3.2.7. Data-Path Validation and Loop Detection

The low-power and lossy nature of LLNs motivates RPL's use of on-demand loop detection using data packets. Because data traffic can be infrequent, maintaining a routing topology that is constantly up to date with the physical topology can waste energy. Typical LLNs exhibit variations in physical connectivity that are transient and innocuous to traffic, but that would be costly to track closely from the control plane. Transient and infrequent changes in connectivity need not be addressed by RPL until there is data to send. This aspect of RPL's design draws from existing, highly used LLN protocols as well as extensive experimental and deployment evidence on its efficacy.

The RPL Packet Information that is transported with data packets includes the Rank of the transmitter. An inconsistency between the routing decision for a packet (Upward or Downward) and the Rank relationship between the two nodes indicates a possible loop. On receiving such a packet, a node institutes a local repair operation.

For example, if a node receives a packet flagged as moving in the Upward direction, and if that packet records that the transmitter is of a lower (lesser) Rank than the receiving node, then the receiving node is able to conclude that the packet has not progressed in the Upward direction and that the DODAG is inconsistent.