Serverless Recipes for Big Data Architectures

Explore seven critical steps for migrating from Spark clusters with Map-Reduce to a serverless architecture in this 47-minute conference talk from Devoxx. Learn from the BBC Audience Platform Data team's experience in collecting, transforming, and delivering billions of daily events from audience interactions with mobile apps and websites. Discover the good, bad, and ugly aspects of the migration process, including valuable lessons learned along the way. Gain insights into achieving real-time analysis at up to 25,000 transactions per second, meeting the demands of high-traffic news days while significantly reducing cloud infrastructure costs. The talk covers key topics such as architecture, scalability, observability, global reconciliation, and security, providing practical knowledge applicable to both existing and new big data architectures.

Introduction
Why are we doing this
Architecture
Scalability
Observability
Global reconciliation
Security
Summary