Vaughn Stewart, VAST Data | SC23
summary
VAST Data argues the old tradeoff between high performance and low cost disappears by combining NVMe flash, disaggregated compute nodes and a software data platform used by CoreWeave and Lambda; the company cites ~700 staff, customers like TACC, and cloud partnerships. Features include Trino pushdown and Parquet pushdown that return only queried parquet columns, claiming up to 20x faster query performance versus tuned all‑flash arrays, a global namespace for hybrid caching to avoid massive data egress.
AI-written summary · about
chapters (12)
- 0:00Welcome and introduction
- 1:05VASTronaut achievement and career journey
- 3:13Storage industry challenges and VAST's vision
- 5:02GPU cloud partnerships and AI disruption
- 8:54VAST Data as a data platform
- 11:20AI revolution and data management
- 13:23Optimizing modern infrastructure
- 15:55VAST Data's secret sauce
- 17:45Disaggregation and architectural shifts
- 19:54Inference as the new web app
- 21:20Favorite customer success stories
- 23:50Making data universal
