The 6LoWPAN Format
The Internet Protocol was designed to enable networking beyond the boundaries of any single network. The Internet is composed of many subnetworks, a number of which are typically traversed by a packet on the way from its source to its destination. For each type of subnetwork, there needs to be an “IP-over-X” specification that defines how to transport IP packets on it. The complexity of such specifications can vary considerably.
At the simple end of the complexity scale for IPv6-over-X specifications is IPv6 over Ethernet [RFC2464], which gets by on little more than four pages of ASCII text (plus boilerplate), mostly describing packet formats and formatsfor addresses to be used in these packets. The reason for this simplicity is that the link model used by IPv6 is closely aligned with the capabilities of Ethernet, enabling a very simple one-to-one mapping. In other cases, more work is required to map the services required by the IP layer onto the services provided by the lower layer. The “IP-over-X” specification can amount to a (sub)layer of its own, often called an adaptation layer.
One such example is defined by IPv6 over the point-to-point protocol (PPP) [RFC5072]. While this specification itself contains only a little more than 10 pages of normative text, it includes by reference the 50-page PPP specification [RFC1661] and, depending on options selected, might invoke specifications as complex as robust header compression (ROHC [RFC3095], more than 150 ...