Our July 2018 meetup - A Report
July 2018 - On July 19, 2018, we hosted a community meetup at LinkedIn focussing on Apache Kafka, Apache Samza, and related streaming technologies. The event featured technical deep dives by engineers from LinkedIn and Uber on the latest in the stream processing space. Here is a brief summary of the talks that were presented.
LinkedIn’s Xinyu Liu presented Beam me up Samza, describing how Linkedin is harnessing cutting edge features of Beam. Apache Beam provides an easy-to-use, and powerful model for state-of-the-art stream and batch processing, portability across a variety of languages, and the ability to converge offline and nearline data processing. In this talk, he discussed the Beam API and its implementation in Samza and the benefits of Beam Runner to the Samza and Beam community. He also explored various use cases of Beam at LinkedIN and future work on it.
Uber operates more than 20 Kafka clusters to collect system, application logs and event data from rider and driver apps. Uber’s Hongliang Xu shared his insignts on Uber’s approch for replicating data between Kafka clusters across multiple data centers. He covered the history behind uReplicator and gave the high level architecture. Furthermore he also discussed the scalability challenges and operational overhead as the Uber exapanded and how did they build Federated uReplicator which addressed challanges at scale
Concourse is LinkedIn’s first near-real-time targeting and scoring platform for notifications. In this talk LinkedIn’s Ajith Muralidharan & Vivek Nelamangala provided an in-depth overview of the design and various scaling optimizations. Concourse has an ability to score millions of notifications per second, while supporting the use of feature-rich machine learning models based on terabytes of feature data.