swecrets

Judgment and Trade-offs

Judgment

Build trade-off vocabulary against ClickHouse's real competitors, and practice architecture-level decision-making.

What this module covers

  • Cloud data warehouses (Snowflake, BigQuery) — cost and latency profile
  • DuckDB — embedded/local OLAP, when local beats a server
  • Real-time OLAP siblings — Apache Druid, Apache Pinot
  • Sampling (SAMPLE clause) as a workload-fit lever
  • Approximate aggregate functions (trade-off discussion)
  • Workload-fit decision framework
  • Sharding (horizontal scale-out) vs. vertical scale-up
  • Distributed table engine

Lessons

  1. 1. ClickHouse vs. Snowflake vs. BigQuery25 min
  2. 2. ClickHouse vs. DuckDB: When Local Wins22 min
  3. 3. ClickHouse vs. Druid vs. Pinot20 min
  4. 4. When ClickHouse Is the Wrong Tool23 min
  5. 5. Sharding vs. Scaling Up: The Decision, Defended24 min
  6. 6. Cold Start Closer: An Architecture Review Scenario25 min