2.8. IPv6 Addressing

💡 First Principle: IPv6 was created because IPv4 ran out of addresses. With 128 bits (vs IPv4's 32), IPv6 provides 340 undecillion addresses—enough for every grain of sand on Earth. But IPv6 isn't just "bigger IPv4"—it's a fundamental redesign. Gone are broadcasts (replaced by multicast). Gone is ARP (replaced by Neighbor Discovery). Gone are variable-length headers.

What happens if you ignore IPv6: Mobile carriers have already run out of IPv4 addresses—most phones use IPv6 by default. Major cloud providers (AWS, Azure, Google) run IPv6 natively. If you configure a server with only IPv4, mobile users may experience slower connections through carrier-grade NAT, or fail to connect entirely. IPv6 isn't future technology; it's current technology that you'll encounter in production.

Consider troubleshooting this: A user reports they can access some websites but not others. You check DNS—resolving fine. Ping works. But curl fails. The issue? The destination has both A (IPv4) and AAAA (IPv6) records. The client prefers IPv6, tries to connect over the broken IPv6 path, and times out. Understanding dual-stack behavior is now a core troubleshooting skill.

The intimidation factor: IPv6 addresses look scary (2001:0DB8:0000:0001:0000:0000:0000:0001), but they follow simple rules for shortening. Once you learn the abbreviation rules, they're no harder than IPv4.