How a French startup embraced open source technology to become the data provider for the Formula E championships

The sports data startup from near Nice in the south of France concentrates on enriching sports' fans experience with live statistics and video to be consumed as 'second screen' applications. It won big with the contract for the official Formula E Championship mobile app last season.

When asked how Intellicore won this contract, a representative said: "Intellicore may be seen as a startup but our data platform is unique on the market and we have long been recognised for delivering best in class apps, so when a rights owner like Formula E evaluates what we have to offer they feel very secure in choosing us to deliver quality products for the fans."


IT powerhouses like SAP and Infosys have built consumer facing, real time data driven applications for sports like the NBA and the ATP tennis tour, respectively, but Intellicore hopes it can avoid competing with these sort of vendors with its unique approach.

Read next: Tennis data: How ATP is using technology to reach new audiences

"We don't really want to become direct competitors to those big companies," said senior software architect Annard Brouwer, who sat down with Computerworld UK at the Formula E finale in London's Battersea Park this month.

"We give our own unique spin on it. We want to add some predictive and gamification, so when we monitor what is going on during the game we can detect a pattern and ask in our app, or in a third party app, something like: 'these two are battling it out, who do you think will win' to engage the spectator into the sport more."

It is engagement, and not gambling, that Intellicore say they want to push through their predictive capabilities.

According to Brouwer: "Personally I don't like [gambling] but there is a possibility. That is not our primary aim. I am sure that at some point we will have to support it but we don't want to be a betting company, we want to find a way to engage the fan directly into the sport [] If there is an opportunity we will never say no, but that's not the primary goal of our company."

Formula E

Brouwer went on to discuss how Intellicore plans on opening up the glut of data being driven by the state-of-the-art Formula E electric racing cars. During a typical 50-minute Formula E race, the cars generate around 1GB of telemetry data per driver at a rate of 400 packets per second, each packet containing up to hundreds of telemetry events. Over 1.2 million packets are received during a race.

Formula E app screenshot

The job of the Intellicore platform is to aggregate, normalise and redistribute that data in real time to the official mobile and tablet Formula E app. Through the app fans get access to data such as a car's top speed, its position on a live map, the video feed of drivers' cockpit and the percentage of lap completion.

Riak TS

Technology wise, Intellicore turned to open-source database specialists Basho, specifically the Riak TS product.

Brouwer explained: "On the backend where all the magic happens we started using Riak to help us with our data processing. We use two warehouses, we have a traditional SQL database for structured data like race schedules, teams, information about drivers. Information which doesn't change very often but where the relationships are very important.

"Then we have the raw computer data and we use Riak for that. In Riak, because it clusters the data in a nice way that doesn't cost any extra performance, access will be as quick as needed. For Riak TS they guarantee that for a specific series of data it will be very closely clustered, so there is still replication, but you don't have to traverse the whole cluster to get to the data."

Read next: Hadoop vs Spark: Which is right for your business Pros and cons, vendors, customers and use cases

The TS part is the important bit for this use case. Meaning time series, this is the database which allows for live data streaming and analysis to take place, and timeliness is vital in a sport where every millisecond counts.

"We were hit by the challenges of volume, latency and cost when it came to our data," said Intellicore CEO Christian Trotobas. "We needed a solution that was able to scale rapidly, and one which wouldn't fail whilst dealing with mission-critical workloads."


By Scott Carey

Zur Startseite