Data Microservices in Apache Spark Using Apache Arrow Flight

Explore the world of machine learning pipelines and learn how to enhance them using Apache Arrow Flight in this 24-minute talk from Databricks. Dive into the fundamentals of Apache Arrow and Arrow Flight, understanding their architecture and key features through examples of Flight servers and clients. Discover the Arrow Flight Spark data source and its potential for building microservices with and for Spark. Compare the performance of Flight against other common transport protocols through benchmarks. Watch a demonstration of a machine learning pipeline running in Spark, powered by Arrow Flight data microservices, to see firsthand how this technology simplifies and accelerates the process. Gain valuable insights into improving data transport efficiency in machine learning pipelines using Apache Arrow Flight and Spark.

Introduction
What Arrow isnt
Building Blocks
What is Arrow Flight
GRPC
Microservices
Parallel streams
Python clients
Spark Source
Spark Source for Flame
Benchmarks