We're Hosting the NYC Apache Beam Kickoff Meetup!
With help from Austin Bennet, the organizer of the successful SF Apache Beam Meetup, Oden Technologies is hosting the first ever NYC Apache Beam Meetup at our office in Flatiron!
Tuesday, October 8, 2019
6:00 PM to 8:00 PM EDT
50 W 17th St 4th floor · New York, NY
One SQL to Rule Them All: An Efficient and Syntactically Idiomatic Approach to Management of Streams and Tables
by Tyler Akidau
What does it mean to execute streaming queries in SQL? What is the relationship of streaming queries to classic relational queries? Are streams and tables the same thing? What sorts of extensions are necessary to support robust stream processing SQL?
Over the last three years, members of the Apache Beam, Calcite, and Flink communities have held numerous discussions on the topic of streaming SQL semantics and syntax. Earlier this year, we formalized these ideas in our SIGMOD 2019 publication, “One SQL to Rule Them All: An Efficient and Syntactically Idiomatic Approach to Management of Streams and Tables”, in which we presented a three-part proposal for integrating robust streaming into SQL: Time-varying relations as a foundation for both classical tables and streaming data. Event-time semantics to ensure correctness despite out-of-order data. Minimal syntax extensions for controlling materialization of time-varying results. In this talk, I’ll give an overview of the contributions in the paper, discuss current implementations in Beam, Calcite, and Flink, and talk through ideas for future work.
Spanning the Edge and Cloud with Apache Beam
by Devon Peticolas
Oden Technologies provides a real-time mission-critical analytics platform to manufacturers. Mission-critical means real-time metrics and insights need to be computed in a way that’s resilient to the network connectivity issues that surround our customer’s factories, may the be in South Africa or a small midwestern town with only one ISP. In this talk we’ll go over how Oden is using Apache Beam’s portability to do duel-deployments of our streaming infrastructure both on Dataflow in Google Cloud Platform and on Apache Flink running on on-premise hardware
Lighting Talk: Running Flink in High Availability mode in Kubernetes
by James Maidment
The current Apache Flink documentation has info on running Flink in Kubernetes, but not in High Availability mode. If your JobManager crashes, you’re in real trouble. In this lightning talk we’ll detail the current State of the Union for running Flink in High Availability mode in Kubernetes, and cover some recently released projects.