Optimizing streaming media workflows to reduce your carbon footprint, Part I

As awareness grows around the importance of reducing an organization’s carbon footprint, media workflow architects are increasingly interested in understanding the environmental impacts of the workloads they design. One overall resource offered by Amazon Web Services (AWS) cloud architects is the Sustainability Pillar of the AWS Well-Architected Framework (WAF), which provides design principles, operational guidance, best-practices, potential trade-offs, and improvement plans you can use to meet sustainability and carbon reduction targets for your AWS workloads.

General best practices help reduce the sustainability impact of your cloud workload by maximizing efficiency and reducing waste. Adding sustainability as a non-functional requirement can result in gaining more value from the resources you use and using fewer of them. This can also translate to cost savings on AWS as you pay only for what you use.

In some scenarios, making changes to meet sustainability and carbon reduction targets may require evaluating trade-offs with respect to traditional metrics such as uptime, availability, latency, and cost. This two-part blog series intends to serve as a primer for the full WAF whitepaper and provides a mental model of how to apply Sustainability Pillar general guidance to live and on-demand media encoding and packaging workflows running on AWS.

Part I of the series focuses on cloud sustainability, the use of managed services, AWS Region selection, and aligning SLAs with your carbon footprint reduction goals. Part II dives into considerations for energy efficiency in codec selection, and implementing AWS Media Services features in alignment with the Sustainability Pillar.

Understanding cloud sustainability

The cloud sustainability section of the Sustainability Pillar says that “when building cloud workloads, the practice of sustainability is understanding the impacts of the services used, quantifying impacts through the entire workload lifecycle, and applying design principles and best practices to reduce these impacts. The Sustainability Pillar focuses on environmental impacts, especially energy consumption and efficiency, since they are important levers for architects to inform direct action to reduce resource usage.”

The Sustainability Pillar also describes environmental sustainability as a shared responsibility between customers and AWS. In the shared responsibility model, AWS is responsible for optimizing the sustainability of the cloud, including delivering efficient, shared infrastructure, water stewardship, and sourcing renewable power. Customers are then responsible for sustainability in the cloud, which includes optimizing workloads and resource utilization, and minimizing the total resources required to deploy your workloads. In this blog post, we focus on the customer’s responsibility for optimizing live and on-demand media encoding and packaging workloads.

Using AWS Media Services

One of the Sustainability Pillar best practices in the cloud is the use of managed services. The whitepaper states that “using managed services shifts the responsibility to AWS, which has insights across millions of customers that can help drive new innovations and efficiencies” and that a “managed service distributes the environmental impact of the service across many users because of the multi-tenet control planes.” With that as guiding principle for this AWS blog, we will focus on the AWS Elemental MediaConvert, AWS Elemental MediaLive, and AWS Elemental MediaPackage managed services from AWS.

AWS Elemental MediaConvert is a file-based cloud transcoding service which allows you to process your media content for playback on a broad array of devices.

AWS Elemental MediaLive is a cloud-based live video encoding service that offers fast, reliable and easy-to-use delivery of high-quality live video streams without the need to manage infrastructure.

AWS Elemental MediaPackage is a highly scalable, video origination and just-in-time packaging service that helps video providers securely, reliably, and cost-efficiently package and deliver live video streams.

In the following sections, we’ll discuss trade-offs architects can evaluate and make when architecting workflows using AWS Media Services.

Choosing your AWS Region

One of the first choices to make as an AWS architect is to determine in which AWS Region to deploy your services. An AWS Region is a physical location around the world where AWS clusters data centers. The Sustainability Pillar explains that the choice of AWS Region for your workload significantly affects its key performance indicators (KPIs), including performance, cost, and carbon footprint.

To effectively improve these KPIs, you should choose Regions for your workloads based on both business requirements and sustainability goals. The AWS Architecture Blog titled How to select a Region for your workload based on sustainability goals goes into further depth on a mental model for considering compliance, latency, cost, and services/features as well as the carbon intensity of the electrical grid for a specific geographic region in which a workload is running.

Media workflows, especially those with live video signals, can be particularly sensitive to latency both for video contribution and distribution, so this is an important consideration to balance. Delivering streaming media using a content delivery network (CDN) with globally dispersed points of presence, like Amazon CloudFront, can help cache files closer to the end user, which reduces latency and aligns with the Sustainability Pillar best practice to optimize geographic placement of workloads based on their networking requirements. For more specific recommendations about how to configure Amazon CloudFront for media streaming, please refer to the whitepaper Amazon CloudFront for Media.

Aligning SLAs with sustainability goals

A key concept discussed in the WAF Sustainability Pillar is about aligning service level agreements (SLAs) with sustainability goals, which involves reviewing and optimizing workload SLAs based on your sustainability goals to minimize the resources needed to support your workload while continuing to meet business needs. We can take an example of live video encoding workflows, where often the default scenario is to choose the highest level of availability to minimize the chance of disruption for end viewers. For certain mission-critical applications, like live broadcast television channels, high-availability is a common requirement; however, non-business critical channels may be able to tolerate unexpected downtime without significant impact to the business.

In situations with SLAs that do not require high-availability and/or 24×7 streaming, users can also apply the best practice to remove or refactor workload components with low or no use within MediaLive by leveraging automation to stop channels when they are not in use either based on a timed schedule or when no input is detected, which can reduce resource utilization, associated carbon footprint, and cost. MediaLive also offers two channel classes: standard and single pipeline. Standard uses two encoding pipelines, one as a primary and the other as a backup in two separate availability zones (AZ). This configuration offers higher availability but also comes with a higher cost and more energy consumption. A single pipeline configuration means that if there are any issues with the primary pipeline, your stream and viewer experience could be interrupted. As noted in the Sustainability Pillar, highly available workloads tend to consume more resources and the best practice is to define or redesign SLAs that support your sustainability goals while meeting your business requirements, not exceeding them.

Another area to evaluate SLAs is with Amazon Simple Storage Service (Amazon S3) configurations and aligning with the best practice to use policies to manage the lifecycle of your datasets, which advises users to use Lifecycle configuration rules to tell the specific storage service to transition a dataset to more energy-efficient storage tiers, archive it, or delete it. Media architects can align to this best practice by implementing Lifecycle configuration to automatically transition media source files to archival storage classes, or delete them entirely, when frequent, millisecond latency access is no longer required. When access patterns are unpredictable, users can also elect to utilize the Amazon S3 Intelligent-Tiering storage class, which is designed to optimize storage costs by automatically moving data to the most cost-effective access tier when access patterns change.

Conclusion

In Part I of our blog post series, we introduced the AWS Well-Architected Sustainability Pillar and discussed how the generalized guidance can be applied through the lens of media by reviewing the sections on cloud sustainability, region selection, and by aligning SLAs to your sustainability goals. In Part II of the series, we review choosing video codecs and implementing service features while considering energy efficiency and carbon footprint reduction.