Explore an on-premise comparison of Databricks and Open-Source Apache Spark in this 21-minute conference talk by Booz Allen's Cyber AI team. Discover how they achieved 10x performance gains on real-world cyber workloads using Databricks Runtime Environment in an air-gapped environment. Learn about the challenges and solutions for implementing data science in sensitive operations, including the service-oriented architecture for capability deployment and high-performance computing project architecture. Gain insights into the results of Spark Open Source vs Spark DBR and valuable lessons for future on-premise installations in data-sensitive environments.
Overview
Syllabus
Intro
The Challenge: Go Fast.... On-Premise?
Solution: A Service-Oriented Architecture for Capability Deployment
Project Architecture: Focused on High Performance Computing
Results: Spark Open Source vs Spark DBR
Lessons Learned for Future On-Premise Installs
Taught by
Databricks
