Deutsche Börse Group launches data analytics platform with AWS
International exchange organisation Deutsche Börse Group gives companies and investors access to global capital markets. After seeing an increasing demand for easier access to detailed historical market data for trading algorithms, Deutsche Börse Group developed A7, a cloud-based data analytics platform.
Once the company had decided they wanted to build this platform, they needed a quick way to do it and to find a way of making it scalable and flexible for its customers. Using Amazon Web Services (AWS), Deutsche Börse Group was able to build and launch A7 in just four months.
“There is a significant group of customers who want simple and easy-to-use access to most granular market data without implementing solutions or the need for raw data,” said Sven Wohlfarth, head of data services operations at Deutsche Börse Group. “This is the group of customers we built A7 for.
“We needed flexible and scalable infrastructure for A7, and building on AWS was the most efficient way to allocate new resources for the application.”
The trading business is a very competitive sector, and AWS allowed Deutsche Börse Group to build and launch its new service in a timely manner, because the historical market data was already stored on Amazon Simple Storage Service (Amazon S3).
“Time to market was really important for us,” explained Zoltan Vekony, lead developer in data services at Deutsche Börse Group.
A7 is now a ready-to-use service that provides fast and direct access to tailored metrics based on Deutsche Börse Group’s most comprehensive market data, without customers having to invest in new technology. It works by using multicast streams to capture and transfer 60TB of historical market data to Amazon S3 buckets.
A7 then uses the managed Amazon Elastic Kubernetes Service (Amazon EKS) to compute the data, transform it, and convert it for analysis. Amazon Relational Database Service (Amazon RDS) is then used for metadata and Amazon Cognito manages the security and authentication of users.
“Without that market data, it would be difficult for clients to form proper decisions or train their algorithms so that they can trade even more successfully,” Wohlfarth said.