Cloud data platforms bill on consumption. The difference between a $5k and a $50k month is rarely “we used the wrong vendor”; it is warehouse size, file layout, clustering that never pays for itself, or a Spark job that shuffles the same 8 TB every hour. These cheat sheets are platform maps: storage, compute, ingest, and the cost levers that actually move a bill.
Use them when you are landing a new workload, comparing a warehouse vs a lakehouse for a specific job, or preparing an interview that will ask you to size a cluster out loud. They are not marketing overviews. Each sheet assumes you already know what object storage is and need the names of the services and the failure modes.
Price a representative query and a representative pipeline, not a list price. Snowflake credits, Databricks DBUs, and BigQuery bytes scanned are not interchangeable units. Run the same workload through the Snowflake cost calculator, the Databricks cost calculator, the BigQuery calculator, and the warehouse comparison tool. Then read the architecture sheets for table-format and modeling choices that dominate cost after year one.
If the interview is “AWS vs Azure vs Snowflake,” answer with the constraint: existing identity, data gravity, and whether you need open table formats. SQL fluency still transfers; IAM and networking do not. Related: SQL cheat sheets and the full library.
← Back to Home