Hi All,
Thanks for making 1st Apache Fluss community call a blast, couldn't be
more happier seeing the community grow and thrive, Looking forward to more
and more otters join us and help us unleash the streams.
For those who couldn't surf along in today's meeting we won't let you miss
out on anything
Meeting summary
-
Python SDK development progressing well, using Rust binding approach for
better performance
-
Iceberg/Lance connector integration underway.
-
First Apache Fluss release targeted for September, focusing on resolving
licensing issues and packaging names
Notes
-
Python SDK Development
-
Using Rust binding approach for better performance than Py4j
-
Basic POC implemented with create table, write data, and read data
functionality
-
Plans to propose FIP for repositories
-
Iceberg/Lance Integration
-
Basic Lance catalog functions implemented and merged
-
Basic Iceberg catalog functions Underway
-
Writer and committer under development for Lance & Iceberg,
-
Initially supporting only LogTable due to Lance Java API limitations
-
Need for support of important AI data types like vector embeddings
and large binary objects for lance
-
Release Planning
-
First version targeted for September release
-
Need to resolve licensing issues and packaging names
-
Plan to release fluss-shaded artifacts this month to replace
Alibaba-published modules
-
Benchmark Integration
-
Proposal to include benchmark in CI process
-
Challenges with finding appropriate CI machine resources for
benchmarking
-
Community Growth and Engagement
-
Several new members introduced themselves and expressed interest in
contributing
-
Encouragement for newcomers to ask questions and get involved
-
People Interested in spark come forward and support connector.
-
Next Steps
-
Complete Iceberg and Lance connector integration by August
-
Drive the release of fluss-shaded artifacts in preparation for first
Apache fluss distribution
Recording
https://fathom.video/share/ULZfyhBJmSK2XKafHSjM4EnV5Et2kmFG
Best Regards,
Mehul Batra
