Craig Nunes & Andre Leibovici, Datrium | VMworld 2017
summary
Claims a storage architecture that separates compute I/O from long‑term capacity so performance scales independently of capacity, using host flash for primary IO and a separate data cluster for persistent copies and cloud archive. Customers see decoupled persistent capacity mapped to public cloud, stateless hosts that survive multiple server failures, and roughly 80–90% backup coverage without separate backup software; bare‑metal containers gain about 10–15% performance. Demos and POCs validate the approach while the company reports ~130% quarter‑over‑quarter growth and targets large VM fleets (~440,000 VMs).
AI-written summary · about
