Bob Muglia, Snowflake Computing | VMworld 2015
summary
Snowflake describes a SaaS cloud data warehouse built from scratch with a multicluster shared data architecture that eliminates traditional admin tasks like vacuuming and distribution key tuning so customers need no DBA. It auto-parses semi-structured JSON/Avro into a variant data type at load time, applies columnar compression and pruning, and runs full SQL queries orders of magnitude faster than brute-force Hadoop scans. Company is three years old, ~85 people, Series C $45M (total $76M), and cites a customer table ~200 TB compressed (1–2 PB raw).
AI-written summary · about
chapters (10)
- 0:00Introduction and welcome
- 1:44Bob Muglia's background
- 3:07Cloud database vs packaged software
- 6:38Comparison with Oracle and Redshift
- 9:30Cost and TCO advantages
- 11:37Migration to cloud and enterprise adoption
- 14:05Workloads and semi-structured data
- 17:58Handling semi-structured data efficiently
- 19:43Multi-model positioning and competition
- 21:28Closing remarks
