RFC8950：实现IPv6网络上的原生IPv4路由

RFC8950规范允许在IPv6网络中直接路由IPv4流量，无需依赖隧道技术1。这一标准为大型网络提供了灵活的架构方案，使得核心网络可以运行纯IPv6、IP传输骨干网仅部署IPv6成为可能1。

根据演示实验，该功能已获得Juniper JunOS、Cisco IOS XR、Nokia SR OS、Mikrotik、Huawei、BIRD、FRR等多个平台的支持1。通过ContainerLab构建的测试拓扑验证了其可行性：在R1和R2仅运行IPv6、无IPv4接口的网络中，C1（源IP段10.0.0.0/24）和C2（源IP段100.64.0.0/24）成功建立了通信1。追踪路由显示，C1到100.64.0.2的数据包经由R1（172.20.20.2）、R2（172.20.20.8）、R3（172.20.20.5）最终抵达C21。在R1的BGP IPv4路由表中，10.0.0.0/24前缀的下一跳被设置为fe80::a8c1:abff:fe44:afb8这一IPv6链路本地地址1，由此实现了纯IPv6路由器对IPv4路由的承载。

RFC 8950 establishes a specification that enables IPv4 traffic to be routed across IPv6 networks without requiring tunnel encapsulation 1. This advancement addresses modern network architectures where core infrastructure operates exclusively on IPv6, such as in large-scale deployments running pure IPv6 at the backbone level and IP transport networks carrying only IPv6 traffic 1.

The specification is supported across multiple networking platforms, including Juniper JunOS, Cisco IOS XR, Nokia SR OS, Mikrotik, Huawei, BIRD, and FRR 1. A practical demonstration using ContainerLab validates the capability by establishing communication between two client routers—C1 and C2—connected through intermediate routers where some links operate purely on IPv6 without IPv4 interfaces 1. In this topology, C1 announces an IPv4 prefix of 10.0.0.0/24 via an IPv6 session, while R3 announces 100.64.0.0/24 1. Traceroute confirmation shows that traffic from C1 to 100.64.0.2 successfully traverses through the IPv6-only intermediate nodes, passing through R1 (172.20.20.2), R2 (172.20.20.8), and R3 (172.20.20.5) before reaching C2 1. The BGP IPv4 routing table on R1 demonstrates the mechanism in operation, displaying the 10.0.0.0/24 prefix with a next-hop address of fe80::a8c1:abff:fe44:afb8, an IPv6 link-local address 1.