COF-C03 Snowflake Architecture: How Storage, Compute, and Cloud Services Work Together

About this video

Snowflake architecture is often introduced as three layers: storage, compute, and cloud services. For the SnowPro Core COF-C03 exam, simply memorizing those layers is not enough. In this video, we break down how Snowflake’s architecture actually works and how each layer affects query execution, workload design, security, performance, and troubleshooting. We’ll walk through the roles of database storage, virtual warehouses, and Cloud Services, then trace a query from authentication and privilege checks through optimization, warehouse execution, storage access, and result delivery. Key topics include: Snowflake’s multi-cluster shared data architecture Persistent storage and micro-partitions Virtual warehouses and compute isolation Scale-up versus multi-cluster scaling Cloud Services, metadata, and query optimization Partition pruning Query execution flow Warehouse queueing and concurrency Local and remote spilling Workload isolation Warehouse privileges versus object privileges Resource monitors and cost control The important exam skill is learning to connect symptoms to the correct architectural layer. A slow query does not automatically mean you need a larger warehouse. High queueing, poor pruning, memory pressure, and authorization failures have different causes and require different responses. Understanding those distinctions will help you reason through COF-C03 architecture and troubleshooting questions instead of relying on memorized feature definitions. #Snowflake #COFC03 #SnowProCore