An open-source storage framework that enables building a Lakehouse architecture with compute engines including Spark, PrestoDB, Flink, Trino, and Hive and APIs
Today there are two Commit Coordinator Client (CCC) interfaces. One in delta-kernel-api and one in delta-storage (that delta-spark uses). Further, the current DynamoDBCommitCoordinatorClient implementation uses the delta-storage CCC interface. This means that any engine that wants to re-use that same DDB-CCC implementation has to depend on delta-storage, which brings in Hadoop and LogStore dependencies. This won't work for all engines (e.g.)
We should Kernel-ize the interface and implementations and make sure that any engine can use the implementations provided in this repo.
Today there are two Commit Coordinator Client (CCC) interfaces. One in delta-kernel-api and one in delta-storage (that delta-spark uses). Further, the current DynamoDBCommitCoordinatorClient implementation uses the delta-storage CCC interface. This means that any engine that wants to re-use that same DDB-CCC implementation has to depend on delta-storage, which brings in Hadoop and LogStore dependencies. This won't work for all engines (e.g.)
We should Kernel-ize the interface and implementations and make sure that any engine can use the implementations provided in this repo.