Dear Authors, I have carefully reviewed draft-ietf-6lo-path-aware-semantic-addressing-13 and would like to share some technical comments and suggestions. Overall, I find the proposal both interesting and thought-provoking. The draft presents a clear design direction for constrained low-power and lossy networks by embedding path and topology semantics into the address structure itself, thereby enabling forwarding decisions to be derived directly from the destination address rather than from conventional routing tables. In my view, this is the most distinctive aspect of the proposal: it redefines the relationship among addressing, topology representation, and packet forwarding, and provides a meaningful alternative to traditional state-heavy routing approaches in highly constrained environments.
From my reading, PASA is particularly well suited to relatively static, hierarchically organized LLN deployments where node locations and parent-child relationships remain stable for long periods. Under such assumptions, the use of semantic addresses and the Tree Address Assignment Function (TAAF) can significantly reduce forwarding state, control overhead, and memory requirements at intermediate nodes. The forwarding logic appears elegant in that prefix relationships in the assigned addresses naturally reflect topological containment, allowing routers to infer whether a packet should be forwarded upward, downward, or locally delivered. In this sense, PASA is not merely a compressed forwarding technique, but rather a specialized addressing-and-forwarding paradigm for stable tree-like networks. That said, I also have several questions and suggestions regarding the current design. First, the effectiveness of PASA seems to rely strongly on long-term topological stability. While this assumption is reasonable for some industrial, building automation, and monitoring scenarios, it may limit the applicability of the scheme in environments where router-level changes, subtree migration, or link reconfiguration occur more frequently. The draft acknowledges that router changes may trigger partial renumbering, but this issue appears fundamental rather than incidental. Since the address structure is tightly coupled with the topology, any structural change may propagate into address reassignment costs. It may therefore be helpful for the draft to further clarify the intended deployment boundary and explicitly distinguish between recommended and non-recommended scenarios. Second, the current TAAF demonstrates the feasibility of semantic hierarchical address construction, but its scalability may deserve further discussion. Because the address length grows with tree depth and sibling indexing, wide or irregular trees may approach the 64-bit limit relatively quickly. This suggests that the present TAAF is an effective proof of concept, but perhaps not yet a universally suitable assignment function for all static LLN topologies. It might strengthen the draft to discuss possible families of address assignment functions, or at least outline how alternative AAFs could better support large fan-out nodes, uneven hierarchies, or future incremental expansion. Third, the parent-selection process appears underspecified. The draft mentions a “first come first served” approach when multiple candidate parents advertise reachability, but in practice parent selection has substantial consequences for logical tree shape, path efficiency, load distribution, and future renumbering risk. Since the resulting address tree directly determines forwarding behavior, parent selection is not merely a deployment detail but an important part of the overall system design. I would suggest that the draft either define a minimal set of parent-selection considerations or provide more explicit operational guidance for implementers. Fourth, I believe the notion of “stateless forwarding” may benefit from more careful qualification. PASA indeed avoids conventional per-destination routing tables, which is a strong advantage. However, the system still depends on several forms of persistent state, such as parent-child relationships, address allocation state, neighbor registration information, and state stored in non-volatile memory. For that reason, it may be more precise to describe PASA as enabling forwarding without traditional routing-table state, rather than as a fully stateless system. A slightly more nuanced formulation may help avoid misunderstanding during technical review. Fifth, the security and privacy implications of semantic addressing deserve continued attention. Because the address structure reflects the path and hierarchical location of a node, it may reveal internal topology information and make address-space inference easier for an adversary. The draft already recognizes these concerns, including the possibility of topology disclosure and address exhaustion attacks, which I think is an important strength. However, given that addressing in PASA carries structural meaning, security considerations may need to be treated as a core design dimension rather than as a secondary add-on. It could be useful to further discuss whether separate internal and externally visible address forms, stronger registration controls, or additional protection against allocation abuse should be considered. Finally, I think the draft would benefit from more quantitative evaluation. The conceptual comparison with RPL and other existing approaches is useful, but additional measurements could make the trade-offs much clearer. For example, evaluations of memory overhead, control traffic, address-length growth under different tree structures, renumbering cost after topology change, and reliability under limited failures would substantially strengthen the case for PASA and help readers better understand both its advantages and its boundaries. In summary, I believe PASA is a valuable and original contribution, especially for static tree-structured LLNs where minimizing forwarding state is a primary design goal. Its core idea—making the address itself carry topological and forwarding semantics—is elegant and potentially impactful. At the same time, the proposal appears best viewed as a specialized framework for well-bounded deployment scenarios rather than as a general solution for all constrained networks. With further refinement in deployment scope, address-assignment generalization, parent selection, renumbering control, security design, and quantitative evaluation, the draft could become significantly stronger both academically and from a standardization perspective. Thank you for your work on this draft. I hope these comments are helpful and constructive. Best regards, Hongyan Chen
_______________________________________________ 6lo mailing list -- [email protected] To unsubscribe send an email to [email protected]
