IPv6 uses a different approach for packet fragmentation compared to IPv4. IPv6 specifies that the source node should attempt to determine the path MTU (Maximum Transmission Unit) before sending the packet. The path MTU is the largest packet size that can be transmitted over the network without requiring fragmentation. This is done using the Path MTU Discovery (PMTUD) technique.
If the path MTU cannot be determined or if the packet size exceeds the path MTU, fragmentation is still possible in IPv6. In this case, fragmentation is performed only by the source node, and not by any intermediate routers. The source node divides the packet into smaller fragments that can fit within the path MTU and adds a new header, called the Fragment header, to each fragment that contains an identification number and an offset value. The Fragment header is added only to the first fragment, while subsequent fragments contain only the identification number.
When the fragments reach the destination, they are reassembled based on the identification number and offset value in the Fragment header. The reassembly process is performed only by the destination node, and not by any intermediate routers.
IPv6 fragmentation is less common than in IPv4 due to the use of PMTUD, but it is still an important feature that ensures packet delivery in situations where fragmentation is necessary.