Playout in the cloud
Comments from James Gilbert, CEO of Pixel Power
Is cloud playout ready today for the transmission of premium 24/7 live channels?
Before we start, I want to be sure we are all absolutely clear about what we mean when we talk about the cloud. To me, there are a number of stages in the evolution of broadcast technology:
1single function devices running in bespoke hardware – the norm for broadcast technology until very recently
2multi function devices implemented largely in software and running on standard computers, perhaps with the addition of application-specific hardware such as video I/O – a simple like-for-like replacement of 1 above
3functionality implemented entirely in software, which can be run on a virtual machine in a server farm environment such as an on-premise data centre – we call this virtualisation
4taking that virtualised software and running it not on site but in a remote data centre run by a third party – this is the cloud.
Step three is the really critical change, the point at which we achieve the dramatic shift in economics and flexibility. Virtualisation means we can make the best use of the hardware: a set of processors could be running a graphics engine at one moment, a transcoder the next. Provided you manage the virtual environment you can deliver a suitable service level with less hardware, consuming less power, needing less rack space and less air conditioning.
The difference between virtualisation and the cloud is only the location of the processing and storage. Cloud solutions depend upon virtualisation. Once you have virtual machine-ready software, you can run it in your own data centre or in the cloud. Virtualisation is inevitable, and for us at least it is well advanced.
The decision to move to the cloud, therefore, is not technical. It is a business decision. What are the relative costs, in terms of capital expenditure, operational expenditure and the availability of suitable support staff? What service level agreement can you get from a cloud provider? How practical and reliable is it to move content to and from the cloud?
To answer your question directly, no, I do not see premium channels migrating to the cloud anytime soon. The SLAs are not available yet.
I suspect the economics will not add up. Cloud computing is best for peaky demand: you spool up processors as you need them, release them to other users when you do not. A 24/7 premium channel is the exact opposite of peaky demand: it is more or less constant, so the supposed cost advantages of the cloud do not yet apply.
What are the chief benefits to a service provider of adopting cloud playout today?
The cloud gives you the capability of almost instant access to almost infinite capacity. So if you want to trial a channel, or set up a pop-up channel, a cloud solution could be the way to go.
Say you wanted to create Glastonbury TV, a constant feed from each of the stages, together with pre-recorded packages. A four channel television service that will only be on air for four days. That is a perfect cloud application, and indeed it would be economically unviable to do it any other way.
The same pop-up channel solution could be used across a wide range of services. A church or other house of worship might want sophisticated broadcast facilities for one day a week. A major global business might want to create a television service around its annual results or a big acquisition.
Niche channels might use cloud-bursting to add functionality for a short while. Such a channel could acquire a new programme and want to promote it heavily. If the playout channel did not support end credit squeezebacks the marketing budget might be used to add the functionality from the cloud for a week or a month.
What are the barriers to adoption?
Technically there are no barriers. At Pixel Power we can deliver it right now. Any broadcast vendor should be moving to all-software solutions which can be run in virtual machines, or they will be left behind. If the software can be virtualised it can be run in the cloud.
There are economic and psychological barriers, though.
Psychologically, broadcasters are really not yet comfortable not knowing where there content is. Letting it go to a third-party business which is not steeped in the law and practice of television is a hard thing to do. Indeed, broadcasters would need to check their contracts to see if producers and content providers allow them to ship programmes out to third party providers, or out of the country.
There are geo-political concerns about handing over content, too. The big cloud players – Amazon, followed by Microsoft and Google – are American. In Western Europe we do not have a problem dealing with America (and vice versa) but that is not the case everywhere in the world. The further you get from the cloud data centre, the bigger the challenges in terms of connectivity: you need a lot of internet bandwidth to get the content from the broadcaster to AWS and the channel output back.
The final psychological issue is the first economic issue. Broadcasters have been tied to capital expenditure for so long it is hard for them to change their way of thinking towards an OPEX model.
Indeed, it can be very hard to calculate just how much a cloud channel would cost. I would suggest it is probably a lot more expensive than we think. Couple that with the move from CAPEX to OPEX and the cost/benefit analysis is tricky.
