How does tunneling enable IPv6 communication over IPv4 networks?

Tunneling is a transition technique in which packets belonging to one network-layer protocol are wrapped, unchanged, inside the packets of a different network-layer protocol so that they can be carried across a network built for the second protocol.

For IPv6 traffic crossing IPv4-only infrastructure, a dual-stack router or host at the tunnel entry takes the complete IPv6 packet (its own header and payload) and places it as the payload of a new IPv4 packet:

1. The source generates an ordinary IPv6 packet with a full IPv6 header and payload.

2. At the tunnel entry point, that entire IPv6 packet is encapsulated as the data portion of a new IPv4 packet, and the IPv4 header's Protocol field is set to 41 to mark the payload as an IPv6 datagram.

3. The resulting IPv4 packet is forwarded hop by hop through the IPv4-only network using ordinary IPv4 routing, since every intermediate router only ever sees a normal IPv4 packet.

4. At the tunnel exit point, the IPv4 header is removed, recovering the original, unmodified IPv6 packet, which is then delivered onward on the IPv6 side.

Cross-check: this is distinct from header translation, which rewrites an IPv6 header into an IPv4 header (or vice versa) instead of wrapping the packet, and from dual stack, where a device runs both protocol stacks natively without needing to wrap one protocol inside the other. Per RFC 4213 (Basic Transition Mechanisms for IPv6 Hosts and Routers), encapsulating IPv6 packets inside IPv4 packets is exactly how tunneling achieves IPv6-over-IPv4 connectivity.