Tuesday, 15 June 2010

Do we really need to transcode video in the cloud?

I hear a lot of talk in the market about the challenge of format proliferation in streaming video delivery. The thinking goes that the number of devices that can stream video - phones, TVs, set top boxes - is growing and so content distributors must transcode into an increasing number of formats to service them. This creates a mare's nest of connections and formats to service consumer demand.

I wonder whether we should also consider another world, where devices increase in processing power and application platforms become standardised across categories of devices (e.g. Android on TVs). In this world, distribution becomes a trial of bandwidth at the edge and possibly also in the core and the devices shoulder the burden of making viewing a rich experience.

A similar thing happened when the mobile web was young. Operators spent tens of millions on hardware and software to enable web pages to be viewed on phones. It didn't take long, however, for phones to gain the processing power to repurpose pages, standards to be developed for designing pages and browsers to reach the handset. It seems to me that telecoms always try and answer emerging demand with big, centralised technology and the answer almost always turns out to be light-weight software that consumes their bandwidth!

For me, the same seems likely to happen with video. There will always be new devices. Creating new 'cloud based' format transcoding regimes for each is just dumb. Right now, if I was a retailer or broadcaster wanting to service non-browser-based demand, I'd be thinking very carefully about how much capital I was going to commit to it. If someone else is willing to do it in the short term, for opex-only outlay, then best to consider it. On a short contract!

Monday, 14 June 2010

Amateur analytics

Reading an article on Microsoft putting Office into the cloud brought 'Directions on Microsoft' to my attention. I'm very used to the idea of analysis firms pointed at particular market segments, but had never encountered one aimed at a particular company.

The more I thought about it, the more this made sense to me. It occurs that the market of today is akin to science at the turn of the 20th Century. There is so much to discover in so many fields that the traditional 'expert' commentators on a field cannot hope to see everything, let alone consider the resulting implications. This effect is becoming more marked as (particularly technology) companies continue their acquisition sprees, bringing many more divisions and products into their portfolios.

Perhaps just as gentlemen astronomers provided great insights into the cosmos, small, pointed analysis firms can help us make sense of the hyper-dynamic, information-saturated market of the twenty-teens.

Something else that worried me about this line of thought was how difficult it must be for a CEO to keep track of everything his or her company is doing, let alone the moves and counter-moves of the competition. A colleague of mine has emphasised many times to me the importance of Management Information reporting and how neglected it is within businesses today. Perhaps the production of rich dashboards will become a key strategic business function in the next decade. Something I'll take up in passing with the clients I meet over the coming months.

Exploiting the intelligence of screens

I've just read the Pew Research Center's annual survey on cloud computing. As ever, it contains plenty of thought-provoking commentary from experts. One that really caught my attention was this quote from Nokia's Davis Fields:

"It's 2010 and I could already basically use only cloud-based applications on my computer. Local storage is already increasingly irrelevant – I have my all my photos stored on Flickr, my address book is in my Gmail and I've got all my emails stored there as well. Apple will likely move iTunes online in the next few years, and streaming movies from Netflix will eliminate the need to download movie files. I use Microsoft Office and Photoshop out of familiarity as my main two desktop apps, but good alternatives already exist online. I predict most people will do their work on ‘screens connected to the web,’ There won't be any sort of ‘computer’ anymore."

Now, I agree with the sentiment that every screen will be connected to the web and the experience on each will be interchangeable - each will be an extension of the other and a portal into data hosted elsewhere. The oversimplification is that each screen is, of course, a computer in its own right. Google TV's tie up with Sony and Intel to produce connected TV's is an example of this. The TV, like the mobile phone, is becoming an interactive computer that enables rich use (as opposed to the computers that perform operational tasks behind the scenes today).

I suggest that the next step in the dynamics of the cloud will be to aggregate up all this spare computing power and use it to form part of the flexible resource. The Seti project and similar initiatives started a wave of this in the early part of this Century, but I can't help feeling that a more market-based approach where resources are requested and drawn dynamically based on 'weather' could open up new opportunities. Could there be money to be made in enabling your spare processor resources to be fed back up to the 'grid' when the machine is idle? As with power generation at home, the actual revenue made is tiny, but psychologically it incents users to participate and may therefore add to the resources available at incremental cost.

Just a thought. Might be far fetched!