Home Analysis Orchestration is the tech-wonder that turns software-defined video into cloud agility

Orchestration is the tech-wonder that turns software-defined video into cloud agility

image1 (9K)
Share on

Rear-of-server-for-web.jpg
Software-based video processing, virtualization and orchestration underpin the TV operations revolution (Pic: Kynny & iStock)

 

An operations revolution is just starting in television, although some platform operators and broadcasters have already begun to transform the way they process, manage and deliver television. The primary ambition is to increase agility so ‘traditional’ video providers can compete more effectively with born-in-the-cloud challengers like Netflix. The new operations paradigm is enabled by software-based video processing, virtualization and orchestration, with the latter unlocking the true power of ‘the cloud’, like the ability to harness compute and storage resources anywhere, avoid under-used hardware infrastructure (and wasted CapEx)  and scale capacity up and down as required.

Multiscreen viewing is possibly the biggest driver behind the trend towards virtualized video processing and the orchestration that goes with it. But it is not the only one. As Brett Sappington, Director of Research at Parks Associates (the market research and consulting firm), points out, “On-demand consumption is also important. On-demand services began to push operators to deliver video in a different way, but it was multiscreen video that really forced the issue.”

Few broadcasters have talked publicly about their transformation plans but Globo, the free-to-air TV broadcaster in Brazil and one of the largest content producers worldwide, views the virtualization and orchestration of network functions and workflows as a way to increase business agility. Last November the company launched Globo Play, a multiscreen offer with VOD and simulcasting and it has been implementing a unified infrastructure to support and streamline different video workflows, reducing complexity and cost.

This project highlights how television operations are starting to converge with IT. The broadcaster is using content management technologies that will be familiar to many multiscreen operations teams, but also a virtualization and orchestration platform from Cisco called V2P (Virtualized Video Processing) that abstracts video processing software from dedicated hardware appliances and moves them into data centers. Globo is also using Cisco Unified Computing System (UCS) servers and Cisco Nexus switches as part of the deployment.

V2P represents a completely new kind of television technology, a management platform for matching software applications to the most suitable compute, storage and network hardware resources. These resources can include dedicated hardware appliances in a traditional headend rack, designed and optimized to run specific software, and generic servers in a data center (on-premise or off-premise). The orchestration platform knows exactly what resources are available at any given moment, where they are, what jobs are already running on them and where future tasks can be assigned. For management purposes, all the different instances of a resource are treated as a single, common asset.

If you are virtualizing video processes to a data center, you need to do more than randomly assign a task to any of the servers present. One of the important capabilities in OpenStack, the open-source orchestration operating system, is the ability to recognize and take advantage of specific platform features in a data center where there could be multiple generations of servers, with increasingly advanced features on the newer ones. One example is only handing jobs to servers with advanced data security.

Enhanced Platform Awareness (EPA) in OpenStack enables deployment of virtual machines (VM) onto server platforms with the specific hardware and silicon capabilities that are optimal for the particular needs of the VM. Despite their apparently generic nature, some server systems can contain specialized silicon and application-specific accelerators designed for different types of workload.

Cisco points out that using the traditional operations model, each screen and form of video requires a separate video production line, using optimized hardware, hard-wired together, often with proprietary interfaces. They often contain the same core functions, like encoding or recording. A product like V2P helps improve TV agility because these separate production lines can be consolidated.

As Gaurav Rishi, Director, Product Management and Business Development at Cisco, explains, you can stitch together the functions needed to create service use-cases. If you are dealing with fully virtualized functions, each function is a software application running on some generic hardware. Creating a new video workflow is as simple as selecting options in the V2P orchestration interface. “You avoid the need to ‘stand-up’ the new services. You avoid equipment-sprawl and problems of manageability, and you improve time-to-market,” he says. “This approach tackles the heart of the problem.”

Rather than ‘device-chaining’ you start ‘service-chaining’. Cisco cites a recent deployment where a customer used existing general purpose compute and storage infrastructure to ‘stand-up’ 12 linear channels, then test one of the channels with encryption before encrypting them all without any site visits. Live-to-VOD processing was then added simply by switching on some software, all using the orchestration platform. “No new hardware or specialist appliances were required,” Gil Cruz, Architect for Video Software at Cisco, points out. “This is a game-changer.”

Other traditional headend vendors, like Harmonic and Ericsson, have their own orchestration and virtualization platforms and this looks like a new competitive battleground for vendors. Tom Lattie, VP Market Management & Development, Video Products at Harmonic, says: “We are seeing content and service providers transitioning to virtualized infrastructure to re-architect their workflows and combine previously discreet media processing stages.” Among its customers, Harmonic has seen a focus on uniting the playout and encoding stages to simplify workflows.

Headend vendors will continue to compete on the basis of ‘point-solutions’, meaning individual functions like encoding or statistical multiplexing. At the same time everyone is going to have to cooperate to ensure their function-software can be integrated into different orchestration solutions. This could place an emphasis on the ‘openness’ of any virtualization and orchestration platform and their APIs. Service providers will want the ability to cherry-pick ‘best-of-breed’ software and integrate their own software into this environment.

Matt Smith, Chief Evangelist at Anvato, which provides a multiscreen solution for automating live video capture,
editing, publishing and syndication, explains how this new approach to video operations could affect the vendor ecosystem. Media companies could use Anvato for live-to-VOD requirements, including clip-creation for instant availability via social media. The company’s software can be run on appliances or virtualized on COTS (common off the shelf) servers.

“More people are asking to take our software stack to run on a server,” Smith reveals. “We are currently working to be integrated with orchestration platforms. It is a way to hook into cable and broadcast infrastructure and provide a seamless OTT capability, running our software on their choice of hardware. For us it is way to grow our business. We do OTT and we do it well. We think this model will emphasize a best-of-breed approach.”

 

Interested in the TV operations revolution?

This story is an edited excerpt from the new 6,000 word Videonet report: ‘21st century TV operations: Ready for anything’, which you can read here, totally free.

You can read about the motives driving the transformation in TV operations including agility, the crucial role of orchestration and how video workflows can be created and consolidated, how virtualization is ready for ‘mainscreen TV’ as well as multiscreen, the potential to unify operations around adaptive bitrate streaming and the different ways in which the new operations environment will be hybrid.

The report considers how companies like Comcast, Globo, nc+, Magine, TVPlayer, Sky and Voo are using virtualization or ‘cloud’ technologies and includes insights from Parks Associates, IDC, Cisco, Ericsson and Harmonic, among others.

View or Download Report

 


Share on