Discovery Builds Direct-To-Consumer Video Platform with SOUTHWORKS and AWS
Executive Summary
Discovery Inc serves passionate fans around the world with content that inspires, informs, and entertains, reaching 220 countries and territories with more than 8,000 hours of original Programming each year across a range of content genres. With consumer demand for on-demand video increasing, Discovery looked to build a new technology platform that would enable rapid innovation and rich direct-to-consumer experiences.
Challenge
With decades of valuable programming in its archive and nearly 1,500 hours of linear and live content delivered daily, Discovery brings a wealth of assets to a streaming video offering. Previously, the company focused its streaming strategy on bundled offerings, striking carriage deals with service providers to distribute its content. In 2019,Discovery Inc broadened its streaming strategy to build and operate a direct-to-consumer platform of its own, taking advantage of its unique content portfolio and dedicated fan following. For its first foray into direct-to-consumer live and video-on-demand (VOD) streaming, an all-new video technology infrastructure would be required.
The new content workflow would need to be reliable and ready to scale quickly, without sacrificing high quality. It would need to be flexible to consolidate hundreds of channels and thousands of on-demand assets into a single workflow. And, it would require reaching outside the organization for the expertise and resources to meet the project’s key requirements in a short timeframe and make the most of audience demand.
Solution
Discovery selected SOUTHWORKS, a Software Development firm and Amazon Web Services (AWS) Partner, to design its new streaming content infrastructure using AWS Cloud services. SOUTHWORKS reviewed Discovery’s key requirements for reliability, scalability, flexibility, and time-to-market and decided to build the workflow around AWS Media Services, a set of managed services dedicated to transporting, preparing, processing, and delivering live and on-demand video content in the AWS Cloud.
Discovery’s live streaming workflow ingests video sources from content providers or media storage over IP Networks directly into the AWS Cloud using AWS Elemental, an Amazon Web Services Company Elemental MediaConnect, a cloud-based transport service designed for live video. MediaConnect pushes live streams to AWS Elemental MediaLive using Real-time Transfer Protocol (RTP) with forward error correction (FEC). MediaLive then encodes and generates multiple versions of the source video as Adaptive bitrate (ABR) outputs to address the full range of viewers’ screens and device types.
Medialive provides live streams to AWS Elemental MediaPackage in the HTTP Live Streaming (HLS) protocol, using MediaLive’s output locking feature to ensure the video pipelines are synchronized. MediaPackage prepares live streams in different distribution formats including HTTP Live Streaming (HLS) protocol and DASH, creating new streams on the fly in response to requests from viewers’ devices in a format and bitrate optimized for the consumer’s screen, player, and internet bandwidth.
Reliability was a critical requirement for Discovery Inc. SOUTHWORKS specified a dual-pipeline workflow architecture, duplicating resources for video transport, processing, and packaging across two AWS Availability Zones to create redundant video streams with real-time failover between resources. SOUTHWORKS and Discovery took this resilient architecture a step further, duplicating the dual-pipeline workflow within a second AWS Region, again with redundant resources across AvailAbility, Inc Zones. This level of redundancy was specified to maximize the availability of Discovery’s live video workflow, even in the event of a service outage at the regional level.
SOUTHWORKS’s approach to achieving resiliency extended to operational systems. To help ensure continuous, uninterrupted viewing for active live streams, the firm engineered a playback and failover console as a single page web application leveraging APIs built with AWS Lambda and Amazon API Gateway and hosted in Amazon Simple Storage Service (Amazon S3). The SOUTHWORKS team also developed comprehensive monitoring dashboards, enabling Discovery’s digital operations team to quickly identify, respond to, and resolve issues.
AWS ELEMENTAL MEDIA CONNECT : -
AWS Elemental MediaConnect is a cloud-based video transport service provided by Amazon Web Services (AWS). It simplifies the secure and reliable transfer of live video streams between different locations and broadcast facilities. MediaConnect offers a flexible and scalable solution for broadcasters, content creators, and enterprises to transmit high-quality video content over the internet.
Here are some key features and benefits of AWS Elemental MediaConnect:
Reliable and Secure Transport: MediaConnect uses AWS's global network infrastructure to ensure reliable and secure video transport. It leverages multiple redundant connections and protocols to minimize packet loss, latency, and jitter, resulting in a stable video stream delivery.
High-Quality Video: MediaConnect supports the transport of uncompressed and lightly compressed video signals, enabling broadcasters to maintain the highest quality of video content throughout the distribution workflow. It also supports a wide range of video resolutions and formats, including 4K Ultra HD.
Flexible and Scalable Architecture: With MediaConnect, you can easily create and manage video transport workflows to suit your specific requirements. It provides the flexibility to add or remove sources, destinations, and processing elements as needed. The service can scale up or down to handle varying workloads and peak traffic demands.
