Author: Tor Blomdell, Head of Products
I’ve just spent a week in Seattle with the Microsoft Azure team at their Executive Briefing Center for ISVs. It’s staggering looking at the talent and knowledge going into the major cloud providers investments as they’re redefining storage, networking, and compute. However, the next frontier for unlocking innovation in the media space is also about improving industry collaboration; to make it easier to deploy hybrid cloud applications and to interoperate between cloud providers, as well as ensuring that the balance sheet investments in appliances can be re-used to ensure easier migration to the cloud applications. Solving this will open up for new customer experiences and hence the ability to monetize consumer eyeballs even further.
Live Media, and especially Video, is different compared to other workflows over cloud, when it comes to cloud transport and processing. With Live, nothing can go wrong, as missed or dropped packages in the workflow will leave you exposed to delivery issues such as audio clipping, or even worse a black screen which will yield bad consumer experience and lost advertising revenues. Further, the nature of live means that we’re always pushing the limits when it comes to latency as no one wants to hear the results from their neighbors or even worse through social media. This is evident looking at Amazon’s acquisition of Net Insight’s consumer streaming service, SYE, which not only will deliver a consistent streaming experience but also open up new advertising possibilities.
In my previous post, we concluded that last mile delivery is broken and things need to change in order to build a reliable, elastic, and secure media pipeline in hybrid cloud environments. But the question is what are the key aspects to solve in order address low latency, high quality, contribution and distribution?
First of all, let’s start with some of the basics; a) media, and especially video, is bandwidth heavy and thus compute, networking, and storage quickly becomes costly if architected in the wrong way, b) the quality of last mile delivery differs and the drive for cost and latency reduction puts enormous strains on the delivery pipes, c) rack space at high value properties is limited and thus the ability to carry traffic from these sites to multiple end points is crucial, and finally d) each vendor, cloud provider, and broadcaster have their own formats and ways of operating leading to interoperability issues across the board which implies even higher costs for operating live media workflows, creating tremendous inefficiencies in the media supply chain.
There are solutions for some of these problems out there, and at Net Insight we’re focused on solving some of these challenges in order to build an hybrid AV pipe that is elastic, secure, and flexible ensuring that media companies, service providers, and cloud providers can both offer and operate high quality media services over any infrastructure.
On-ramping to any of the cloud providers, they all offer dedicated lines (for example Azure ExpressRoute or AWS Direct Connect), however, all of them have similar characteristics with jitter and lost packages. These characteristics are resolved with retransmission (think MediaConnect with AWS) or Forward Error Correction to ensure a safe on-ramp.
However, just offering a reliable on-ramp is not enough when building low-latency workflows over cloud infrastructure, as we’ve experienced building out multi-regional infrastructure. You get drops in the VMs and as you move across your virtual private clouds, hence you need to secure your transports between your Dockers and Pods to ensure a safe transport even between your regions to ensure a safe transport.
One major concern when moving high value media over un-managed infrastructure is the question about security; moving from on-prem to cloud infrastructure, media companies are giving access to their most valuable assets, the content. A breach of a pre-release movie results in large revenue losses and the cloud providers have taken actions to secure their data centers and networks, through MPAA certification etc, which is also benefiting Live workflows.
Further, ensuring end to end encryption, secure signaling, firewalls, and rotating keys is good practice to ensure that you protect your assets throughout the services. But there’s more to reflect on; moving into the cloud you need to think through how you share and manage certificates, how you manage access control and entitlement to ensure that the right groups have access to the content and that you’ve got audit capabilities to understand what went wrong and what corrective actions you can take in real time.
Manageability and Programmability
Once the cloud is ingested into the cloud infrastructure, then comes the next challenge on how you manage the workflow in a consistent way. All the major cloud providers offer their flavor of multi-cloud orchestration and deployments, but ensuring integration and manageability into your live workflow is crucial so that ingest, processing, quality assurance, and delivery can all be managed in a coherent way. Thus you can ensure an unified workflow when handing of to your current appliances, another customer’s VPC, or ingesting into an OVP.
The importance of building a system based on lightly coupled functions, that are scalable, controlable, and monitorable is crucial to ensure end to end programmability and thus the ability to integrate into your current workflow.
One of the major challenges for any media company, is the investments sitting on their balance sheets. The ability to interoperate with multiple solutions and vendor and leaving no device behind is crucial to ensure a safe on-ramp to a new infrastructure.
However, here comes one of the major challenges with the cloud vendor community; interoperability. AWS were early with their support for Zixi, which offers a world leading protocol and a 500+ strong vendor community, they did then follow-up with adding support for RIST the ARQ standard promoted by VSF. Following, Azure and Alibaba promoted SRT and the SRT Alliance with it’s 300+ members. Google is still to announce their strategy for a safe cloud on-ramp.
What strikes me looking at the vendor community is that everyone is picking different protocols and solutions, while no one is offering a full support for the ARQ protocols and interoperability. Thus, limiting the ability for media companies to move their content into hybrid cloud deployments, and also risking their customers’ balance sheets. Even worse, it might lead to a fallback to the lowest, common denominator which would be FEC and RTMP flows, which is really not a good protocol or solution for high quality media over any infrastructure.
Future IP based media workflows will open up new innovation utilizing cloud technologies
At Net Insight, we’ve been solving the toughest media problems for more than 20 years, carrying the world’s most valuable content over any infrastructure and ensuring consistent workflows for our customers; that being distribution, contribution, or remote production. As cloud infrastructure moves closer to the Edge, looking at AWS’ Local Zones or the work that Edge Gravity has been doing for the last couple of years, it’s evident that media technology needs to adapt to a hybrid cloud infrastructure. The future holds a lot for the companies that can master retransmission technologies and cloud orchestration across any infrastructure; while ensuring elasticity, security, as well as monitoring and control at scale.
These are some of the challenges that we’re putting a lot of our engineering talent on and if you’re interested in joining the movement. Then please reach out and be part of unlocking the next level of media innovation (Net Insight)
Are you making the switch to IP in your studio? Here’s a cheat sheet for things to consider.