Comparison of Router Processing Full-Scale RPKI Data and IPv6-Only RPKI Data through Experimental Evaluation

This study provides a comprehensive experimental analysis of router performance when processing full-scale Resource Public Key Infrastructure (RPKI) data versus IPv6-specific RPKI implementations, addressing critical operational challenges posed by IPv4 exhaustion and accelerating IPv6 adoption. Through a controlled testbed utilizing EVE-NG network emulation and modified rpstir2 software with enhanced RPKI-to-Router (RTR) protocol capabilities, we quantitatively evaluated five key metrics: storage, synchronization time, CPU utilization, bandwidth efficiency, and packet loss. Our findings reveal that IPv6-only RPKI configurations achieve 73.8% memory reduction (11 MB vs 42 MB), 81.6% faster synchronization time (4.222 s vs 22.959 s), and 78.5% lower bandwidth consumption (0.246 Mbps vs 1.145 Mbps) compared to full-scale RPKI implementations. These improvements stem from processing 81,333 IPv6 Route Origin Authorizations (ROAs) versus 452,650 total ROA in full deployments. The results demonstrate a significant reduction in memory space and shortening of synchronization time when using an IPv6-only cache compared to a full-scale RPKI data cache. The study aims to provide a clear quantitative analysis of the performance implications for routers handling full-scale versus IPv6-only RPKI data, which is of great significance for the deployment of IPv6-only network in the future.