Details
-
New Feature
-
Status: Open
-
Major
-
Resolution: Unresolved
-
3.1.0
-
None
-
None
Description
People in Spark Scalability & Reliability Sync Meeting have discussed a lot about remote (disaggregated) shuffle service, and plan to do a reference implementation to help demonstrate some basic design and pave the way for a future production grade remote shuffle service.
There are already two pull requests to enhance Spark shuffle metadata API to make it easy/possible to implement remote shuffle service (PR 28616, PR 28618). Creating a remote shuffle service reference implementation will help to validate those shuffle metadata API.
Attachments
Issue Links
- is related to
-
SPARK-25299 Use remote storage for persisting shuffle data
- Open