Are the concerns of playout suppliers and broadcasters being addressed by vendors?
That is a tough question for me to answer. I hope that we are addressing their concerns, because we spend a lot of time talking through the issue and making it clear that there are advantages and disadvantages.
A number of prominent vendors, though, are feeding the frenzy that the cloud is the next big thing and we must all rush towards it. A lot of them were the same prominent vendors who told us we would all be watching 3D television by now.
The cloud is not something to get excited by for its own sake. It may offer economic and practical benefits for some or all of your tasks. Right now at Pixel Power we can offer software products which are virtualisation-ready and cloud-ready, and we can offer the support to help you understand when it would be best for you to make moves in this direction.
What are the challenges in delivering Ultra HD from the cloud?
Ultra HD simply means more pixels – 4k resolution, or maybe 8k – and/or better pixels – more bits per colour for high dynamic range and extended colour gamut.
If you have a software-enabled playout chain then Ultra HD just means more ones and zeroes to be handled. You need a few more processor clock cycles, a bit more storage space, and maybe an even bigger pipe between your office and the cloud. But other than that, Ultra HD is simply a matter of switching it on in the software.
Trial or experimental Ultra HD transmissions would actually be a good use of the cloud, particularly as no-one is yet clear whether it is going to be 4k, 8k, HDR, HFR or some combination of them all. Where Ultra HD content is available – maybe premium events at the Rio Olympics – a channel could offer them to viewers via a cloud delivery service, reverting to the existing HD hardware at other times.
What type of media organisation or application will move to cloud playout first?
I think this is bound up in the changes in broadcasting, away from the linear channel to the television screen in the corner of the living room, and towards the TV Everywhere concept. Video on demand is a natural for cloud services because it has unpredictable peaks and troughs. Both Netflix and BBC iPlayer are hosted by Amazon Web Services, for instance.
Next will come the pop-up and trial channels I have already described. Video conferencing and webcasting are already widely accepted and they are cloud services, so extending their scope is readily understood.
Will linear channels ever migrate to the cloud? It will depend upon economics. In Western Europe we already accept outsourced playout. Could it be that in five years’ time will the cloud service provider be the new playout centre of today? Or will today’s playout centre develop to become an industry-specific cloud provider, allowing it to deliver the service levels we expect, like five nines availability?
What technology do you have in this area?
We have IP based and virtualised products available right now for master control playout and branding. We have a rolling programme to move all our products to virtualisation-ready software applications. At IBC we launched the next generation of playout solutions, StreamMaster, which can run on a dedicated PC, as a virtual machine in a data centre or in the cloud.
The really important point about this is that an operator should not need to know where the process is running. The software should look and perform identically whether it is under your desk, in the basement or thousands of miles away.
We have other products which can be run in the cloud, including our core graphics engines and Pixel Factory, an automated content creation system. They are ideal cloud applications because they are not running 24/7, but when you want something from them you want it immediately. This can bring great benefits such as use-based-pricing.
Moving our product portfolio to the cloud also enables us to deliver demonstrations to customers in a more efficient and timely manner anywhere in the world.
How are your clients reacting to cloud playout?
I think there is a risk of an emperor’s new clothes problem here. The clamour around the cloud is so loud that it is easy to think it is something everyone has to implement. It really is not the case: cloud is just one way to deliver the necessary functionality. Just because it is shiny and new (in our industry, at least) does not mean that it is inherently better.
My experience is that premium channels are showing little interest in the cloud. We say this as a company that can deliver cloud functionality right now. So who’s up for it?
Given the barriers I talked about earlier, I see cloud playout becoming a reality in the two to five year timeframe, and even then only for niche channels, trials and pop-up services. Ten years ago the channel in a box was the big thing: is this same market now ‘channel without a box’ ?
Will premium channels ever go there? Will the economics ever deliver a sufficient benefit? We need to wait and see. At Pixel Power we can help you make that full transition to IP/Cloud based broadcasting right now, but are also happy to talk to you about the real benefits of all the hardware platforms that can support your playout requirements.