AI / Apache Paimon Interview questions
What is the difference between Paimon's and Apache Iceberg's core design philosophy?
Both are open table formats supporting ACID transactions, schema evolution, and time travel on top of object storage, but they optimize for different primary workloads:
| Aspect | Apache Paimon | Apache Iceberg |
| Core storage engine | LSM-tree per bucket, designed for frequent streaming upserts | Immutable Parquet/ORC/Avro files with metadata-tracked changes, historically append/merge-on-read oriented |
| Update model | Native, high-throughput streaming updates via merge engines | Row-level updates via merge-on-read or copy-on-write, evolving toward native deletion vectors |
| Origins | Flink community (FLIP-188 / Flink Table Store) | Netflix, focused on large-scale batch analytics correctness |
In short: Paimon's LSM-based design was built streaming-first with batch analytics as a compatible side effect, while Iceberg's file/metadata-log design was built batch-first with streaming/CDC support added on top — a difference in starting point that still shows up in which workload each performs best for out of the box.
Paimon's on-disk storage engine is built around:
Which project originated primarily to serve large-scale batch analytics correctness at Netflix?
More Related questions...
