Virtualization Meets Video Processing at NAB 2014
If anything is certain about the video business, it’s this: the volume of change is daunting and every change tends to make life more complicated, not less.
This is certainly true at the sharp end of the business – digital video processing – where “multiscreen” video, new video formats and new video technologies are together creating a perfect storm of complexity. Once there was SD over MPEG2 delivered to TVs. Now there is SD, various flavors of HD and, soon, 4K; and MPEG2, AVC and now HEVC; plus a wealth of encapsulation schemes and DRMs; And even more screen sizes and resolutions as the number of device to be supported grows ever larger.
The number of permutations of all these options is truly dizzying. Every permutation is a potential video “workflow” to be implemented – and the number of permutations is expanding rapidly, apparently endlessly and it’s exponential. Today Cisco deals with some media companies that have over 80 video workflows for their content. One more video format – for instance 4K – and this potentially doubles to 160. Another compression scheme – HEVC perhaps – and now we have 320. And so on.
Keeping track of all these “workflows” is one thing, but actually implementing them using today’s approaches is even another. For, today, separate workflows are most often implemented as separate “production lines” with each line having its particular separate set of specialized equipment that must be separately bought, scaled and managed. It does not take much imagination to realize that this approach is simply not practical going forward. It is expensive to buy, install, maintain and scale many discrete workflow elements. As bad, this level of complexity means that the people Media companies, broadcasters and service providers hire to make them money by delivering exciting programming to viewers are not doing this at all – instead they are spending all their time installing, configuring and managing pieces of equipment.
It does not take a rocket scientist to see that as an industry we have to find a better way!
And this is where cloud technologies and virtualization come in. These technologies make things better because virtualization is best understood as enabling any piece of hardware to do any one of a wide variety of functions. The “personality” that the hardware takes on is determined by the software that is loaded onto it. So instead of needing to buy, manage and scale separate pieces of equipment, in a virtualized environment it is only necessary to manage a single pool of hardware.
In a traditional environment, workflows are often instantiated as a chain or rail of servers. This makes it especially difficult to create new workflows as each workflow needs a set of physical hardware. In a virtualized environment, each workflow is created in software and it is therefore far faster, easier and cheaper to create new workflows. It is also easier to automate the creation and elastic scaling of workflows. In a virtualized environment no new single purpose physical hardware is required to create new workflows and scaling is accomplished by allocating more hardware from a central pool. Virtualized environments automate the task of allocating new hardware, instantiating software on the hardware and connecting the allocated hardware into workflows. This frees staff from the drudgery of workflow creation to focus on making money for the company.
In sum, virtualized environments reduce capital and operations costs, make it far faster to create new revenue producing services and free staff to concentrate on money making activities rather than on installing, configuring, maintaining and scaling hardware.
All of the above is generally true of virtualized environments. These advantages have obvious applications to the challenges media companies and service providers are facing in the area of Video Processing. So we at Cisco have been working on applying virtualization techniques, which we have pioneered for networking applications (see our announcement of the Evolved Services Platform, here), to video processing environments.
At NAB we are announcing our approach to applying the advantages of virtualization to video processing. We call it Videoscape Virtualized Video Processing, and it consists of three elements:
- Our Virtualized Video Processing Portal easily configures even the most complex workflow from a single screen.
- The Virtualized Video Orchestrator allocates software and hardware resources from a common pool – which can be a mix of physical and virtual resources – to satisfy each workflow request. It embeds business and technical logic to instantiate, as needed, additional software and to allocate additional hardware.
- The processing workload for each workflow is performed by the hardware and software elements of the V2P. Each piece of hardware, whether it is a performance optimized appliance or virtualized into an industry standard server, is multi-function; its functions are determined by the software elements instantiated by the Virtualized Video Orchestrator.
As we developed this strategy and discussed it with our media and service provider customers one thing became apparent – our Virtualized Video Processing must be evolutionary not revolutionary. More simply put, our customers wanted a solution, which would leverage the equipment they had purchased rather than obsolete it. For this reason our V2P can be built from our popular AnyRes, DCM and 9036 families.
I hope this blog helps explain our V2P announcement at NAB. The number of video processing workflows is going to continue to increase. V2P will, we believe, help manage this complexity so as to reduce costs, free staff to focus on money making tasks and increase the speed and agility of delivering new money making video services to market.
Like so many things in our increasingly virtualized lives you need to learn more about it. Come by our NAB booth N9332. We’ll gladly walk you through Virtualized Video Processing!