Simple Configuration and Management: MediaConnect offers an intuitive web-based console for configuring and managing video streams. You can quickly set up sources, destinations, and the desired network flow between them. The console also provides real-time monitoring and troubleshooting tools to ensure smooth video delivery.
Enhanced Collaboration: MediaConnect allows multiple users to simultaneously access and work on the same video streams, making it ideal for collaborative video production workflows. It supports secure and granular access controls, enabling different team members or external partners to contribute to the video production process.
Integration with AWS Services: MediaConnect seamlessly integrates with other AWS services, such as AWS Elemental MediaLive, AWS Elemental MediaPackage, and AWS Elemental MediaStore. This integration enables you to build end-to-end video processing and distribution workflows using a combination of AWS services.
Pay-as-you-go Pricing: AWS Elemental MediaConnect follows a pay-as-you-go pricing model, where you pay only for the data transferred and the network resources used. This allows you to scale your video transport infrastructure based on demand without upfront commitments or long-term contracts.
Overall, AWS Elemental MediaConnect offers a robust and scalable solution for transporting live video streams, making it easier for broadcasters, content creators, and enterprises to deliver high-quality video content across different locations and platforms.
RTP PROTOCOL :-
RTP stands for Real-time Transport Protocol. It is a network protocol specifically designed for real-time transmission of multimedia data, such as audio and video, over IP networks. RTP provides mechanisms for the efficient and timely delivery of real-time data, ensuring that audio and video streams are transmitted and received with minimal delay and loss.
Here are some key characteristics and features of the RTP protocol:
Payload Agnostic: RTP itself does not specify the format or content of the data being transmitted. It is a generic protocol that can carry various types of multimedia payloads, including audio, video, and other real-time data.
Timing and Synchronization: RTP includes mechanisms for timestamping and sequence numbering, allowing receivers to reconstruct the correct timing and order of received packets. This ensures proper synchronization of audio and video streams.
Packetization and Framing: RTP divides the multimedia data into smaller packets for transmission over the network. Each packet contains a payload, along with RTP headers that provide information about the packet's timing, sequence number, and other control information.
Error Detection and Correction: RTP includes a checksum mechanism to detect transmission errors in received packets. It does not provide error correction itself, but it can be used in conjunction with other protocols, such as Forward Error Correction (FEC), to recover lost or corrupted packets.
Quality of Service (QoS): RTP allows applications to specify different levels of priority and QoS requirements for their real-time streams. This helps prioritize the delivery of time-sensitive data and ensure a consistent quality of experience for the end-users.
Multicast Support: RTP supports multicast transmission, allowing a single stream to be delivered to multiple recipients simultaneously. This is particularly useful for scenarios where the same multimedia content needs to be distributed to a group of receivers, such as in video conferencing or live streaming applications.
RTP Control Protocol (RTCP): RTP is often used in conjunction with the RTP Control Protocol (RTCP). RTCP works alongside RTP to provide feedback and control information about the ongoing transmission, including statistics about packet loss, jitter, and network conditions. It helps monitor and manage the quality of the real-time stream.
RTP is widely used in various real-time applications, such as VoIP (Voice over IP) systems, video conferencing, streaming media, and telephony. It provides a standardized framework for the efficient and reliable transmission of real-time multimedia data over IP networks, enabling smooth and synchronized playback of audio and video content.
HLS PROTOCOL:-
HLS stands for HTTP Live Streaming. It is a streaming protocol developed by Apple for the delivery of live and on-demand multimedia content over the internet. HLS breaks down multimedia content into small, manageable segments and delivers them using standard HTTP protocols, making it compatible with a wide range of devices and platforms.
Here are the key components and workings of the HLS protocol:
Content Segmentation: The multimedia content, such as video or audio, is divided into short segments, typically 2 to 10 seconds in duration. Each segment is a standalone file with a unique URL.
Media Playlist (m3u8): HLS uses a master playlist (m3u8 file) that contains references to different variant playlists. The variant playlists, also in the form of m3u8 files, represent different versions or bitrates of the content. Each variant playlist consists of multiple segment references.
Variant Playlists and Adaptive Bitrate: HLS supports adaptive bitrate streaming, where the client device dynamically adjusts the quality of the stream based on the available network conditions. The variant playlists provide different quality levels or bitrates of the content, allowing the client to switch between them based on the network bandwidth.
Index Files: HLS uses index files (m3u8 files) that provide information about the available segments, their order, and duration. These index files help the client device request and retrieve the segments efficiently.
HTTP Delivery: HLS leverages HTTP as the transport protocol, which allows it to take advantage of existing content delivery networks (CDNs) and caching infrastructure. The segments and playlist files are delivered over standard HTTP connections, making it firewall-friendly and enabling widespread compatibility.
Adaptive Bitrate Switching: The client device monitors the network conditions and selects the appropriate bitrate variant based on available bandwidth and device capabilities. It can dynamically switch between different quality levels during playback to ensure smooth streaming without buffering.
