A cache-friendly IPv6 LPM with AVX-512 (linearized B+-tree, real BGP benchmarks)

A cache-friendly IPv6 LPM (Longest Prefix Match) algorithm using AVX-512 has been developed. It uses a linearized B+-tree data structure to improve performance. Real-world BGP benchmarks show significant improvements. This could lead to faster and more efficient routing in IPv6 networks. Engineers can explore the code on GitHub.