Building an Efficient Data Pipeline for Data Intensive Workloads

Moving data through the pipeline in an efficient and predictable way is one of the most important aspects of modern data architecture, particularly when it comes to running data-intensive workloads such as IoT and machine learning in production. This talk breaks down the data pipeline and demonstrates how it can be improved with a modern transport mechanism that includes Apache Arrow Flight. This session details the architecture and key features of the Arrow Flight protocol and introduces an Arrow Flight Spark data source, showing how microservices can be built for and with Spark. Attendees will see a demo of a machine learning pipeline running in Spark with data microservices powered by Arrow Flight, highlighting how much faster and simpler the Flight interface makes this example pipeline.