Navigation and Service



07 July 2020 CEST | Market + Data services

New A7® platform offers “analytics as a service”  Interview with the A7 team

With A7®, Deutsche Börse’s Market Data + Services team has recently launched a new cloud-based analytics tool for high precision market data. Who are the people behind the new analytics platform, what does A7 offer to our clients, and what makes it so special? Our colleagues Zoltan Vekony, from Data IT, as well as Taras Beletskiy and Aneliya Hristova, from Market Data + Services, have the answers.

Congratulations on the launch. How long have you all worked on the new platform and what team is behind this?

The A7® platform is the result of a close cross-divisional collaboration between Market Data + Services and Data IT. Some may remember that we presented the prototype during the OpenDay forum last year and now, a couple of months later, we are ready to offer the final version to our customers. 

We would like to take this opportunity to acknowledge the extreme dedication of the whole team. It was a privilege to work with colleagues so motivated and willing to go the extra mile to deliver a solution which will help customers to address some of their pain points.

A7® is part of the Deutsche Börse 7 Market Technologies. How and where does it fit in?

Yes, our A7® analytics platform uses the market data feeds from the T7 trading system from the 7 Market Technologies. This tight coupling of services enables customers to work with exact the same, unchanged data formats and feed characteristics which they use in their trading environment. This means that, for example, the back-testing of customer algos will work exactly the same way with real-time trading data feeds. This lack of side effects increases the quality and validity of the results, as well as making it easy to port algos into trading environments.

What is A7®’s unique offering to our clients?

There are several key unique features: usage of un-normalized order-by-order market data in nanoseconds and granularity for all traded instruments on Eurex and Xetra. One must remember that these premium market data feeds are only available at co-location, which requires significant client investment. A7® already includes historical data from these premium market data feeds, as well as an algo workbench to develop and back test execution algorithms, including cross-product and even cross-market strategies. A7® also provides easy access to market data via state of the art interfaces. The customer can decide which tool he chooses – from C++ to even MS Excel – to integrate market data. 

But what makes our A7® platform really unique is the technology stack: we built the core system from scratch using C++ to be able to manage large amounts of data – something which can create performance issues for common ‘off-the-shelf’ data analysis tools.

Why did you decide on the cloud?

We have chosen a cloud-based infrastructure because of its scalability and flexibility. It allows us to scale the infrastructure according to the current needs we have and use the advantages of infrastructure as code. For instance, one day of Eurex EOBI data amounts to half of a terrabyte of data. Cloud is able to provide a solution for handling that.

What makes this data so valuable to its users?

A7® provides instant access to more than a year of historical public market data from our Eurex and Xetra trading systems. However, the real value we provide to the customer is not the data itself, but the business layer we have added on top. Customers do not need to build their own big data and machine learning infrastructure, but can use ours instead. 

We offer “Analytics as a Service”. As such, time-to-market and costs for building and operating such a platform can be significantly reduced. In addition, the generation of insights, as well as the back testing of algos, will come at a lower price compared to painful surprises during testing in real trading.

How easy to access is our new platfrom?

We have an easy to use web-based front-end that suits the needs of different client groups. In addition, all functionalities and all market and derived data on the platform can be accessed via the REST API. This protocol is generally accepted and widely used. Regarding the customer algos – one of the core elements of the system – the customer has to be familiar with the C++ programming language.