Come learn about the latest changes in Apache Flink 1.5 and how we made stateful stream processing more powerful, more expressive, and more flexible to support applications that were previously difficult to realize.

Apache Flink 1.5 is the biggest Flink release the community has ever published. Not only does it contain many enhancements for its APIs and libraries but it also comes with major improvements to three of its core components. We reworked the network stack and added a credit-based flow control mechanism which results in even better throughput/latency trade-offs and more efficient checkpoints. Fast local recovery helps to speed up the recovery of applications with large state. Finally, the redesign of Flink's distributed architecture makes Flink fit naturally onto Kubernetes, Yarn, Mesos, and standalone setups with support for resource elasticity.