Time to put primetime’s head in the cloud

TV Connect

Software-defined video processing, virtualization and cloud operations mean less time and cost building infrastructure and more time improving services. 

https://knect365.com/media-networks/article/9895cd3f-2ce6-4afb-92b4-eda0fdd6f070/time-to-put-primetimes-head-in-the-cloud?utm_campaign=Share+Widget&utm_medium=Article+Share&utm_source=twitter

Despite the pioneering efforts of companies like Discovery to take traditional broadcast operations into the cloud, broadcasters and pay TV operators have been reluctant to migrate from the digital TV head ends that have served them well for the past two decades.

Understandably, the largest operators are unlikely to abandon their billion-dollar investments in their own infrastructure overnight. There are, though, signs that more are willing to embrace cloud-based approaches as, in many cases, they already have cloud experience from other video workflow functions and business operations, often in multiscreen services.

“Nobody should be investing in hardware/head ends,” says Paul Stapley, MD at cloud services provider Nordcloud. “The CapEx, RoI and fixed nature of these investments means that your audience will be moving faster than you are. Remember, iPads were not available seven years ago and, in that time, they have become the regular device for watching catch-up and live content, with the internet acting as the delivery mechanism.

“Was your head end ready for that change?” he asks. “Will your next hardware/head end be ready for the next, as yet unknown, technology shift? Whatever it is, cloud will provide a more appropriate infrastructure on which to build new solutions than a hardware approach.”

Virtualised hardware has provided tremendous economies of scale against conventional on-premise solutions in multiscreen TV roll-outs. Public cloud networks, in particular, provide a competitive landscape for media companies that do not own local infrastructure to use pay-as-you-grow compute, storage and networking for their OTT services.

Outsourcing infrastructure should be an attractive proposition to most organisations. Avoiding CapEx and the cost of maintaining internal systems can help an organisation become leaner and more focussed on their core competences. Similarly, the time needed to add channels can go from months to days, providing a new competitive workflow to reach consumers using OTT web access.

However, concerns remain about moving the entire broadcast workflow – which still accounts for the bulk of television revenues – into a cloud-based operation. Traditional broadcasting technologies were specifically designed and optimised for delivering content on a large scale to locations where internet connectivity was unavailable – all from a central broadcast location.

“By design, the standard broadcast head-end systems were also intended to ensure (close to) 100% uptime, given that failure would impact all, or the majority of, viewers,” explains Steve Huin, VP of Strategic Partnerships at content protection solutions provider Irdeto. “The challenge is to create a guaranteed, smooth and secure migration path from traditional broadcast head ends to its cloud-based equivalent. All while providing high-quality, highly presented playout without black to air.”

The move into a virtualized and cloud-based operational model requires new technologies, new ways of working and organisational change. Legacy systems, workflows and attitudes are a hurdle.

“The vast majority of video functions within the typical live and linear workflow have remained on-premises because of the optimised scale and performance of available products and also a very established operational model,” says Keith Wymbs, CMO for AWS Elemental. “These services are mature, yet highly premium, carrying significant content and advertising revenue in many cases.”

Broadcaster and pay TV operations teams have always kept their broadcast infrastructure away from their IT counterparts and emphasised the skills and knowledge differences between the two. This has led to two separate organisations. The horsepower to run the compression algorithms has until recently only been available from dedicated hardware.

In-house video network engineers and operations teams are used to a physical video network that they can see and touch and that have been operationalised with monitoring, reporting and diagnostics. Virtualizing video workflow into private and public cloud raises concerns by staff about how they ensure it’s working, reporting and diagnosing problems quickly to minimise mean-time-to-repair.

This is changing as new technologies arrive on the market. For example, virtual network routing has vastly improved to allow uncompressed video payloads to be transported over IP. GPU compute power can now be virtualized and is also now being made available by leading cloud providers.

