IPv4 Parcels and Advanced Jumbos (AJs)

All aspects of are imported as normative specifications for IPv4 parcels and AJs, with the exception of the following differences:

The IPv6 header includes a "Payload Length" field defined as the: "Length of the IPv6 payload, i.e., the rest of the packet following this IPv6 header, in octets". The IPv4 header instead includes a "Total Length" field defined as: "the length of the datagram, measured in octets, including internet header and data". These differences have no bearing for parcels/AJs, for which both the IPv6 Payload Length and IPv4 Total Length values carry identical codes. The same as for IPv6, IPv4 parcels encode the length "L" of the first segment in Total Length, while IPv4 AJs encode a jumbo type value in Total Length.

The IPv4 "Time To Live (TTL)" and IPv6 "Hop Limit" values are treated in exactly the same way in both protocol versions. In particular, the source sets the TTL/Hop Limit to an initial value and each router in the path to the destination decrements the TTL/Hop Limit by 1 when it forwards the parcel/AJ.

The same as for IPv6, the Parcel Payload Length field in the Parcel Payload Option and the Jumbo Payload Length field in the Jumbo Payload Option of IPv4 parcels/AJs encode the length of the IPv6 extension headers plus the length of the {TCP,UDP} header plus the combined length of all concatenated segments with their per-segment headers/trailers. Therefore, the length of the IPv4 header itself is not included in the Parcel/Jumbo Payload Length field the same as for IPv6. The IPv4 header length for IPv4 parcels and AJs is instead calculated from the IPv4 header Internet Header Length (IHL) field the same as for ordinary IPv4 packets.

Whenever an IPv4 address needs to be coded in an IPv6 address field, the address is coded as an IPv4-compatible IPv6 address as specified in .

When IPv6 parcels become subject to packetization, the node performing packetization inserts an IPv6 Extended Fragment Header in each packet produced to allow the final destination to restore the original parcel. Since some IPv4 paths may not be able to forward a packet that includes an IPv6 Destination Options header with an Extended Fragment Header option, and since IPv4 destinations that do not recognize the option may be inclined to drop the packet, the Identification field and Parcel Index/P/S codes must be encoded in a different way for IPv4. For each IPv4 packet created during packetization, the node instead includes a single 8-octet Identification Extension "pseudo-option" at the end of the IPv4 header, where the first octet encodes option Type value '0' ("End of Option List") , the next octet encodes the Index/P/S octet the same as for IPv6 (see: ) and the final 6 octets encode the 6 most significant octets of the parcel Identification. The node then writes the 2 least significant octets of the parcel Identification into the IPv4 header Identification field and sets the Don't Fragment (DF) flag to 1. The IPv4 destination then performs restoration by gathering up IPv4 packets that arrive with the same upper layer 5-tuple and with a pseudo-option coded as above where the Identification components in the IPv4 header and final 6 octets of the pseudo-option both match the other packets. The pseudo-option Index field determines the ordinal position of each packet segment to be concatenated into the restoration buffer, i.e., the same as for IPv6. (Note: if the IPv4 destination does not recognize the pseudo-option encoding, it simply processes the packet as a singleton IPv4 packet. This would result in correct behavior, but would fail to take advantage of Generic Receive Offload (GRO) benefits.)

When an IPv4 router or destination receives an intact Parcel/Jumbo probe, it can return an immediate Parcel/Jumbo Reply if necessary in an ICMPv6 Packet Too Big (PTB) message prepared the same as for IPv6 but wrapped in UDP/IPv4 encapsulation headers. The Reply will then traverse the IPv4 network on the reverse path to the source. When an IPv4 destination receives an intact Parcel/Jumbo probe, it should instead return an MTU Report "pseudo-option" in any IPv4 packet to be returned to the source. For each such IPv4 packet, the node includes a single 12-octet option at the end of the IPv4 header where the first two octets encode the option Type value '0' ("End of Option List") , the next 4 octets encode the MTU value to report to the source (i.e., the same as for IPv6) and the final 6 octets encode the 6 most significant octets of the probe Identification. The destination then writes the 2 least significant octets of the probe Identification into the IPv4 header Identification field then sets the Don't Fragment (DF) flag to 1. When the IPv4 source receives a packet that includes an MTU Report pseudo-option prepared as above, it first matches the IPv4 header Identification field with the least-significant 16 bits of the probe Identification. If the bits match, the source proceeds to match the most-significant 48 bits of the probe Identification with the value found in the final 6 octets of the pseudo-option. If these bits also match, the source then marks the destination as "Parcels/Jumbos supported" and records the MTU value found in the pseudo-option as the path MTU.

To support the IPv4 parcel/AJ header checksum calculation, the network layer uses modified versions of the {TCP,UDP}/IPv4 pseudo-header found in . Note that while the contents of the two IP protocol version-specific pseudo-headers beyond the address fields are the same, the order in which the contents are arranged differs and must be honored according to the specific IP protocol version. The IPv6 pseudo-header is found in , while the IPv4 pseudo-header is shown in . The similarities between the two pseudo-headers allows for maximum reuse of widely deployed code while ensuring interoperability.

Note: The same as for IPv6, the Index/P/S and Parcel Payload Length fields in the IPv4 parcel pseudo-header are replaced by the single 4-octet Jumbo Payload Length field in the IPv4 AJ pseudo-header.