AWS for M&E Blog

Stream with Evertz SaaS playout platform

This blog was coauthored by Martin Whittaker (Evertz), Jeremy Blythe (Evertz), and Noor Hassan (Amazon Web Services).


Driven by industry-wide transformation, broadcasters, content owners, and content creators continue to reinvent ways of creating and sharing content in pursuit of greater efficiencies and faster and more effective processes. Evertz, an Amazon Web Services (AWS) Partner, recently announced; a multi-tenant Software as a Service (SaaS) platform that provides cloud video services for content owners and creators. is built using AWS and is a validated solution through the AWS Foundational Technical Review program.

In this blog post, we’ll touch on how not only serves the traditional broadcast playout market, but is built to natively support the creation and management of Free Ad-based Streaming Television (FAST) channels. Keep reading to learn how adoption of cloud-based SaaS video services benefits broadcasters, content owners and vendors alike.

What is is a multi-tenant SaaS platform with discrete applications that allows customers to stream, transform, or engage with content. engineering fully manages platform maintenance, backend engineering, and cloud scaling, allowing customers to focus their resources entirely on operations.

In this blog post, we focus on the newest application available from the platform, Stream. Stream provides customers with all the features of enterprise-class broadcast playout systems within a fully managed SaaS architecture, with user-managed channel configuration and pay-as-you-go pricing. Channels can be tailored for over-the-top (OTT), FAST, or traditional broadcast with a choice of features, including playout of live sources, video and audio mixing, comprehensive multi-language audio with captioning support, and full graphics and branding capabilities. playlist user interface

FAST TV channels are generally available free for viewers to watch without a subscription and are supported by advertising revenue. In recent years, the market for FAST channels has exploded with modelling by nScreenMedia predicting that ad revenue generated for FAST channel providers in 2023 will reach $4.1 billion. Stream is designed to capitalize on this expanding market, allowing content owners to maximize revenue and achieve the monetization potential of their assets. OTT and FAST channel launches can be driven from pre-built channel recipes for services like Roku, Samsung TV+, YouTube TV, and Tubi, making it easy and affordable to stream specialty or even hyper-targeted channels that would previously have been too costly or too niche to launch.

Why is designed based on lessons learnt from deploying complex and large-scale broadcast systems in the cloud and with a new approach to managing channel configuration. eliminates the need for vendor setup by giving users control over channel configuration and eliminating the undifferentiated heavy lifting of setting up underlying resources, allowing customers to focus on the technology.

With automated channel orchestration, once you’ve configured your channel it can be spun up at the press of a button. It takes between 5 to 10 minutes to go from cold to channel playout. If you want to turn that channel off, again it’s just a button press and manages spinning the channel down.

Updates to channel configuration are managed through the user interface. You can make an update and publish it to the channel and in the background manages changes and provides a transparent blue/green deployment of your channel without the need for downtime.

With, you can deploy in minutes rather than weeks or months. channel configuration wizard channel configuration wizard

What are the key challenges a SaaS can solve?

Improved time to market

Thanks to the pre-established infrastructure of the platform, customers can move straight to configuration without time-consuming hardware implementation. Customers can also improve the speed of their configuration by leveraging pre-defined templates and ‘recipes for success’. Channels can be configured and on-air in under 15 minutes. Overall, reduces engineering effort and enables additional focus by customers on up-stream business systems and content monetization.

Scalability, agility, and flexibility
Stream customers have full control over channel configuration and channel launches, source and destination routing, service regionalization, and user management through the user interface. This means that customers can grow and scale their business if they see success or quickly shut down services if they don’t. This leads to simplified capacity planning and allows tenants to make use of the collective engineering know-how embedded in the SaaS.

