Axel Hanikel created OAK-10311:
----------------------------------
Summary: Optimize SegmentBlob#equals for segment blobs that
originate from the same blob store
Key: OAK-10311
URL: https://issues.apache.org/jira/browse/OAK-10311
Project: Jackrabbit Oak
Issue Type: Story
Components: segment-tar
Affects Versions: 1.52.0
Reporter: Axel Hanikel
Fix For: 1.54.0
[SegmentBlob#equals|https://github.com/apache/jackrabbit-oak/blob/jackrabbit-oak-1.52.0/oak-segment-tar/src/main/java/org/apache/jackrabbit/oak/segment/SegmentBlob.java]
can be optimized for blobs originating from the same external blob store: In
that case, SegmentBlob#getBlobId() can be used to determine if the blobs are
equal or not. That change should avoid falling back to a byte-by-byte
comparison if the two blobs aren't equal and therefore don't have the same
ContentIdentity, which can be very expensive for larger blobs.
The optimization should be placed behind a code/feature toggle or system
property and disabled by default. The reason is that the contract of
[Blob#getContentIdentity|https://github.com/apache/jackrabbit-oak/blob/cc0521e1cf8dc10ad7a8d41a9f2d3fd2905e5c9b/oak-api/src/main/java/org/apache/jackrabbit/oak/api/Blob.java#L80-L82]
does not mention the special case where Blobs reside in the same blob store.
As part of this story, integration and benchmark tests should be created to
demonstrate gains in performance and correct behavior.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