There is of course no one-size-fits-all cloud. Media organisations have valid business reasons for needing to explore a mix of private and public cloud approach. Since the majority of current workflows for live and linear television have not yet embraced any cloud approach, and use dedicated appliances and applications in broadcast facilities, most cloud infrastructure providers and solutions vendors support all options.

AWS Elemental, for example, supports private (virtual data centres), public and hybrid instances reporting that dozens of its customers have deployed its software in their own private cloud. However, the company argues that the full benefits of cloud can only be gained in the public cloud.

“Building video workflows on AWS public cloud creates the most dynamic, efficient and high-performing business operations for anyone in media,” says Wymbs. “This flexibility allows for experimentation, innovation, spinning-up of new channels and packages, the creation of new on-demand offerings – and all with a ‘pay-as-you-use’ approach reducing business risk. It has no comparison with traditional approaches to video infrastructure.”

An important part of the cloud movement globally, and across different industries, is having access to on-demand, pay-as-you-go cloud compute, storage and application resources, rather than planning, building and operating owned cloud infrastructure.

“Public cloud takes the expenditure model beyond OpEx to pay-as-you-go: no more need for capital or fixed payments,” underlines Huin. “It provides elastic compute and storage.”

For linear playout, for example, the increased agility offered by cloud means there’s no need to build out permanent infrastructure, so new channel variants and operational models become possible.

For example, around key events such as Glastonbury or FIFA World Cup, a broadcaster may want to spin up one or several pop-up channels for a short period of time, pay for them only when they are used and then tear them back down with no ongoing cost. Doing this conventionally would cost millions of euro.

The same can be said for additional disaster recovery facilities, which become flexible in a way that has not been possible before. Here, virtualized assets can be copied to a public cloud environment, spun up and then played out, with the broadcaster only paying for actual use.

One critical difference between physical and cloud distribution is scalability. Physical networks tend to focus on maximising performance and capacity of equipment. Public clouds, on the other hand tend, to focus on smaller, efficient and flexible instances that can scale with replication.

The largest cloud providers offer multiple regions and availability zones, allowing organisations to mirror operations in two separate geographical regions and also more locally within a country, state or city. Media and metadata can be automatically mirrored across these data centres, meaning that anyone can take over operations in real-time for truly mirrored services. This dispersed, yet connected infrastructure on-demand is unique to public cloud provision.

AWS Elemental stresses a clear direction within the industry towards microservice-based video applications that can be deployed in private or public cloud infrastructure. Microservices – or componentised software modules – perform a certain function that scale independently and can be assembled and managed independently.

“The ultimate focus is to have public cloud native services that can scale on demand, and be leveraged in a true pay-as-you-use model,” says Wymbs.

Many of the key functions used in pay TV and broadcasting can be deployed in the cloud (both private and public) today. These range from ingest and storage to scheduling and playout, compression – including statistical multiplexing – encryption with conditional access (CA), quality assurance, monitoring and then modulation and broadcasting (or direct IP delivery).

The fundamental benefits of cloud that of financial gains from an OpEx model are the same for broadcasters and pay TV operators, but there are some differences between them. Pay TV providers, for example, need to encrypt their content services and add in DRM. They also tend to be a magnitude larger than broadcasters in terms of the multi-channel scale of their operation.

Live workflows are more challenging than non-live, from a cloud perspective, because of the way content is captured, the delivery architecture and the need to function in real-time, 24x7. Even here, AWS Elemental says its customers are able to deploy its on-premise appliances close to the live sources as well as leverage AWS Elemental Cloud for core video processing functions.

The need for lower cost production and distribution models is unarguable and requires infrastructure that is programmable and agile to enable broadcasters and pay TV operators the ability to continuously adapt to new market demands and compete with Silicon Valley’s internet-based giants.

It would be foolish to ignore the challenges, both technical and organisational, in migrating to the cloud but it seems we have now reached the point where public cloud, and the vendor ecosystem that must reside in it, are ready for primetime linear broadcasting.