blog: Best of both- Designing hybrid networks for the cloud
Best of both- Designing hybrid networks for the cloud
By Dan Muchmore, Sales and Marketing Director, Tyrell
There are many good reasons why content providers are moving to the cloud. The cloud has the potential to deliver everything broadcasters currently do, with more control and much more efficiency. But broadcasting in the cloud doesn’t obey the same rules as a traditional broadcast system, so you need a technical partner who knows the difference.
The cloud is an aspirational space and it is full of potential. It can provide scale to various workflows, as well as more flexibility than traditional fixed infrastructures can ever dream of. It delivers an efficient SaaS model which can reduce capital expenditure, slash travel and support budgets, and centralise management of everything from content creation to delivery.
But while it is filled with possibility, the cloud is so vast that there’s a lot of room left over for uncertainty. Much of it is still to be explored and many of the skillsets required to design cloud workflows around a set of unique requirements are still being developed. Content producers openly talk about how difficult it is to guarantee latency and bandwidth, while its ability to deliver reliable and deterministic signals remains a worry for many.
Here’s the good news. All of these technical challenges can be overcome. We know this because we’ve done it.
The biggest blocker to reliable cloud access is, in fact, people.
More specifically, your people.
A different set of skills
Although many broadcasters are confidently working with technology vendors to push the boundaries of the cloud, most of these proofs of concept are built on existing SDI or IP networks. Building a traditional IP network is relatively simple; everything is well-defined and as long as the IP settings are right, everything will talk to everything else.
In the cloud, the rules are completely different and network designers need to build the entire infrastructure from scratch to truly embrace its scalability. A provider like AWS provides the same service blocks to every paying customer, but those blocks need to be put together in a way which suits your specific workflow, and in a way that can be tweaked and adjusted when requirements change.
This requires broadcast engineers to embrace new skillsets; it requires dev ops, full-stack developers and cloud engineers. Network infrastructures need rigid version control, and the need to be able to carry out iterative changes without having to rebuild the whole network.
The issue is that in-house engineers who already have the requisite knowledge to support these systems are not only rare, but they are also often prohibitively expensive.
Going to the cloud is not easy and as the broadcast industry incorporates this new generation of technology into its existing designs, it is increasingly important to get support from a System Integrator (SI), like Tyrell, who understands them.
Some crossover
Tyrell has been working in broadcast environments since it was founded in 2000, and it has been helping clients move into the cloud for several years. But cloud environments are rarely pure cloud; moreover, moving production wholesale into the cloud isn’t necessary to provide the efficiencies that content providers are looking for.
Cloud broadcast networks will always be hybrid environments. There will always be glass on a lens that feeds a camera; there will always be on-prem infrastructure to deliver signals into the cloud; and there will always be convertors to maintain the integrity of the signal path. For these reasons, Tyrell is committed to working with partners that can add value to all these elements with innovative and scalable products designed to embrace them. Traditional broadcast system integration is not the same as cloud system integration. Tyrell has expertise in the on-premise world but we are also pioneering cloud system integrations in media and broadcast.
Latency is still key
Other broadcast challenges can be mitigated with intelligent network design directly in the cloud.
Latency is still broadcast’s biggest consideration in real-time broadcast environments like live sports. This is still true wherever the processing is located but working hand in hand with clients in the field has enabled Tyrell to mitigate latency in several ways, with clever architecture design and cloud vendor relationships making the biggest difference.
Unlike Microsoft and Google Cloud, AWS has a big enough media portfolio to support real-time pipelines like these andhas services within its portfolio which actively help to minimise latency. Secondly, Tyrell uses not only use the bestencoders on-prem to convert the signals to SRT, but it ensures the infrastructure employs trusted broadcast products at the end of the AWS pipeline to create a private Content Delivery Network (CDN). This guarantees signal delivery to the set-top box at the lowest possible latency and works in a similar to WebRTC, except with the added benefit of being natively designed for mass delivery.
Guaranteeing uptime
Redundancy is also still critical. Traditional broadcast infrastructures often employ dual redundancy, but when designing in the cloud also means having to accommodate contingencies to upgrade AWS services and broadcast products which are hosted on that service with no downtime. Creating dual environments which allow users to phase products, services or platforms in and out independently are vital.
It’s a fundamental difference between system declaration in the cloud and traditional system integration because it’s not about creating main/failover networks but ensuring the automation processes can move and switch between both networks whenever it is required to do so. Dual cloud networks also enable users to upgrade elements and phase part or all of it into the other.
Cloud networks also allow SIs like Tyrell to continually monitor the AWS environment to identify potential issues. Because while IP packets are routinely dropped and signal architectures are naturally self-healing, it is important to understand how AWS errors apply to the real world and act accordingly.
Making the leap into the cloud
Transitioning into the cloud from on-prem environments is a big step, and it is something that many content providers are tentative about. But they shouldn’t be.
The cloud can guarantee low latency and robust streaming, more efficient delivery, and centralised control and management. But it is far from a traditional approach.
An experienced SI can help ease the way by creating a cloud infrastructure with equipment that you already trust, built-in network governance, and transparent workflows which do the job more efficiently.