Change Description
The Kunpeng BoostKit OmniData feature pushes operators on the compute side to storage nodes and transmits the procession results to compute nodes through the network, reducing the amount of data transmitted over the network.
Version 1.0.0 has the following changes:
- Offloads Filter, Aggregation, and Limit operators.
- Accesses HDFS, including TXT, ORC, and Parquet file formats.
- Supports Spark.
Version 1.1.0 has the following changes:
- Supports connection to the homogeneous acceleration framework (HAF) to replace the server/client interfaces of the original gRPC communication pushdown framework and implements pushdown using annotations.
- Supports operator pushdown to Ceph storage nodes for processing.
Parent topic: V1.1.0