IBM deploys mission-critical IOT system on AWS for DB Systel GmbH

By and Aris Tsakpinis | 3 minute read | May 3, 2021

Photo: DB AG

About DB Systel GmbH

DB Systel GmbH, headquartered in Frankfurt am Main, Germany, is a wholly owned subsidiary of Deutsche Bahn AG and the digital partner for all DB Group companies. Currently around 5,100 Systel employees are employed at the main locations in Frankfurt, Berlin and Erfurt.

Deutsche Bahn Group (DB Group) is a leading supplier of mobility and logistical services with a clear focus on rail transport in Germany. Its headquarters are in Berlin. About 340,000 employees are employed by DB Group, including over 200,000 in the integrated rail system. By integrating transport and rail infrastructure, as well as through the economically and environmentally intelligent linking of all modes of transport, DB Group moves both people and goods. DB Group achieved 40.4 billion EUR revenues in 2019 and transported 13.4 million passengers per day.

Systematic digitization is of crucial importance to creating a strong rail system. In a joint undertaking with all of DB’s business units, DB Systel is working on the digitization of countless rail-related business processes to create high-performance user features that deliver maximum customer value. With this goal in mind, it has created a holistic, customer-specific range of services that meet the highest IT standards, and it is launching effective and efficient customer solutions that are based on innovations such as the cloud, big data, internet of things and artificial intelligence.

As a partner for long-term undertakings, DB Systel contributes extensive rail and IT expertise. Going further, DB Systel wants to anticipate customer wishes and maintain flexibility, so it carefully selects first-class business partners who can support it. These strategic partnerships are crucial to its collective success.

 Client Context

IBM supports DB Systel in the course of setting up a generic IoT Data Solution on Amazon Web Services (AWS) and therefore supports the IoT Platform approach of DB Systel.

The overall goal of the project is to create a digital representation of DB Group’s operational ecosystem and to enable data-driven use cases around the field of Intelligent Connected Operations and beyond.

Business Challenge

If passengers rely on public transportation, they expect trains to be on time. If customers engage, e.g. DB Cargo for their goods to be delivered, they expect timely delivery as well. These expectations are subject to unforeseen und unexpected operational challenges. Still precise and accurate information about the delay (e.g. duration and reason) is needed and to be communicated.

To provide latest train updates, good customer experience, timely maintenance, timely prediction of the potential problem and other operations, a seamless communication between assets like trains, stations, maintenance sites etc. across all business areas is of paramount importance. The solution was developed with the goal of enabling a smooth integration within DB’s evolving platforms.

Solution

The IBM team’s area of responsibility includes the design and implementation of highly available and hyper-scalable data engineering pipelines. These pipelines receive the sensor data stream originating from the assets, homogenize it and enrich it with context, e.g. from various enterprise or geofencing systems. Subsequently, the data is made available to operational systems in near real-time. Moreover, a data lake is fed containing historical batch data sets.

The implementation is based on AWS, using common AWS reference architectures for serverless IoT and Big Data solutions as well as the AWS “Well Architected Framework”.

Results

The successful go-live of data pipelines for several asset types was achieved. Today, the solution processes over 2,000 telemetry events per minute originating from DB Cargo wagons, summing up to almost 3 million events per day only for a single asset type. With the use of this data, use cases like the calculation of wagon parking fees, could be automated.

By applying DevOps concepts, the productive operation of the solution consisting of over 50 microservices can be managed parallel to its further development. For example, by abstracting individual components of the data pipelines in a context-independent manner, the solution could be extended to an application platform.

Author: Aris Tsakpinis