Moreover with a purpose to parallelize the switch even if blocking I/O is used, the code tries to write a small quantity of knowledge to each fd in a loop, so that the kernel will ship the packets in the background to multiple slaves at the same.
Also visit my web-site;
ufa656