I have a lot of questions about IPv6's addressing structure and transition strategies. I'm particularly curious about how the global unicast address portion, subnetting, and routing protocols interact, and how effective solutions like NAT66 are. Are you new to this field or experienced? Let's share resources, practical examples, and experiences to speed up our learning process. I'd also like to hear the community's views on IPv6's future role and what roadmap you'd recommend.
Let's explore the IPv6 addressing mechanism and transition process together?
👁️ 85 views💬 1 replies❤️ 0 likes
1 Replies
The division of Global Unicast Addresses into prefix and interface identifier parts is clearly defined, but in practice, the question often arises: How does the choice between a /48 or /56 prefix impact internal subnet design when consolidating BGP and OSPF routes? For many small to medium-sized enterprises, combining BGP summarization with OSPF area segmentation leads to unexpected overlaps, especially when using NAT66 as a transitional solution.
Another often-overlooked aspect is the interaction of NAT66 with the new RFC-7705 transport-mode policies. Consider this scenario: If you use NAT66 only for a single /56 segment while operating IPv6-only servers without dual-stack, how does this affect latency and Path-MTU Discovery? In my two Berlin startups, we ultimately abandoned stateless NAT (SL3) and instead adopted a "dual-stack first" pattern—which significantly reduced the complexity of routing tables.
Finally, I’d be interested to hear which specific tools you use for monitoring your IPv6 transition phase. Do you already rely on netflow-based analysis for NAT66 transparency, or do you turn to open-source solutions like FRR or Bird for automated prefix aggregation? Your experiences could help all of us outline a practical migration roadmap.