The current generation Internet protocol IPv4 is nearing address exhaustion, and wide deployment of its successor IPv6 is just around the corner. Some ISPs have already begun to migrate their customers to IPv6, either natively or by some transition mechanisms like DS-Lite. Even though many people would like to stick to IPv4 for good reasons, I’d like to give three practical reasons why early deployment of IPv6 at home could make sense.

Better routing than your provider

An IPv6 tunnel might give you slightly better latency to major sites like google.com or facebook.com. My ISP (Unitymedia) peers directly with Hurricane Electric, an IPv6 tunnel broker and maintainer of the world’s biggest native IPv6 network. The distance to their nearest tunnel server tserv1.fra1.he.net is just 15 ms (5 hops), and latency to google.com via IPv6 is just 16 ms (7 additional hops). In comparison, latency via Unitymedia’s IPv4 network is about 19 ms (13 hops), or about 20% more than via IPv6. This is presumably due to suboptimal routing within Unitymedia’s network.

Overall bandwidth is unaffected: In theory, protocol overhead is minimal. The tunelling mechanism used by Hurricane Electric, 6in4, only has about 1.5% protocol overhead (MTU of 1500 bytes for IPv4 vs. 1480 for IPv6). In practice, download and upload speeds are the same, whether I use IPv6 or IPv4. I conducted simple tests that maxed out my 64/5 MBit/s cable connection at home, and the Hurricane Electric tunnel can easily provide that bandwidth.

One DNS for all

I have several appliances at home which I access from my computer (PVR, Fire TV, mobile phone etc.). I used to access them by IP address, which is cumbersome and doesn’t work from outside. Using IPv6 allows me to assign globally valid DNS names that resolve to globally routable IPv6 address. So if you have a Fire TV e.g. at 2001:db8::1234, you could add an AAAA record for firetv.example.com. that resolves to this address. If you want to access your device, you can use this DNS name regardless of whether you’re in your LAN or outside at another IPv6 enabled site (if you so wish).

Before IPv6, i.e. with IPv4 + NAT, this was only possible with splitted DNS where you have a local DNS server at your site that resolves to RFC 1918 addresses, and a public DNS server that announces for example the IP address of your router.

Shorter IP addresses thanks to unique local unicast addresses

Even if you have neither an IPv6 tunnel nor a local DNS server, or you still want to use site-local IPv4 addresses for whatever reason, you can still profit from IPv6. One of the counterparts of RFC 1918 addresses in IPv6 are Unique Local Unicast Addresses (ULA) defined in RFC 4193. The IPv6 specification reserves the block fc00::/7 for numbering of local networks. These are addresses local to a specific site and should never be routed, similar to the RFC 1918 blocks known from IPv4 (10.0.0.0/8, 172.16.0.0/12 and 192.168.0.0/16).

If you e.g. advertise the prefix fd33::/64 within your network, you would effectively shorten the length of IP addresses you have to type to reach your target host. Instead of typing e.g. 10.8.0.1 to reach your router, you could assign the address fd33::1, which is actually a character less to type. However, if you want to use these addresses in URIs to access them via browser, you have to wrap them in brackets ([]), which might make these addresses longer. Also, this only works if your devices support manually assigned IPv6 addresses or you use DHCPv6, which is not always possible or desired.

Conclusion

This post should give you three simple benefits of IPv6 at your local LAN. Are these the most important advantages of IPv6? Of course not. Globally routable addresses and Stateless Address Autoconfiguration are other benefits. Should you deploy IPv6 at via a tunnel broker? If you like to tinker, go ahead. If your life depends on IPv4, probably not. Are there real disadvantages for IPv6 in local LANs? Of course, but this is a topic for another blog post.