System design interview question
Design a Distributed Multi-Tenant Data Warehouse and Query Optimization Platform for Enterprise Analytics (Like Snowflake at Scale)
Design a distributed, multi-tenant data warehouse system that ingests structured data from 10,000+ enterprise customers, each with 1-100TB of historical data. The system must support real-time data ingestion (100K events/sec globally), complex analytical queries (JOINs, aggregations, window functions), and enforce strict tenant isolation with sub-100ms query latency for 95th percentile. Customers span 6 geographic regions with varying regulatory requirements (GDPR, CCPA). The platform must handle concurrent query execution from thousands of users, manage storage costs across hot/warm/cold tiers, and guarantee data consistency while allowing horizontal scaling of compute and storage independently. Design the query execution engine, metadata management, resource allocation, and consistency guarantees.
Practice this question step by step with an AI interviewer that reviews your requirements, your architecture and your trade-offs, and scores the design.