Unleash the Magic of Salesforce Genie - in Action
summary
Claims a production data cloud that handles petabyte scale with a zero-copy architecture that can virtualize Snowflake and other warehouses, stores raw data on Iceberg + Parquet on S3, and runs ETL and streaming with Spark stream processing plus Airflow, Kafka and CDC. Reported capacity numbers include handling 50 billion profiles, 1 trillion engagement events per customer, 3M compute hours on Black Friday and activation of 43 billion engagements; profile API sustains 30k qpm at 375 ms while multitenancy, data skew and union-heavy queries remain operational challenges.
AI-written summary · about
chapters (22)
- 0:00Introduction to Data Cloud
- 1:02Why companies need Data Cloud
- 3:53Data Cloud use cases explained
- 6:24Benefits and challenges of Data Cloud
- 8:52Real-time vs batch processing
- 10:17Key challenges in Data Cloud
- 13:06Value and ecosystem of Data Cloud
- 15:35Governance and practical considerations
- 16:57Introducing Salesforce Data Cloud
- 17:41Zero-copy architecture explained
- 18:47Salesforce Data Cloud architecture overview
- 20:55App layer capabilities
- 23:23Data flow and architecture deep dive
- 24:24Change feeds and segmentation
- 25:06Scale and performance metrics
- 25:49Ingest challenges and solutions
- 27:14Eventing and streaming optimizations
- 28:17Processing data at scale
- 29:01Event-driven processing approach
- 30:25Delivery and activation challenges
- 31:28Query performance and constraints
- 31:48Query troubleshooting example
related talks
30mArtificial Intelligence More Like Artificial Stupidity
04mInterview
29mData foundation for analytics and AI
29mSimplifying data flows
07mThomas Gerber Interview
13mLightning Talk: Highlights of the Snowflake Elastic Data Warehouse
28mScalable AI & Web Insights with GitLab, Snowplow & Snowpipe
05mUnleash the Power of Data at Scale