Case Study: Salesforce DMP doubles processing speed and scales petabyte‑level global data with Amazon Web Services

Salesforce - Customer Case Study

Salesforce DMP (formerly Krux) is a leading data-management platform that captures and activates audience data across devices and channels to drive real-time, personalized experiences. Operating on petabyte-scale data—interacting with over three billion devices and processing billions of events monthly—the company needed always-on, low-latency infrastructure that could scale without limits and support flexible, customer-driven audience taxonomies.

To meet that demand Salesforce DMP built its platform on AWS, using Amazon S3 for storage, Amazon EMR with Apache Hadoop and Spark for batch and ML workloads, AWS Data Pipeline for orchestration, Amazon DynamoDB for segment storage, and EC2 Spot instances for cost-efficient compute. The result: a 100% increase in processing speed for data science and ETL jobs, major cost savings (up to 40% vs. Reserved Instances and up to 80% vs. on‑demand), and the flexible, scalable capacity needed to deliver real-time personalization and rapid product innovation.

Salesforce

Roopak Gupta

Vice President, Software Engineering