One of the most startling suggestions at NAB 2013 came from Quantel, suggesting that we are making a false step in moving to file-based architectures. The company’s principal software architect, Dr Richard Cartwright, thinks that instead we should look to the way that Google relies on hyperlinks, and regard our content as “an internet of frames”. “Every frame is valuable” is the mantra, so every frame should have its own identity, with the devices that use those frames relying on links between them. As a practical example, this makes simple the ability to work at multiple resolutions, with content stored in different places as browse resolution, SD, HD and potentially in the future 4k. In an “internet of frames” solution, an operator or process would identify the frames required either by their content or by their links to other frames – they come next in the sequence. The device might pick up the browse proxy first, then fill in the required frames from higher resolutions at the point of delivery. “Files are the wrong kind of abstraction for our industry,” said Cartwright. “The do not work well with large amounts of data. The IT industry is moving towards cloud-based storage of blocks of data, taking advantage of the ‘scaliness’ of the internet. That needs a different identity model.” SDI is a streaming format, but so is HTTP, he argues, so why do we need to block move files? Thinking of each frame, in each resolution, as an individual entity – with its own URL – opens up the prospect of “just in time” video production. It depends upon open standards if it is going to work: the implementation in Quantel’s own RevolutionQ and QTube products depends upon the MXF AS02 format. If widely adopted, it should make interoperability and systems integration easier.
