OmniRuntime
The big data features of Kunpeng BoostKit OmniRuntime are presented in the form of plugins to improve the performance of data loading, computing, and exchange from end to end.
Starting from Here
Application Acceleration Features
- OmniOperator
Uses native code (C/C++) to implement big data SQL operators to improve query performance.
- OmniStream
Uses native code (C/C++) to reconstruct the logic of Flink SQL and DataStream operators, improving query performance.
- OmniStateStore
Introduces lightweight modifications to Flink. It leverages techniques such as state caching and filtering to reduce Flink's reliance on RocksDB and improve end-to-end performance.
- OmniScheduler
Optimizes the Hadoop YARN Capacity Scheduler to enable more balanced resource allocation and higher utilization efficiency within a cluster.
- OmniShield
Performs data encryption and decryption within a hardware-based TEE environment, ensuring that the data remains secure and private even on the REE side.
- OmniMV
Recommends optimal materialized views from historical SQL queries using AI algorithms, reducing redundant computations and improving query performance.
- OmniAdvisor
Achieves automated parameter tuning and recommendations for Spark tasks through AI tuning algorithms, expert rule–based tuning algorithms, and operator acceleration algorithms, enhancing end-to-end parameter optimization.
- OmniShuffle
Reduces drive I/O overhead and enhances data analysis timeliness and cluster resource utilization through key features such as unified memory pool addressing, data exchange in memory semantics, and shuffle convergence.
- OmniHBaseGSI
Accelerates SingleColumnValueFilter queries by storing index data in an independent index table, thereby improving query performance.
- OmniData
Pushes operators of the big data engine to storage nodes to implement near-data computing, which reduces network bandwidth consumption and improves the query performance of the query engine.