Query-side
score
Pause
Fresh-data path cost
Loaded dataset
Newly
written
Ordered
raw data
Ordered pre-
aggregated
Query cost
Accumulated
query runtime
×
Read-side
compute price
Fresh
data
continuously
Query
runtime
Fast answers
continuously
2
3
FOR CONTEXT
Where query cost and query runtime come from
QUERY-SIDE SCORE
=
2
×
3
· LOWER IS BETTER
ClickHouse
BEST
Snowflake
32×
WORSE
PERF PER $
Databricks
23×
WORSE
PERF PER $
BigQuery · Capacity
101×
WORSE
PERF PER $
Redshift Serverless
57×
WORSE
PERF PER $