Watch a client route survive a relay failure. The workbench separates detection,
path repair, reply selection, and resumed traffic so learners can see what the
mesh changes and what it cannot fix.
Core ideaAlternate paths prevent single relay failure
Routing basisLowest airtime-style cost, not just hop count
Failure signalMissed probes plus poor link metrics
LimitPartitioned nodes still need another link
TrySelect Kitchen relay offline with Probe interval 1.0 s and 3 Missed probes before down, then Play.
ObserveTraffic pauses until failure detection, then route cost shifts to an alternate path; Patio isolated never reaches root. Step confirms this readout.
ExplainHealing requires neighbor-loss detection and a complete lower-cost path; redundancy helps only when another connected route exists. Step exposes this mechanism.
Colour keywifi mobile identitycurrent / primaryreference / datasuccesscautionerror / failure
Scenarios
Detection Timing
Repair Sequence
Mesh Route Map
A relay has failed. Step through detection, route discovery, and traffic recovery.
Route repaired
Active routeCamera -> Patio -> Garage -> Living -> Root
Hop count3 mesh hops
Recovery estimate4.2 s
Airtime cost891
Diagnosis
Forwarded dataPath request or warningPath replyFailed relay or broken route
How To Read The Animation
Green data pathThe route currently carrying client traffic toward the root gateway. A changed green path means traffic has been rerouted.
Red dashed pathThe route that was good before the failure but now crosses a failed relay or unusable link.
Orange request waveA simplified path request broadcast. Real systems limit and rate-control these messages to protect airtime.
Purple replyA path reply travels back along a selected route. The selected route should have a lower cost than other available choices.
Quick Technical Reference
HWMPIEEE 802.11s defines the Hybrid Wireless Mesh Protocol, combining proactive root announcements with reactive path discovery.
Airtime metricWi-Fi mesh routing usually prefers links that consume less channel time. A cleaner high-rate link can beat a shorter poor link.
Detection delayThe workbench estimates detection from probe interval multiplied by missed probes. Real devices also use beacons, retries, and vendor logic.
Partition limitSelf-healing needs at least one complete path. If every path is broken, the correct outcome is no route, not instant recovery.
Try These Checks
Step the kitchen failureWatch the old route fail through Kitchen, then compare the new Patio -> Garage -> Living path.
Raise retry pressureHigher retry pressure increases the estimated recovery window and can change the path cost calculation.
Choose Patio isolatedThe learner should see that route discovery can fail when there is no connected path back to the root.
Click a mesh nodeClicking any non-root mesh node toggles its failure state, making it easy to test your own topology case.
Model Boundaries
This is a teaching model. It uses a small graph, a simplified airtime-style
cost, and RREQ/RREP-style names to show the idea of mesh repair. Actual
consumer and enterprise systems may use proprietary controllers, band steering,
dedicated backhaul radios, or controller-assisted decisions. The key concept
remains the same: detect a bad next hop, choose a better available path, and
resume forwarding only when a connected path still exists.