Snowflake's Three Layers
Describe Snowflake's architecture: what are its three layers, and what does separating them buy?
Answer out loud in under a minute: a definition, one concrete example, and one pitfall or trade-off.
Hints
Hint 1
Think about what happens when two teams run heavy queries on the same data at the same time.
Solution
Show the approach
The short version
Three layers. Database storage holds data reorganised into a compressed columnar format in micro-partitions. Compute is virtual warehouses: independent clusters that run queries and share no resources with each other. Cloud services coordinates everything — authentication, access control, metadata, query parsing and optimisation.
It is a hybrid of shared-disk (one copy of the data, reachable by every warehouse) and shared-nothing (each warehouse's compute is its own). The consequence that matters: storage and compute scale separately, and many warehouses read the same data without contending.
Example
A data-science team's warehouse runs a long training-set scan while the finance team's warehouse runs its month-end reports on the same tables; neither slows the other, and each is sized and billed on its own.
Pitfall
Independence is not free: each warehouse warms its own local cache, and metadata operations go through the shared services layer, which is where cross-warehouse coordination lives.