Resilience and security
Nobody likes downtime, but in traditional software models unless you’ve heavily invested upfront in a dual system, redundancy downtime for upgrades and maintenance is inevitable. solves for this in a several ways. Geographic resilience is built into the core of the multi-tenant SaaS architecture, with a simplified update and upgrade procedure with a blue/green deployment model. This means tenants no longer need to manage and schedule downtime. Upgrades happen seamlessly with no downtime. incident stream user interface engineering  

All in on AWS

Evertz made the decision early in the design of to go all-in on AWS. We wanted to take advantage of its full power and scalability by going cloud-native.

To offload the undifferentiated heavy-lifting of managing infrastructure, our approach has been serverless-first where AWS Lambda, AWS Step Functions, Amazon DynamoDB, and Amazon EventBridge are the key components of our microservices. However, working with video is challenging and can involve long process times or continual 24×7 run times. Our second option is containers and we use AWS Fargate and AWS Batch for processes that are not a best fit for Amazon Lambda. When we need 24×7 run time and particular hardware specifications, we turn to Amazon EC2. is built from multiple independently deployable microservices, each of which deploys into our development, test, and production environments via AWS CodePipeline.

Following is a high-level view of the architecture: architecture diagram on AWS

Here you can see the three classes of microservice: Serverless; Serverless with a longer runtime component; and 24×7 or specific hardware requirement. Video enters and exits via AWS Elemental MediaConnect flows and low-resolution stream previews straight to the browser via Amazon Kinesis Video Streams. All assets, schedules, as-run logs etc. stay in the customer’s account under their control. is granted secure access via AWS Identity and Access Management (IAM). architecture diagram on AWS per channel

Channel origination is one of the “serverful” parts of the architecture. For each channel this consists of three running servers:

  • Automation Service: Manages the playlist, provides an API to the UI, and calls the API on the Playout Service
  • Playout Service: Produces the output stream, interacts with the Flow management service for video i/o, and reads assets directly from the customer Amazon S3 bucket
  • Intelligent Monitoring: Detects and alerts on stream output mismatching the schedule

High-performance engineering

SaaS provides benefits to customers and vendors alike. It enables the engineering team to follow the practices and principles of high-performance teams: high deployment frequency (<daily), low lead time for changes (<an hour), fast recovery, and low change failure rate.

Complex customer-hosted systems tend to stagnate on old versions with batches of changes released at a frequency measured in months. The customer decides when and if they wish to upgrade and since batched changes are higher risk, the tendency is to avoid them.

In, we push small changes from trunk into production, multiple times a day. Engineers get immediate feedback while their change is still fresh so they can make further changes as needed right there and then. The risk is much lower and the reward much higher. observability diagram

Observability graph showing 12-hour period in the production environment

The previous graph, from our observability data, is a typical 12-hour period in our production environment. Each CI/CD deployment is captured as a marker (vertical pin). Observability is a fundamental tenet from the Operational Excellence pillar of the Well-architected framework. Our code is instrumented using Open Telemetry and fed to, an AWS partner. Honeycomb is our primary engineering interface into Its powerful querying, visualization, and monitoring allows our developers and Site Reliability Engineers (SREs) to understand how services are performing and to troubleshoot when necessary.

When deploying a service within the channel origination chain, stream orchestration launches a new or partial chain alongside the live one. The new chain is synchronized with the old and two identical streams are provided to the outbound AWS Elemental MediaConnect. If the Intelligent monitoring and health checking systems determine that the new chain is viable, MediaConnect is instructed to switch over. Finally, the old chain is terminated.


In this blog we explored key drivers for SaaS Playout solutions, and how the platform leverages cloud-native best practices to enable customers to accelerate workflows on AWS.

We recommend you read more about patented technology for video observability as discussed in this blog post, or watch how easy it is to set up a channel in the platform in this blog post.

You can also reach out to the team to learn more and test drive the platform yourself.

Noor Hassan

Noor Hassan

Noor Hassan - Sr. Partner SA - Toronto, Canada. Background in Media Broadcast - focus on media contribution and distribution, and passion for AI/ML in the media space. Outside of work I enjoy travel, photography, and spending time with loved ones.