Cloud Native Solutions

Let’s meet to hear and discuss two real cloud native cases from the media industry. Christian will present the on-going transformation at Berlingske Media, and Jan-Erik & Thomas will present their talk from KuBeCon China on how they used Kubernetes for a live TV show.

A cloud-native data platform in both batch and real-time for Business Intelligence, Data Science and system-integrations

A talk by Christian Selch, Head of Data Intelligence at Berlingske Media

At Berlingske Media we’re undergoing a transformation of the company towards being organized in a way that supports working data-driven and include both data usage and capture into everything we do. Concurrent with this organisational transformation we’re going through a process of decommissioning a large number of legacy-applications (operational systems as well as BI applications) to support this new way of working.

In order to support this we’re in the process of building a 100% cloud native environment from the ground up on AWS (using both managed services by AWS such AWS Batch, ECS, CodeBuild/CodePipeline, RDS & SpotFleets in addition to open source technology such as Apache Airflow, Apache Spark, Scala, Python, Docker & Terraform) to help us collect, standardize and join data across operational systems (delivering data in both batch and real-time) to support analytics & reporting as well as make the data (and derived analytical insights) available in operational systems to support data-driven actions or in a fully automated, algorithm-driven way.

Using Kubernetes for Handling Second Screen Experience of European TV Show

A talk by Jan-Erik Revsbech & Thomas Hector, CTO and Head of Operations at Peytz & Co

Danish TV station TV 2 runs a nationwide, very popular entertainment show called “Fuld Plade”. The show is basically a bingo show, where the audience can participate in the bingo game live on their smartphones, tablets, or computers. The nature of live TV shows add a very interesting workload scenario where the application needs extremely high scalability. During the hour of the show, the load on the system increases by a factor of 100, and is therefore a very good case for highly scalable solutions.

The presentation will show how we at Peytz & Co solved the task of delivering this kind of solution, using GoLang deployed to a Kubernetes cluster, and how using Prometheus helped us gain valuable insight into the performance of the application.