- IPv6-specific DNS records
- Reverse resolution
- Part summary
In modern networks, the DNS (Domain Name System) translates domain names into IP addresses that machines can use. With the introduction of IPv6, DNS had to adapt to support 128-bit addresses while maintaining backward compatibility with IPv4. This coexistence is especially important in dual-stack environments, where both IP versions operate simultaneously.
IPv6-specific DNS records
To associate a domain name with an IPv6 address, DNS uses a AAAA (quad-A) record, analogous to the "A" record for IPv4 addresses. The AAAA record explicitly maps a domain name to an IPv6 address.
Example:
ipv6.mydmn.org. IN AAAA 2001:66c:2a8:22::c100:68b
This record indicates that the domain
ipv6.mydmn.org resolves to the IPv6 address 2001:66c:2a8:22::c100:68b. It is possible, and even recommended for maximum compatibility, to associate the same domain name with several IP addresses, whether IPv4 (via an A record) or IPv6 (via an AAAA record). This allows IPv6-compatible customers to prefer IPv6, while ensuring IPv4-only clients remain supported.In addition, DNS supports reverse resolution, meaning it can look up the domain name associated to a given IP address. In the case of IPv6, this operation uses PTR records placed in the
ip6.arpa zone. This zone is specifically reserved for IPv6 reverse resolution. For IPv4, it is the in-addr.arpa zone.Reverse resolution
Reverse resolution of an IPv6 address follows a strict process:
- Expand the address into full hexadecimal notation (16 bytes, i.e. 32 hexadecimal digits). Example:
2001:66c:2a8:22::c100:68b
Becomes:
2001:066c:02a8:0022:0000:0000:c100:068b
- Reverse the order of each hexadecimal digit (nibble), separate them with dots and append
ip6.arpa:
b.8.6.0.0.0.1.c.0.0.0.0.0.0.0.0.2.2.0.0.8.a.2.0.c.6.6.0.1.0.0.2.ip6.arpa IN PTR ipv6.mydmn.org
This structure ensures standardized, unique reverse lookups in the IPv6 address space.
Please note: DNS queries can travel over either an IPv4 or IPv6. The transport protocol used has no effect on the type of records returned.
For example:
- A client connected over IPv6 can request an IPv4 record.
- A client connected over IPv4 can request an IPv6 record. The DNS server simply responds with the records it has, regardless of the query's transport protocol.
When a hostname is mapped to both IPv4 and IPv6, the choice of which address to use is governed by RFC 6724, which defines an address selection algorithm based on factors such as prefix preference, scope, and reachability. By default, IPv6 is generally preferred unless overridden by system or network configuration.
Important reminder: when embedding an IPv6 address in a URL (Uniform Resource Locator), it must be enclosed in square brackets (
[]). This avoids confusion between the colon (:) inside the IPv6 address and the colon separating the hostname from the port in the URL.Valid example:
http://[2001:db8::1]:8080
This ensures that the URL is processed correctly by both browsers and web servers.
Integrating IPv6 into the DNS system therefore relies on new record types, a strict method for reverse resolution, and precise selection and formatting rules to ensure routing compatibility and consistency.
Part summary
In this section, we explored the fundamental principles of IPv6 addressing. We began by examining the structure of IPv6 address: its 128-bit length, hexadecimal notation, and the simplification rules used to shorten repetitive sequences of zeros. This design enables IPv6 to overcome the limitations of IPv4's address space, while guaranteeing scalability and efficient hierarchy.
We then looked at the different categories of IPv6 addresses: unicast, anycast and multicast, detailing their scope, typical use and representation in the address space.
Next, we reviewed the methods of assigning IPv6 addresses within a local network, whether by manual configuration, via the DHCPv6 protocol, or using stateless autoconfiguration mechanisms such as those offered by NDP. These approaches enable devices to automatically generate their own address from the provided prefix and their MAC address (via EUI-64), while offering flexibility in terms of lifetime management and privacy.
We've also detailed how address blocks are allocated, starting from IANA, which distributes them to the five RIRs (Registered Internet Regions), and then to the ISPs, who redistribute them to their customers as subnets (often in /48, allowing 65536 /64 subnetworks). The distinction between Provider Aggregatable (PA) and Provider Independent (PI) blocks helps manage multihoming or provider-change scenarios.
We saw that DNS has adapted to IPv6 with the introduction of the AAAA record, and that reverse resolution mechanisms now rely on the
ip6.arpa zone. Importantly, DNS remains independent of the transport protocol used (IPv4 or IPv6), ensuring seamless interoperability in a dual-stack environment.IPv6 is therefore not just an incremental improvement over IPv4, but a complete redesign of the addressing system, built to meet both current and future challenges of the global Internet.
In the final part of this NET 302 course, we will move into practice and focus on network diagnostic tools.
Quiz
Quiz1/5
net3024.4
What type of DNS record is used to find the name associated with an IPv6 address?