GRID Optimized Game Data Ops for Live Esports Events by Migrating to AWS

The GRID Data Platform has grown over the years, scaling the number of rights holders from 1 to over 90 and commercial clients to over 350 across 20 use cases. In 2018, when the company started, it used multiple cloud service providers for different use cases, including AWS for storage. At the end of 2021, GRID discovered the benefits of using AWS products in its stack when it started a new project for secure video data feeds for its betting clients.

The company used AWS Elemental MediaLive, which encodes live video for broadcast and streaming to virtually any device, to ingest video of live tournaments and develop an embeddable video feed accessible by its large network of partners. GRID also used AWS Elemental MediaPackage, which prepares and protects video for internet delivery, to repackage and distribute video to customers for use in their products. “We had access to experts from AWS who spent time making sure that we were using AWS services in the right way,” says Chris King, chief technology officer and cofounder of GRID. “The level of support that we received on this project was important to us.”

After the success of the video data project, GRID began an ML project using Amazon SageMaker, which is designed to build, train, and deploy ML models for nearly any use case with fully managed infrastructure, tools, and workflows. The company used Amazon SageMaker to develop and train ML models that can predict what is likely to happen during esports matches. GRID also used Amazon SageMaker to put in place a pipeline that uses its proprietary game-agnostic data format, offering standardized data across the GRID platform to streamline and automate it for near-real-time data processing and management.

With the success of its projects using AWS and the positive experience working alongside AWS solutions architects, GRID decided to migrate fully to AWS in 2023. “The GRID Data Platform offers a variety of services, from data extraction to generating predictions and distributing data, often in real time and on a dynamically changing scale,” says King. “After testing various options, we are confident that AWS services are the most comprehensive and reliable for the type of operations that we have, and we are comfortable relying fully on them.”  

For the company’s initial video project using AWS, GRID used Amazon CloudFront, a content delivery network service, to securely deliver content with low latency and high transfer speeds. Within weeks, the solution was rolled out to several of GRID’s data partners. With this first solution, GRID used AWS to stream live tournaments of the games Counter-Strike: Global Offensive and Dota 2 to quickly provide video for its customers in betting and media.

“We kept close contact with solutions architects and the AWS team and talked about GRID’s aspirations and a lot of different use cases where GRID fulfills a pioneering role in its creation of data solutions for the esports industry,” says King. “Our second AWS project, involving ML, was born out of one of these conversations.” GRID has biweekly meetings with the AWS team about its ML project, which provides deeper data insights, optimizes the predictive models, and creates new types of predictions.

After gaining positive experiences in these biweekly meetings and working alongside AWS solutions architects, GRID began migrating fully to AWS in April 2023. The company used the AWS Migration Acceleration Program (AWS MAP), a comprehensive and proven cloud migration program. “The AWS MAP program is great, and we use it to lower costs as we scale up, taking a careful approach to avoid disruption to our end users,” says King.

The company has over 200 containerized services running in production in addition to a staging environment and makes more than 10 deployments a day across each one. This is all handled using Amazon Elastic Kubernetes Service (Amazon EKS), a trusted way to start, run, and scale Kubernetes. Using AWS, the company has reduced infrastructure management workload by 50 percent, and its developers can adopt services that support quick feature development without compromising quality.

GRID has also improved its efficiency and agility by migrating to AWS. The company has increased model-training-process efficiency by 200 percent, and since it increased its use of AWS in 2023, GRID has onboarded over 50 new partners. “The core of what we do technically is help our partners to extract a huge amount of data and then process, enrich, and distribute it—in near real time and with as close to 100 percent reliability as possible,” says King. GRID is also achieving high availability, which is important for its edge layer to support data integrity for its customers.