Flink Forward 2016 is over and for the second time we’ve enjoyed an awesome conference with interesting sessions and inspiring discussions. Hopefully, you met old friends and made new ones as well, gained exciting insights and had a pleasant time in sunny Berlin. The Flink Forward team is happy about this successful event and wants to thank all speakers, attendees, volunteers, media partners and sponsors for their support and involvement.

Before we take a look back at three exciting days full of Apache Flink and related streaming technologies we want to ask you to help us make Flink Forward even better: Please take a minute to fill out our feedback survey.


Flink Forward kicked off on Monday with Kostas Tzoumas’ and Stephan Ewen’s Keynote “The maturing data streaming ecosystem and Apache Flink’s accelerated growth”. It provided an interesting overview on several large-scale stream processing use cases. On the following two days of conference 350 attendees visited Kulturbrauerei. Over 40 speakers offered 25 hours of technical talks on 3 parallel tracks. This year’s program had a broad range of data stream processing topics. Among others the conference presented talks on the design and implementation of Alibaba’s Flink-based system, Apache Zeppelin including Flink, the basics of Apache Beam, the data streaming infrastructure of Netflix, and the future of Apache Flink.

The second day started with Ted Dunning’s Keynote “How can we talk Flink Forward?” followed by the panel “Large Scale Streaming in Production” with leading experts in the field from data Artisans, Alibaba, MapR, Netflix, and Uber who talked about the challenges they have faced and the solutions they have discovered while deploying stream processing in production at very large scale.

Overall, Flink Forward 2016 was a testament to the maturing and growing Flink community. The scale at which companies reported to use Flink in production processing tens of billions of events per day and scaling Flink to 1000s of nodes and 100s of gigabytes of state while guaranteeing exactly once was astonishing. Equally impressive was the amount of improvements done in the Flink platform itself as presented by several committers and contributors at the conference.


After the conference officially ended on Tuesday evening more than 75 participants learned more about Apache Flink in hands on training sessions on Wednesday. Starting with an introduction into Flink’s DataStream API to lectures on concepts of processing time, participants gained in-depth insights into Apache Flink.

10In case you missed some sessions at the conference you have the chance to watch the talks online. The video recordings from all talks from all three stages are uploaded on our YouTube channel. All videos are also added to their corresponding session description on this website, along with the presentation slides.

You can also find a collection of photos from this year’s conference on Flickr – please take a look!

Last, but definitely not least we would like to thank our program committee, the data Artisans team, David Anderson for supporting and organizing the training as well as the team of Kulturbrauerei Berlin who were involved in the organization of Flink Forward for an inspiring time. It’s been a blast! See you at Flink Forward 2017!


Photos: CC BY 2.0 by iStream