Public Cloud is Ready for Live Production at Scale

NAB

Delivering and monetizing the live experience reliably at scale has always been one of the hardest challenges that media operators, broadcasters and content owners face but technology providers maintain that their expertise and cloud more broadly is now ready for prime time.

article here

“Delivering live streaming is relatively straightforward, but delivering it at scale, in the highest quality, and with latency equivalent or better than traditional broadcast is where the challenge comes in,” MediaKind says in a white paper, “Live without Limits: Streaming at Scale.”

MediaKind says that live content is somewhat behind the transition compared to non-linear content in its migration to public cloud, “and it is likely that much of the very top tier of live production will remain on-premises in a more dedicated environment for many years.”

However, as the ability to produce live content in public cloud matures, it becomes the obvious way for adding flexibility to production capabilities: “no more limits on production due to the number of studios — simply create them on-demand and tear them down again afterwards.”

This fits well with a remote production approach, in turn minimizing both infrastructure and operational production costs, thus enabling a wider range of content to be made available.

“Originally, high-availability and the custom connectivity from broadcasters to their transmission infrastructure were the main reasons to retain it all on premises. However, as availability is now similar to on-premises, and highly-available connectivity is readily possible over IP (for example via SRT — Secure Reliable Transport), there are few remaining arguments as to why the entire chains for both streaming and traditional broadcast cannot be delivered through public cloud.”

MediaKind makes a case for the viability today of public cloud as the infrastructure for hosting the biggest live sports events.

Among other benefits, the use of public cloud to build production and publishing workflows as needed can radically reduce deployment times, the vendor says.

“Building new processing capabilities using on-premises physical infrastructure can take weeks or more, whereas creating a new channel or production environment can be done in minutes, including monitoring. Automation through orchestration is the way to make this timely and reliable — orchestrate all the components to be instantiated exactly when they are needed, connect them, and go.”

The automation and version management through Kubernetes means that “it is possible to replicate the exact environment with certainty,” which is a prerequisite for being able to use automation to instantiate and tear down media applications and services with confidence.

Yet public cloud environments are very different from the more traditional on-premises fixed-function world. Cloud technology requires a new set of skills that are not closely aligned with the traditional broadcast approaches, meaning either training or recruitment are a prerequisite to making a transition if expertise is remaining in-house.

Hence the pitch for media organizations to work with MediaKind or other tech partners.

“While it is certainly possible to work in a best-of-breed manner, it comes with significant and recurring costs in terms of engineering and operations. Therefore, it will be increasingly typical to leverage vendors’ expertise to deploy and maintain code in public cloud environments, with more of a sub-system approach rather than individual components.”