AWS for M&E Blog
Solving OTT workflow challenges with AWS
Authored by Tom Schaeffer, Founder, President & CEO of Float Left Interactive. The content and opinions in this post are those of the third-party author and AWS is not responsible for the content or accuracy of this post.
This post discusses the history and evolution of over-the-top (OTT) streaming and outlines how my company, Float Left Interactive, uses AWS to create a more intuitive and simplified approach to OTT management using our Application Management Portal (AMP). Readers will learn about some of the challenges introduced during the evolution of online video and how our development approach can help customers overcome those challenges to create engaging OTT experiences. For instance, we recently developed an easy and intuitive user navigation tool called Convert+ (available in the AWS Marketplace) making it easier than ever to streamline the transcoding process and leverage the power of AWS Elemental MediaConvert.
About Float Left
For over a decade, Float Left has been building and delivering OTT technology solutions for some of the world’s most prolific media brands. Our suite of products and services have matured with the industry to meet our customers’ ever-growing demands.
Over the last several years, one of our primary focus areas has been building management software to improve our customers’ workflow. In this context, a workflow is the process in which administrators publish new content, push app updates, and observe viewer behavior. When configured correctly, it’s a system that enables OTT service providers to respond quickly to changes in viewer behavior.
A brief introduction to OTT
For the uninitiated, OTT is an acronym that means “over-the-top,” and it’s a bit of a legacy term that describes additional, direct-to-consumer TV services “riding over the top of” traditional cable TV. The evolution of OTT is often referred to as “cord-cutting” or “cord shifting.” For the consumer, OTT means watching live and on-demand TV via apps on devices such as Roku, Apple TV, Amazon Fire TV, and most modern smart TVs.
OTT has been and continues to transform the TV landscape. For video-first media brands, OTT is the brave new world and offers opportunities to understand viewer behavior in ways that were not possible with traditional cable TV. These new insights grant providers the ability to create a more personalized content catalog, which increases engagement and reduces churn.
Over the years, OTT services have become much more sophisticated, and viewers expect high-quality content delivered in exceptionally great user experiences. Consider what Netflix has done to drive the industry forward. They have created the gold standard for the subscription-based, video-on-demand TV because they understand viewer behavior better than any other TV service in history. As the services evolve, so too must our ability to manage them.
Online video platform evolution
In the mid to late 2000s, most consumers watched digital video through web-based video players. With only one delivery destination, the workflow was simple. Today, that story has changed significantly considering the number and variety of connected TV devices available to consumers.
As technology continued to increase and new consumer devices entered the market, OTT service providers required more efficient ways to manage the various distribution channels.
Online video platforms (OVPs) entered the scene and provided a simple, packaged solution that included media storage, encoding profiles, content management, analytics, and more. OVPs grew rapidly with more consumers shifting their viewing habits to online video.
The demand for OVPs exploded after 2010 as digital video demand was surging. The challenge for OVPs was that as consumer devices, like Roku, started to gain traction, OTT providers were looking to add these platforms to their services. But the OVPs’ solution for managing video was too web and mobile-centric to support these types of devices. With so many customers, many OVPs were slow to pivot and adapt to their customer’s needs.
This increase in complexity is where Float Left saw an opportunity to develop an OTT-first management solution.
The OTT technology stack, at a glance
OTT workflow software manages the underlying OTT technology. To better understand these workflows, let’s look at the technology stack required to properly operate an OTT service:
- File encoding and transcoding – Raw media files are uploaded and transcoded (converted from their source format) into the required file formats. Different devices sometimes require different file types and configurations.
- Media storage – After the raw videos are processed, the various renditions are stored and made available for delivery.
- Content Delivery Network (CDN) – CDNs are needed to ensure requested video files are always available and delivered quickly. CDNs ensure uptime, quality of service, and redundancy of media files.
- Content Management System (CMS) – The CMS is a middle layer that sits between the CDN and the front-end applications that allow for content organization and metadata management (content title, description, runtime, thumbnails, etc.).
- Customer Resource Management (CRM) – CRMs are the primary method for subscription-based services to collect and track payments. The CRM also provides an authentication layer to the front-end application to ensure viewers have the proper access to the content available.
- Ad Server – Ad servers are responsible for selecting a relevant video and delivering upon request to the video player. These are usually in the form of “pre-rolls” (before the content plays), “midrolls” (during the content playback), and “post-rolls” (after the content finishes). Ad servers can also act as a programmatic method for locating and delivering the correct ads and the proper time.
- Application Programming Interface (API) – APIs provide the required data to the front-end applications when requested, i.e., loading the home screen with the data necessary to “draw” the home screen.
- Databases – Databases are at the heart of all software systems. In OTT services, databases provide the backbone to the APIs by allowing basic CRUD (“create, read, update and delete”) instruction to store features for cross-platform viewing, i.e., start watching a movie on your Roku and finish watching it on your iPhone.
- Analytics – OTT analytics provide reports on service usage. These reports provide an in-depth understanding of content consumption and are essential in providing a quality service. This information also becomes critical for personalization features, i.e., content recommendations
- Apps – TV apps are the last mile in the stack. Apps bring all of these pieces together in fractions of seconds to deliver an exceptional viewing experience to the viewer. Viewers expect low latency, high-performance TV experiences. Brands that offer a less-than-expected experience risk losing viewers forever.
