Sportradar Accelerates Data and Video on AWS, Launches Simulated Reality Innovation in 11 Days

Sportradar—a global leader in delivering sports analytics and content—faced a scaling challenge. As the company’s growth accelerated, Sportradar found that its on-premises architecture couldn’t keep up with demand from media companies, sports federations, and the betting industry. Migrating to Amazon Web Services (AWS) helped Sportradar scale its services, continue to innovate its offerings, improve the fan experience, and mitigate risk. By using AWS services such as Amazon Kinesis Data Streams and Amazon Elastic Compute Cloud (Amazon EC2), Sportradar increased data reliability by a factor of 10, decreased recovery times, and quickly launched its new product, Simulated Reality, at a time when live sporting events were being canceled.

Sportradar’s live data intelligence services provide game statistics and play-by-play action as well as real-time analysis of data streams from sporting events. The company also distributes live video from a vast array of rights holders to betting operators around the world. Customers choose Sportradar for its depth and quality of data and for the breadth of sports it covers—as many as 20,000 live sporting events and matches a month through its engagements with more than 1,000 sports leagues. 

The company processes and distributes streams of real-time data to customers in 80 countries. During a typical game, a Sportradar representative works on location, continuously documenting the action through a specially designed app that immediately uploads every input to the company. Through automated data-processing workflows, Sportradar performs instant data analysis and then transmits the information to customers in seconds. Customers receive this intelligence as data feeds that integrate into their applications or betting systems or as live video streams distributed over internet protocol (IP) networks. On peak days, Sportradar processes data from as many as 1,500 live events simultaneously. 

Prior to building a solution on AWS, Sportradar maintained duplicate failover servers in its data centers for compute and database workloads. If its production servers went down, these hot standby servers could take more than 30 minutes to come online, leaving customers without valuable data in that time.

To achieve the level of data-processing performance its applications require and to reduce the burden of maintaining and updating on-premises databases, Sportradar migrated its data-processing and analytics workloads to the managed database and analytics solutions offered by AWS. The company turned to Amazon Aurora, a MySQL- and PostgreSQL-compatible relational database that combines the performance and availability of traditional enterprise databases with the simplicity and cost-effectiveness of open-source solutions. Sportradar also made use of Amazon DocumentDB (with MongoDB compatibility), the fully managed document database service that decouples storage and compute, letting them scale independently. The company found a suitable Elasticsearch solution in Amazon Elasticsearch Service (Amazon ES), a fully managed service that makes it easy to deploy, secure, and run Elasticsearch cost effectively at scale. Meanwhile, Amazon ElastiCache enabled Sportradar to seamlessly set up, run, and scale popular open-source-compatible, in-memory data stores in the cloud. 

In parallel with its database migration, Sportradar shifted its compute infrastructure to Amazon EC2, a web service that provides secure, resizable compute capacity in the cloud. The company integrated an array of services—including Amazon Kinesis Data Streams, a massively scalable and durable real-time data-streaming service, and Amazon Elastic Kubernetes Service (Amazon EKS)—to enable containerized application environments. Sportradar also deployed AWS Lambda to enable serverless compute close to the network edge and to enhance performance and cut latency for customers. 

The company took advantage of AWS Regions to optimize the performance of its services around the world. Now, Sportradar’s databases are distributed across the AWS global network, and its customer applications are run locally, using the Amazon EC2 compute pods located closest to users. Applications access data within the same Region, lowering latency while limiting the impact of potential outages. Sportradar further enhanced its service performance using AWS Global Accelerator, which provides static IP addresses that act as fixed entry points to the company’s data-processing application endpoints. 

To enhance the stability and performance of the content it receives from remote locations, Sportradar adopted AWS Elemental MediaConnect, a high-quality live-video transport service, for signal contribution. The company uses foundational AWS services and AWS Elemental MediaConnect to achieve an error-resilient signal delivery, avoiding the dedicated fiber connectivity it used to rely on. Now, Sportradar can capture live video feeds from sports leagues and other content owners through AWS and integrate them into its video operations center. From here, the company can relay the video feeds to customers or further process them and incorporate real-time statistics and analytics prior to distribution. AWS Elemental MediaConnect enables Sportradar to reliably transport live streams through the AWS network at bitrates of 20–30 Mbps with built-in error correction. 

More than 1,000 businesses depend on data and video from Sportradar. The company is now capable of executing up to 50,000 betting transactions per minute and performing up to 10,000 match simulations per second to determine how games might play out whenever odds change. It runs its infrastructure in an active-active framework, with duplicate workflows distributed across multiple AWS Regions and Availability Zones for a 10-times increase in resiliency and failover and for recovery times as low as 15 seconds. As a result, Sportradar no longer requires duplicate on-premises standby servers or 30-minute recovery times.

Technical innovation is at the forefront of Sportradar’s commitment to its customers and stakeholders. Having upgraded its data and video infrastructures, the company again turned to AWS to facilitate a series of innovation initiatives that held strategic value to its business, starting with its real-time analytics infrastructure. Sportradar launched a project to redevelop this critical technology using AWS Data Lab with a 4-day intensive engagement between a team of Sportradar developers and AWS technical experts. 

“Sportradar engaged AWS Data Lab for guidance on developing a modernized, low-latency data analytics pipeline and workflow to power real-time statistical models, feature extraction, and inference using machine learning and real-time dashboards,” says Vassili Patrikis, manager and solutions architect for AWS Data Lab. “Using the collective team’s strong technical skills, Sportradar built its real-time streaming analytics platform in just 4 days. Not only was the Sportradar team able to accomplish all of its goals through the AWS Data Lab engagement, but the team also set the bar in terms of time from prototype to production. Sportradar truly embodies the AWS Data Lab model of ‘Think big, start small, scale fast.’” 

The company also engaged the AWS Europe, Middle East, and Africa Prototyping Lab team in building a new Internet of Things (IoT)–based voice recognition system. This enables the company’s scouts to automatically process events—such as soccer goals, corner kicks, or offside violations—using voice recognition, even in noisy, stadium-like environments. The solution uses embedded IoT hardware, a microphone, and a mobile phone that runs machine learning algorithms to detect specific words related to sports reporting and to automatically relay these events to the AWS backend in real time.

Deploying its workflows on AWS has given Sportradar newfound operational agility, easing the onboarding of customers and media partners and accelerating the deployment of new offerings. In 2020, as the COVID-19 pandemic sidelined live sports across the world, Sportradar moved quickly to serve customers with a new product. Dubbed “Simulated Reality,” Sportradar’s breakthrough live sports offering uses artificial intelligence to simulate matches, using real-world rosters, teams, and leagues in professional cricket, tennis, and soccer. Game action plays out over the same time line as a live-action event, and betting operators facilitate wagering on outcomes and in-game moments just as they would with live sports. Using its AWS infrastructure, Sportradar was able to bring Simulated Reality from concept to revenue in only 11 days. 

“Global leaders in sports trust Sportradar with the integrity and value of their products,” says Ben Burdsall, Sportradar’s chief technology officer. “To uphold that trust, our services must be fast and reliable whenever and wherever our customers need them. AWS gives us a resilient, flexible foundation to deliver high-performance data and video services our customers can depend on.”