Summing up SME2009 - Where's technology heading?

HTTP Delivery and Dynamic Rate Adoption

The buzz on the technology side clearly was HTTP as delivery method instead of classic streaming protocols. Recently Microsoft released their HTTP-based Smooth Streaming product. Apple declared HTTP Streaming is the only supported way of delivering streaming video to the iPhone. With Adobe announcing an HTTP streaming technology of their own, every big streaming technology player is betting on an HTTP streaming technology. HTTP delivery has the advantage of looking similar to normal HTML traffic, and thus passing firewalls and mobile networks. As opposed to traditional streaming, the player client itself handles the request of the right data chunks for the given time. This takes out a lot of logic from the streaming servers. Because of that HTTP delivery of video-on-demand clips can be handled by normal HTTP servers with little to no logic. If content has been prepared accordingly in beforehand.

Another technological trend is the actual arrival of bitrate adaptation on the broad market. All major players now offer that in their current delivery formats. For Microsoft and Apple it is part of their new HTTP delivery offering. Adobe has integrated that into their RTMP streaming protocol, but their upcoming HTTP delivery won't probably miss out on that one. A caveat is currently that in all those implementations you will need to manually provide the server with all the different bit-rates. The actual news here is that dynamic data-rate adaption actually works now. Real Networks has had the technology for ages, but it never really seemed to work so well in real-world services. Of course also the demand for such a technology, with streaming coming more and more to the mobile space, has promoted the appearance on the market heavily.

Format Convergence

The big hope is of course that those different HTTP delivery specifications would converge to a single omni-compatible standard. Software vendor Microsoft is not going that road, keeping their specs closed and limiting their Smooth Streaming to their own IIS and Windows Media / Silverlight technology. The HTTP delivery release seems to tune in with their current attention-seeking for their Silverlight technology. Apple in contrast openly filed their HTTP Streaming specification as an RFC draft to the IETF. Apple being a hardware vendor tries to shift revenues from the streaming solution market to their sky-rocketing device sales with that move by trying to short-cut the development from a device-user's perspective. I don't guess that Adobe is going to follow Apple's initiative. But it might be that Apple's approach will become a de-facto standard being supported in most other video playback clients. And then we have to see how Adobe and Microsoft are going to react.
In theory the format convergence would be possible as H.264/AAC has been pretty much established as the de-facto industry standard. And that in spite of the yet unclear future of MPEG-LA licensing terms, in addition to the double-licensing of both codec implementers and content distributors. Perhaps there is a chance that we are going to have a de-facto standard platform for web video streaming in an foreseeable time. That would be a major breakthrough for streaming media. But we should be ready for some nasty surprises as well.

So when we take an overview of convergence today between devices, desktops and appliances, we can see that a top-down convergence is still hard to find. Sure most targets now support some form of H.264, but things like delivery methods, ad-standards and DRM-requirements are still very fragmented. Apple offers convergence inside the walled garden of their device and service offering, but of course the bigger the convenience difference between their stuff and the rest of the world is, the more value it creates for them. Apple does not have a big interest in bringing the general convergence development forward, as is visible in their non-presence in initiatives like DECE. Their marketing power enables them to. I am not sure if there is a lot to expect from Microsoft in that respect either. Adobe has done some movement towards opening up with the Open Screens project, but the process is slow and we have to see where they are heading to. There is a lot of strategic manoeuvring happening between the main players, to the disadvantage of small content-producers and users. But it could be worse, a monopolist position of one of the big players would be deadly for the market and very disadvantageous for the end user.

Open Source in Video Streaming

An effective and proven solution to accelerate development and consolidation on the streaming market is to openly develop extensible specifications and publish implementations under widely recognized Open Source licenses. Adobe made an interesting case of that with the introduction of their MPL-licensed Open Source Media Framework (OSMF). The player component in Flex is basically replaced by the OSMF component offering of a modular plugin framework that allows all kinds of third-party extensions to hook into the playback life-cycle. That would be useful to mix-and-match different statistics-, authentication- or ad-integration modules connecting the player to different services. This is of course currently limited to the Flash platform, but perhaps this interesting development catches on to other systems.

Another interesting thing that struck me on the Open Source streaming front is the business model of the streaming platform developer Kaltura. They offer a video delivery software that allows others to build own video portals. The offering of three different editions of their platform is very interesting. There is a completely open community edition, a for-pay supported premium edition with proprietary modules, and a hosted version of that, which takes away most of the duties of running a video platform from the customer. This range of "totally independent" to "totally care-free" is a very interesting Open Source business model. The platform is developed in php and uses Flash as premier target client platform which probably targets the broadest range of the market on both sides of the wire. This can definitely facilitate smaller businesses to enter the streaming market with their own video site while the solution grows as they do. Kaltura advertises that it adds another option to the basic build-or-buy decision, an intriguing idea.