OTT workflow challenges
There are several key challenges that businesses face when considering developing OTT workflows:
- There is no “one-size fits all” solution. OTT solutions must be architected in a way that supports flexibility throughout the supply chain. If the provider is not providing a specific supply-chain item directly, it must support quick and easy integration with partner services.
- Depending on the service offering, large tech teams are often required to meet the service providers’ demands to deliver feature-rich TV apps.
- The device market is incredibly fragmented, and each device platform requires knowledge of different programming languages. To create scalable solutions, development agencies need to architect frameworks that adapt and scale to rapidly changing requirements.
- OTT workflows are never the same between OTT services. Even subtle differences in service options can drastically modify the workflow. After years of working with large media brands, we’ve found that each has created their unique method for uploading and managing the content, placing ads, and updating metadata (title, runtime, director, actors, etc.).
Introducing AMP
At Float Left, we faced these challenges head-on, not by adding more engineers but getting smarter by developing new technology. Float Left’s AMP (Application Management Portal) is an OTT-focused management solution to the OTT workflow challenges.
AMP is the culmination of years of research and work, understanding the consistencies and differences of our customer’s workflows. We realized the many OTT technology providers were failing to pivot fast enough from web and mobile-centric solutions to full-service OTT solutions that included support for the CTV devices (the devices used to watch content online, like a streaming device, smart TV, or gaming console).
AMP’s primary goal is to be an OTT-first solution focused on brands that are launching OTT services across a multitude of fragmented platforms.
AMP’s primary purpose is to make managing OTT services more intuitive for brands distributing across a wide range of devices and platforms. It is a powerful combination of content and application management. Whereas traditional CMS solutions tend to focus on assets and content playlists, AMP is built around the concept of modifying screen layouts.
How does it work?
Administrators start with a blank screen and populate the screen with playlists that are made up of either single videos or a collection of videos – referred to as a series. This is helpful when creating a list of TV shows that contain seasons made up of individual episodes. This method of management takes a hybrid approach to OTT app management.
We believe that this approach will make service management much more efficient for our customers.
Building and deploying such a sophisticated architecture would take a dedicated team many months to build. Today, we’re able to make product decisions in a matter of days to weeks as opposed to months to take advantage of immediate market opportunities.
The AWS product suite has been a staple of our infrastructure for years. AWS allows us to offer a flexible architecture that adapts to each of our customer’s needs and frees our engineering team of the worry of dealing with architecture schematics, load balancing, and the human resources required to maintain such a system.
AWS products we use
AMP is built on several key pieces of AWS products:
Amazon Simple Storage Service (Amazon S3)
We use S3 primarily for the storage and availability of files required by the TV apps to run. We’ve moved much of the app UI into configuration files stored here and used at app runtime. We create “buckets” that contain specific data depending on its intended use. At runtime, our apps request the particular file, load the data, then draw the screens as instructed. By leveraging S3, we now have server-side control over the applications.
CloudFront is the content delivery network that acts as a software layer that sits on top of our S3 buckets and creates redundancy. Its primary function is ensuring that requested resources in our S3 buckets are always available.
Amazon Elastic Cloud Computing (Amazon EC2)
EC2 instances are virtual servers that are created and destroyed as they are intended to perform specific jobs. Once a job is started, a server will be instantiated and then terminated once the job is complete. This allows us to leverage powerful computing in the cloud when needed without maintaining servers indefinitely.
Lamda is relatively new to our infrastructure. Lambda “functions” are routines translated into executable code that perform essential functions. It allows us to query databases for specific information and complete automated jobs processing data for various outputs.
Lambda functions can be written in various languages, which means development teams can leverage existing knowledge to integrate nicely with existing technology.
All of the APIs that we use for sending and receiving information from the front-end applications are serverless. Using serverless technology, leveraging Lamda, and the Serverless architecture, we can build APIs that will infinitely scale.
DynamoDB is a “NoSql,” non-relational database that leverages “key-value” pairs instead of traditional database tables. Once integrated, DynamoDB provides a powerful, fast, and efficient way to ensure the data necessary for our APIs at scale. DynamoDB offers us the flexibility to scale without worrying about hardware.
A “serverless” framework allows us to define our infrastructure via software and AWS, specifically all of our APIs, driven by Lambda functions, talking to the DynamoDB instances for quick delivery of data to the front-end applications.
Before AWS, this type of infrastructure would have taken months into years to design, test, build, and deploy.
Looking towards the future
In the future, OTT services will become more personal and tailored to individual viewers. Data will become more critical than ever, not only to understand how the services are performing but also to improve the overall experience for the consumer. With sufficient data, advertisers will be able to send the proper ads to the right person at the right time.
As OTT services continue to evolve, data, machine learning, and artificial intelligence will be critical for service providers to understand their audience in a highly competitive marketplace.
In closing
When we started over a decade ago, the idea that we could do more than offer OTT app development was far-fetched. We stuck to what we knew, learned everything we could about the entire tech stack required to be the backbone of an OTT service, and waited until the timing was right.
Taking too long to develop a product that is either not relevant anymore or too far behind its competitors’ versions prevents us from capitalizing on critical opportunities. AWS services have made it possible for small teams like ours to quickly build sophisticated products and services at scale, efficiently and effectively.