[shaklee3]

This is also the reason DPDK uses a highly-efficient ring buffer. I believe (old benchmarks) it's faster than LMAX.