OmniShuffle Combined with OmniOperator

Combine OmniShuffle and OmniOperator to leverage both features.

OmniOperator improves the operator execution efficiency and OmniShuffle optimizes the data interaction process to improve the end-to-end engine query performance.

OmniShuffle connects to the OmniOperator columnar vector data structure and stores shuffle data in the memory pool of the Open Computing Kit Daemon (OCKD) process, reducing drive I/Os and improving query performance.