Introduction
The multi-CPU distribution functionality can be configured to place data in 1 – 128 different host buffers with dynamic segment size, and in 1 – 64 different host buffers with static segment size. The algorithm used by the accelerator for placing a captured frame in a host buffer is based on flow information (see Ports, Flows, Host Buffers and Streams) as shown in this figure.
If the specified hash key cannot be calculated for a frame, the frame is distributed using hash key type 31, round-robin.
Illustration
This figure illustrates the multi-CPU distribution functionality.
