Benoit Tellier created JAMES-4224:
-------------------------------------
Summary: BlobId control the entropy
Key: JAMES-4224
URL: https://issues.apache.org/jira/browse/JAMES-4224
Project: James Server
Issue Type: Improvement
Components: s3
Reporter: Benoit Tellier
h3. Why ?
Today dedup storage identifier is a 256 bit hash encoded as base64url. With GC
info (1_690_) it takes 44 + 5 = 50 chars
Duplicated again and again in Cassandra (6 occurences) and s3 metadata store
128 bit entropy means 1,5 × 10⁻¹⁹ collision chance accross 10^10 objects
128 bit entropy means 21+6 = 27 char names so a 46% reduction...
This empowers a 5% db size reduction on Cassandra (tiered!) and a dramatic size
reduction of the Rados gateway metadata DB... Nice to take when it comes at low
cost.
This property is appliedonly at blob creation so can dynamically be changed.
h3. What ?
Give operator control, default to 256 but allow a lower 128
-Djames.blobid.entropy=128
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