Media Segment Encryption: HLS supports the use of encryption to protect the content from unauthorized access. It uses the AES-128 encryption standard and supports key rotation to enhance security.
Seeking and Playback: HLS supports seeking within the content by allowing the client device to request specific segments based on timestamps. The client buffers and plays back the downloaded segments in a sequential manner, providing a seamless playback experience.
HLS has become a widely adopted streaming protocol due to its compatibility, adaptability, and robustness. It is supported by various devices, including web browsers, mobile devices, smart TVs, and streaming media players. Many streaming platforms and content providers use HLS for delivering high-quality live and on-demand video content over the internet, enabling smooth and reliable streaming experiences for viewers.
ABR and DASH :-
DASH and ABR are two related concepts in the context of streaming media delivery:
DASH:
DASH stands for Dynamic Adaptive Streaming over HTTP. It is a streaming protocol that allows for the adaptive delivery of multimedia content over HTTP. DASH breaks down multimedia content into small segments, similar to HLS, and dynamically adjusts the quality of the stream based on the available network conditions. DASH utilizes media presentation description (MPD) files, which contain metadata about the available streams and segments, to enable adaptive bitrate switching during playback. DASH is an open standard and is widely supported by streaming platforms and devices.
Key features of DASH include:
Adaptive Bitrate Switching: DASH supports adaptive streaming, where the client device can dynamically switch between different quality levels or bitrates of the content based on network conditions. This allows for smooth and uninterrupted playback, adjusting the video quality in real-time to match the available bandwidth.
Efficient Bandwidth Utilization: DASH dynamically adjusts the video quality by selecting the appropriate bitrate variant for each segment. This ensures that users receive the best possible quality while minimizing buffering and maximizing bandwidth utilization.
Wide Device and Platform Support: DASH is compatible with various devices and platforms, including web browsers, mobile devices, smart TVs, and streaming media players. It allows for consistent streaming experiences across different devices, adapting to the capabilities and screen sizes of the viewing devices.
ABR (Adaptive Bitrate):
ABR refers to the technique of adapting the bitrate of a video stream in real-time to match the available network conditions and optimize the viewing experience. ABR is commonly used in streaming protocols like DASH and HLS.
Key aspects of ABR include:
Multiple Bitrate Variants: ABR involves encoding the video content into multiple bitrate variants, each representing a different quality level. These variants are typically pre-segmented and stored on the server.
Bitrate Switching: The client device continuously monitors the network conditions during playback and selects the appropriate bitrate variant based on factors such as available bandwidth, latency, and device capabilities. It can dynamically switch between the variants to maintain a smooth streaming experience.
Seamless Transitions: ABR aims to provide seamless transitions between different bitrate variants without noticeable buffering or interruptions. The switch between bitrates occurs during segment boundaries to ensure smooth playback.
Quality Optimization: ABR balances video quality and network conditions to provide the best possible viewing experience. When network conditions are favorable, higher-quality variants are selected, and when conditions deteriorate, lower-quality variants are chosen to prevent buffering and interruptions.
ABR, combined with streaming protocols like DASH, allows for adaptive delivery of multimedia content, ensuring optimal quality and smooth playback by dynamically adjusting the bitrate based on the viewer's network conditions.
Results
The live video workflow SOUTHWORKS delivered for Discovery quickly made a significant positive impact on the network’s business. In only six months, SOUTHWORKS delivered a complete workflow capable of supporting hundreds of live channels; now, Discovery has the capability to deliver a broad array of new channels and subscription-based services to audiences across the world.
Previously, preparing a new channel for a live event would take days or weeks. Its new video platform means Discovery can add live event channels in less than 30 minutes, including channels in 4K ultra-high definition (UHD) format. With automatic conversion of live TV to VOD content, time spent manually converting files for archive has been eliminated. The redundant workflow has proved reliable, maintaining high quality for live streams even during periods of peak demand, such as high-profile sporting events.
One of the first new offerings Discovery built on top of its live streaming workflow was Eurosport Player, a mobile and web app that achieved more than six million views in its first month. The app has been so successful that Discovery plans to use it as the streaming video platform for major global sporting events, hosting up to 45 UHD channels of live and on-demand sports action. Discovery also deployed its new video platform to build Food Network Kitchen, an interactive app that connects subscribers with popular cooking shows and celebrity chefs. With SOUTHWORKS and its video workflow built using AWS services, Discovery was able to bring Food Network Kitchen from concept to launch in only 10 weeks.
“Our live streaming platform has to deliver what our audiences have come to expect: a robust choice of engaging, one-of-a-kind content, delivered reliably and with great quality,” said Avisar Ten-Ami, Vice President, Video Delivery Platform at Discovery. “With the expertise of SOUTHWORKS and the foundation of Amazon Web Services, we’re able to bring the complete Discovery viewing experience to viewers on any screen.”