1.4IPv6 and address resolution

IPv4 address exhaustion is already a real-world constraint, and network designers are forced to judge whether to "keep expanding new sites or services on IPv4," "migrate to IPv6," or "run both side by side." Beyond IPv6 fundamentals, this section covers choosing migration technologies—how to combine limited IPv4 resources with future-proof IPv6.

1.4.1IPv6 address notation and types

• An IPv6 address is 128 bits, expressed as eight groups of 16 bits each in colon-separated hexadecimal (e.g., 2001:0db8:0000:0000:0000:0000:0000:0001). A run of consecutive 0000 groups can be abbreviated once, using ::, and leading zeros within each group can be omitted (2001:db8::1).
• A global unicast address is an IPv6 address uniquely reachable across the internet (in the 2000::/3 range). A link-local address (fe80::/10) is valid only within the same link (the same segment) and is never forwarded across a router. IPv6 interfaces are automatically assigned a link-local address, which serves as the basis for functions such as neighbor discovery.

1.4.2ARP and NDP (address resolution)

• ARP (Address Resolution Protocol) is the protocol used in IPv4 environments to resolve the MAC address corresponding to a destination IP address by broadcasting a query. It is used within the same segment to identify the MAC address of the next device to forward to.
• NDP (Neighbor Discovery Protocol) provides the IPv6 equivalent of ARP's function, using ICMPv6 messages (Neighbor Solicitation/Neighbor Advertisement, etc.); the key difference from ARP is that it performs address resolution via multicast rather than broadcast. It also underlies automatic address configuration via router advertisements (SLAAC).

Suppose you are the network designer at an ISP-like operator wanting to allocate additional IPv4 addresses to existing customers, but facing the constraint that the available IPv4 address stock is nearly exhausted. One option is NAT/NAPT, sharing private addresses so a single global IPv4 address serves many customers, but this carries side effects such as port exhaustion and compatibility issues with certain applications (e.g., P2P communication). Another option is to gradually migrate new customers to IPv6, but this would break communication compatibility with existing IPv4-only services; adopting dual stack (where a single device holds both IPv4 and IPv6 address stacks simultaneously and chooses which to use depending on the peer's protocol) lets you accommodate new customers primarily on IPv6 while preserving backward compatibility with IPv4 services, curbing IPv4 address consumption while migration proceeds. If a situation later arises where IPv6-only devices need to communicate with IPv4-only devices, conversion technologies become necessary, such as tunneling (carrying IPv6 packets encapsulated over an IPv4 network) or NAT64 (translating between IPv6 and IPv4). Rather than "migrating every customer to IPv6 right now," weighing compatibility with existing assets, migration cost, and future address demand and choosing a phased migration centered on dual stack is the realistic design judgment under IPv4 exhaustion.

1.4.3Section summary

• An IPv6 address is 128 bits in colon-separated hex; a run of zeros can be abbreviated once with ::
• ARP resolves via broadcast; NDP resolves via ICMPv6 multicast and underlies SLAAC
• Under IPv4 exhaustion, center the approach on dual stack and combine NAT/NAPT, tunneling, and NAT64 as required