Meshtastic opens review of next-hop routing improvements

Meshtastic opened review on June 21 for PR #10735, Next Hop improvements, from collaborator NomDeTom. The work landed in the part of meshtastic/firmware that decides how packets move from node to node, and that is where reliability shows up fast: if pathing is clean, messages arrive with fewer retransmits, but if it is not, users see delays, duplicates, and missed handoffs.

The concrete question running through this update is simple: what kinds of messages or node behaviors get better if next-hop routing improves? For direct messages, routers, repeaters, and mixed-role networks, the answer is likely fewer wasted transmissions, better path selection, and more dependable delivery when several devices can hear the same packet.

Meshtastic has already been pushing in this direction since version 2.6, when direct messages moved to next-hop routing instead of relying only on managed flooding. The project said that work had taken nearly 1.5 years of preparation, implementation, simulations, and real-life testing, and that the design had to handle mobile nodes, asymmetric links, and packet loss. The system also repurposed two unused bytes in the unencrypted header to track both the current relayer and the expected next hop.

The demand for that change was visible before the rollout. In a September 2024 GitHub discussion, a user in Portland, Oregon said direct messaging was poor and argued that next-hop-based routing would be more spectrum-efficient because it would reduce unnecessary repeats and keep paths more consistent. A September 2025 bug report then showed how much ride-or-die routing logic matters in the field: nextHop and next_hop could be set from the return traceroute path instead of the forward path on asymmetric routes, a failure that the reporter said made Meshtastic unresponsive for many users.

Meshtastic framed its 2.6 direct-message routing as the first release of its new routing algorithm and one of its most feature-packed releases since 2.0 in November 2022. By November 2025, the project had already added Zero-Cost Hops for favorite routers, underscoring how much value the community places on preserving hop budget and staying within the 7-hop limit on long paths.

That is why PR #10735 matters. In a busy mesh, the whole story comes down to whether the right relay gets chosen on the first try, so a packet reaches its destination without spending the network’s airtime on the wrong path.