Self-Designing Approach for Managing Large Datasets
This startup emerging from Stratos Idreos’ Data Systems Laboratory aims to commercialize a first of its kind, self-designing data system that can take trillions of forms to perfectly and continuously adapt to the data, queries, and usage patterns of customers as well as to the hardware and pricing policies of cloud providers. The system minimizes cloud costs and eliminates tuning, administration, and performance regression costs. It maximizes system performance (by more than 10x compared to existing data systems) and evolves with the application over time as new features and usage patterns emerge. In addition, it supports SQL and Key-Value APIs and applies to transactions (OLTP), real-time analytics (HTAP), in- memory systems, as well as data management for AI.