Raj Yadav, Regional Sales Head for South Asia, Imagine Communications
All the talk in the broadcast industry at the moment is about IP. The move to IP connectivity is extremely important, of course, but in truth it is one step on the path toward the ultimate goal.
The real win comes when we move to a broadcast architecture which is fully software-defined, running all the functionality on standard, relatively inexpensive commercial-off-the-shelf (COTS) hardware. The advantage we get from this is a completely flexible infrastructure: we assemble the functionality we want for the workflow we are handling. For the next workflow, we assemble different functionality.
To gain this benefit, we have to take a completely fresh look at the way we design systems. In the past, because we needed bespoke hardware to meet our needs, the industry built monolithic devices which performed rich functionality. Those devices were then built into rigid systems which were tailored for the workflow at the time of the design, but were not flexible enough to do much different.
In a software-defined architecture, we can treat each small process as a separate device, running as a software application on standard hardware. Because it is extremely compact and is designed to do just one thing, we call them microservices.
The system orchestration layer simply calls up the microservices it needs for a workflow and, as soon as they have completed their task, they are released. So processing cycles are minimized because you are not wasting energy running functionality you never actually need, and licensing costs are reduced because you can share functionality across multiple workflows at different times.
Agility is increased because if you want to do something new, you just have to add the relevant microservice or microservices. In the past if, say, you wanted to move from H.264 to HEVC delivery, you would have to unbolt a big, expensive box labelled encoder from the rack and put it on a shelf to gather dust, while you bolted in another big expensive box, most of whose functionality – apart from the output codec – was being performed perfectly well by the original.
Now, in a microservices architecture, you simply licence the HEVC output codec and the upgrade is complete.
We have eliminated those big, expensive boxes taking up rack space, power, and air conditioning, and replaced them with software running in virtual environments. The Zenium microservices environment from Imagine Communications encourages third parties, and even users, to create microservices to work alongside those from Imagine, to create precisely the functionality you require.
This virtualization also makes the most of the hardware. The processors in your data centre are also dynamically allocated to run microservices. Unlike traditional broadcast architectures, you do not have online waiting for equipment for the moment you need it: you simply have a pool of computing and storage which is used as required.
Which brings us to the cloud. Microservices are essential for efficient use of the cloud, where you are paying for processing time. Minimizing the use of resources, while knowing it can scale to whatever you need, is the central benefit of the cloud.
When you are talking to vendors who may claim their software is cloud ready or even cloud native, ask them if it is a microservices architecture. If it is not, then it is most definitely not cloud native, and is unlikely to deliver the full advantages promised.
The most recent broadcast technology study by the IABM and Devoncroft Partners found that 85 percent of broadcasters thought they would be using the cloud in the next 2-3 years, with 28 percent already using it. The cloud is not a magic cure-all, but with the right software products – cloud native and built on a microservices architecture – it will deliver unprecedented levels of flexibility, scalability, and efficiency.