamed across the network. The server then has to drop out of the conversation and leave the two handsets still in touch with one another.
It hardly seems worth it to send someone a video that it supposed to represent them in some way, and you can have fun thinking up all sorts of scenarios whereby the wrong video arrives at the handset. Lewis confirmed that yes people keep ten or more videos and use them for different people or put them on random selection. We’d hate to see what would happen if the video intended for your wife, perhaps a tasteful selection of baby shots of your first born, instead arrived at your bosses phone. There are worse scenarios and we’re sure you can make up your own.
We don’t get video Ring tones, but then again personally we never got the original musical ring tones market and that is a multi-$1 billion market today.
‘The set up fee to enable the handset can be $5 and the service can be $2 a month and then there can be a charge, say $1 to download a new video and add it to your site,’ explains Mitch. In effect this is one special instance of the new service creation environment, one that Dilithium has already created.
The Vivas environment uses a new version of the video gateway that got Dilithium its start in life to handle transcoding, but one that can handle 1200 simultaneous transcodes instead of 120, and a media server and media gateway all in one, plus a SIP signalling server.
‘We have other applications we have made with Vivas. In Dubai the police have cellular links built into security cameras, which take video snapshots of suspects and send them to a remote face recognition system.
‘Less ominously Singtel is using the same capabilities to allow parents to dial from their cellphones to existing security cameras at dayschools and check on their children.’
In Indonesia the Vivas environment is being used for video mail and messaging, and the operators there have carried 3 million video streams to date, while Vodafone and KPN are supporting the sending of videos to news services or operating TV video call ins, Lewis said.
‘The next steps are applications like video calls to a technical help desk so that the caller can show video of just what’s happening with whatever devices that need fixing, and ensure that if an engineer is needed they take the right parts,’ says Lewis.