Compare

Apache Iceberg vs Apache Paimon

Apache Iceberg and Apache Paimon are free, open-source table formats for analytic data. Both support table snapshots, time travel, schema evolution, self-hosted deployment, and bring-your-own storage.

Compare them if you need a table format that fits your existing query engines, streaming or batch work, storage systems, and team support options.

Round by round

  1. Price and licensingEven

    Both Apache Iceberg and Apache Paimon are free and provide a free plan. Iceberg is open source under the Apache License, Version 2.0. Paimon is described as an open-source data lake platform, with connectors and storage plugins chosen for your environment. Neither lists a paid plan or a free trial. Since the facts show the same free pricing position for both, neither has a price advantage.

  2. Core data featuresApache Paimon wins

    Iceberg is an open table format for analytic datasets and huge analytic tables. It supports SQL merges, updates, targeted deletes, schema changes without rewriting, hidden partitioning, snapshots, rollback, atomic updates, serializable isolation, concurrent writes, and compaction. Paimon supports ACID transactions, time travel, schema evolution, fast scan planning, incremental clustering, primary-key streaming updates, configurable merge engines, automatic small-file merging, and Z-order compaction. It also lists vector, full-text, blob, and global-index features. Paimon wins because its facts cover both analytic table management and more realtime and multimodal workloads.

  3. Engine and data integrationsApache Paimon wins

    Iceberg tables can be used at the same time by Spark, Trino, Flink, Presto, Hive, and Impala. Its documented integrations include Spark, Flink, Kafka Connect, Hive, and catalogs such as AWS Glue, JDBC, and Nessie. Paimon lists Flink, Spark, Hive, Trino, Presto, StarRocks, and Doris, plus CDC pipelines for MySQL, PostgreSQL, Kafka, MongoDB, and Pulsar. Supported operations vary by engine, and Paimon can publish Iceberg metadata for Iceberg readers. Paimon wins for the broader listed engine, CDC, and interoperability options.

  4. Setup and operationEven

    Both products are self-hosted and use storage that you provide. Iceberg supplies source archives, runtime JARs for Spark and Flink, and cloud bundles for AWS, Google Cloud, and Azure. Paimon supplies engine JARs, filesystem plugins, and a Java API bundle. Paimon engine integrations need a matching connector or built-in integration, catalog and warehouse access, and shared storage for distributed clusters. Iceberg also depends on engines and catalogs, but the facts do not establish that either product is simpler overall, so this is a tie.

  5. Where they runEven

    Both list API, Linux, macOS, self-hosted, and Windows platforms. Iceberg is delivered through source archives, Spark and Flink runtime JARs, and AWS, Google Cloud, and Azure bundles. Paimon offers engine JARs, filesystem plugins, and a Java API bundle for query engines or embedded Java use. Paimon also documents a Python client whose core table reads and writes do not need a JVM or a running Flink or Spark cluster. The listed platform coverage is otherwise the same, so neither wins overall.

  6. Community supportApache Iceberg wins

    Iceberg provides community discussions through a developer mailing list and an Apache Iceberg Slack workspace. Its community page also lists public user and developer mailing lists. Paimon directs users to its user mailing list and GitHub issue tracker for help and issue reporting. Both provide public ways to ask for help, but Iceberg lists an additional dedicated Slack workspace alongside mailing lists. On the stated support options, Iceberg wins.

  7. Security and encryptionApache Iceberg wins

    Iceberg supports table encryption with predefined AWS, Azure, or Google Cloud KMS clients, or a custom KMS client. Its catalogs must preserve the table's encryption key ID throughout the table's lifetime. Paimon says trust and authorization boundaries are generally enforced by the surrounding catalog, engine, service, operator configuration, and storage authorization. Its filesystem guide documents server-side encryption headers and settings for AES256, KMS, or SM4. Iceberg wins because its facts define table-level encryption choices and a key-ID requirement, while Paimon's controls rely more on surrounding systems.

The facts side by side

  • Free tierPriceFree tier
  • —Public API—
  • —Open source—
  • YesFree tierYes
  • NoFree trial—
  • —Web—
  • ✓Windows✓
  • ✓Mac✓
  • ✓Linux✓
  • —iOS—
  • —Android—
  • —Free plan✓
  • tablesData scopetables
  • ✓Dataset branching✓
  • ✓Point-in-time rollback✓
  • tableSnapshot granularitytable
  • bring_your_ownStorage backendbring_your_own
  • self_hostedDeployment modelself_hosted
  • #6Standing on Data Version Control Tools#5
  • Pick Apache Iceberg if
  • You need a free Apache-licensed format for large analytic tables.
  • You want SQL merges, updates, deletes, snapshots, and rollback.
  • Your team uses Iceberg’s listed engines, catalogs, or Slack community.
  • Pick Apache Paimon if
  • You need streaming updates, CDC pipelines, or configurable merge engines.
  • You want multimodal, vector, full-text, or blob table capabilities.
  • You plan to use PyPaimon with Python, Ray, PyTorch, or Pandas.

By what you need

  • Apache IcebergLarge analytic tables. Choose Iceberg for its stated focus on huge analytic tables, SQL changes, hidden partitioning, snapshots, and concurrent writes.
  • Apache PaimonRealtime lakehouse updates. Choose Paimon for primary-key streaming updates, merge engines, changelog producers, and batch and streaming processing.
  • Apache PaimonCDC from operational systems. Choose Paimon for listed CDC pipelines covering MySQL, PostgreSQL, Kafka, MongoDB, and Pulsar.
  • Apache PaimonBroad query-engine choices. Choose Paimon for its listed Flink, Spark, Hive, Trino, Presto, StarRocks, and Doris integrations.
  • Apache IcebergTable encryption controls. Choose Iceberg for documented KMS client options and its requirement to preserve the table encryption key ID.

Verdict. Choose Apache Iceberg for a free, Apache-licensed table format centered on large analytic tables, SQL data changes, snapshots, rollback, concurrent writes, and documented encryption controls. Choose Apache Paimon for realtime lakehouse work, primary-key updates, CDC pipelines, multimodal data, and Python-based access. Both are self-hosted, run on the same listed platforms, use storage you provide, and support table-level time travel and schema evolution. Your existing engines, connectors, catalogs, and storage setup should decide the final choice.

Questions

Are Apache Iceberg and Apache Paimon free?

Yes. Both have a free plan. Iceberg is open source under the Apache License, Version 2.0; Paimon is described as an open-source data lake platform.

Which supports realtime updates?

Paimon’s primary-key tables support large-scale streaming updates with configurable merge engines. Iceberg supports concurrent writes and SQL merges, updates, and deletes.

Which has more listed integrations?

Paimon lists Flink, Spark, Hive, Trino, Presto, StarRocks, and Doris, plus CDC pipelines. Iceberg lists Spark, Trino, Flink, Presto, Hive, and Impala, plus catalogs and Kafka Connect.

Does Paimon work with Iceberg readers?

Yes. Paimon can publish Iceberg metadata so applications can read existing Paimon data files through Iceberg connectors. Writers and maintenance remain in Paimon.