Node Placement Strategy Principles and methods for selecting repeater sites that maximise coverage and network reliability. Repeater Placement Principles The Three Rules of Repeater Placement Every successful LoRa mesh deployment rests on three placement rules. Violate any one of them and the network will under-perform regardless of hardware quality, antenna gain, or software tuning. Rule 1 - Height Above Terrain LoRa operates in UHF spectrum where propagation is quasi-optical. The band depends on region: 902 - 928 MHz in the United States, 863 - 870 MHz in the EU, and 433 MHz in some regions. In the US, use 902 - 928 MHz — 868 MHz is the EU ISM band and is not available for Part 15 LoRa here, and 433 MHz US operation is tightly restricted under Part 15. (Note: 433/868/915 MHz are all UHF; SHF begins at 3 GHz.) The radio horizon - not the visual horizon - sets the hard ceiling on range. Antenna elevation is the single most powerful lever available to a network planner. Doubling antenna height does not double range; it multiplies it by roughly √2 (~1.41×). Using the standard refraction-corrected (4/3-earth) radio horizon formula d(km) ≈ 4.12 × √(h in metres), a node at 2 m AGL (above ground level) has a radio horizon of about 5.8 km on flat earth. Raise it to 20 m and the horizon extends to ~18 km. Raise it to 50 m and you reach ~29 km. Every metre of height is worth more at low elevations than at high ones - the return diminishes as you climb. (The purely geometric/optical horizon uses the smaller coefficient 3.57 × √h; the 4.12 figure includes normal atmospheric refraction.) Practical target: For a community repeater, aim for a minimum of 10 - 15 m AGL. A rooftop mount, water tower, or hilltop installation should be the first option explored. Rule 2 - Line-of-Sight to the Coverage Area A repeater serves areas it can "see" far better than areas it cannot. Because LoRa uses relatively low frequencies compared to Wi-Fi, it can diffract around moderate terrain features - but non-line-of-sight coverage via diffraction is weak and unreliable, and penetration through significant ridgelines or dense urban canyons is limited. The Fresnel Zone (a football-shaped region around the direct path) must be at least 60% clear for reliable propagation; 100% clearance of the first Fresnel zone is the gold standard. At 915 MHz the first Fresnel zone radius at the midpoint of a 10 km path is approximately: r₁ = 17.32 × sqrt(d / (4f)) [equivalently 8.66 × sqrt(d / f)] = 17.32 × sqrt(10 / (4 × 0.915)) ≈ 17.32 × sqrt(2.73) ≈ 28.6 m (d in km, f in GHz) By the 60% rule, an obstacle begins to impair the link once it intrudes more than ~40% into the first Fresnel zone - i.e., once it rises into roughly the central ~17 m of the ~29 m radius at midpoint (about 11 - 12 m of intrusion). Full clearance of all ~29 m is the gold standard. Use terrain analysis tools (HeyWhatsThat, SPLAT!, CloudRF, or the terrain profile view in Google Earth) to verify clearance before climbing a tower. Rule 3 - Overlapping Coverage with Adjacent Repeaters A mesh node at the edge of one repeater's coverage zone must be within range of at least one other repeater. This overlap provides two things: redundancy if one repeater fails, and smooth routing as a mobile node moves between coverage zones. As a planning rule of thumb, target roughly 20 - 30% geographic overlap between adjacent repeater footprints. These are illustrative heuristics rather than measured thresholds: too little overlap (very roughly under ~10%) tends to leave dead corridors, while very heavy overlap (over ~50%) adds little extra reliability for the extra sites. The Minimum Viable Repeater Principle One well-placed hilltop or rooftop repeater at 30 m AGL can cover 300 - 800 km² of flat to rolling terrain. Ten ground-level nodes scattered across the same area will collectively cover far less, generate more channel congestion, and create a fragile, multi-hop routing mess. This principle is counterintuitive for operators coming from Wi-Fi, where more access points always help. In LoRa mesh, channel time is shared by all nodes. Each additional ground-level node adds noise and contention while contributing minimal coverage. Rule of thumb: Before deploying any new ground-level node, ask whether a single elevated repeater could solve the coverage problem instead. In almost every case, it can. Dead Zone Identification Methods Dead zones are coverage gaps where a mesh client cannot hear any repeater. Identifying them systematically before deployment prevents costly re-work later. Method 1 - Terrain Masking Analysis Load a digital elevation model (DEM) of your target area into a tool such as SPLAT!, Radio Mobile, or CloudRF. Run a viewshed analysis from each proposed repeater site. Areas outside all viewsheds are your primary dead zones. This takes minutes and costs nothing. Method 2 - Wardrive / Walk Survey Deploy a mobile node running MeshMapper or a GPS-tagged Meshtastic client. Drive or walk the target area while logging signal reports (RSSI and SNR) from known repeaters. Overlay the log on a map. Holes in the RSSI heat-map are dead zones. This method catches dead zones that terrain analysis misses (e.g., urban canyons, underground parking, indoor spaces). Method 3 - Predictive Modelling with Terrain Obstruction Factor After estimating coverage radius by terrain type (see the Coverage Radius Estimation page), draw coverage circles around each repeater. Grid cells covered by zero circles are dead zones. Adjust for known terrain features: a ridge intruding the line-of-sight path adds diffraction loss that depends on the ITU-R P.526 diffraction parameter (the ridge's height relative to the direct line, its position along the path, and frequency), so model it per-path rather than applying a fixed percentage. A ridge near one endpoint, at the midpoint, or barely grazing the Fresnel zone all produce very different losses. Method 4 - Community Spot Reporting For an existing network, a simple form where users report "no connectivity" locations aggregated on a shared map is surprisingly effective. Each report is a data point; a cluster of reports identifies a dead zone. The disadvantage is latency - it only works after deployment. Using Terrain Analysis to Find Optimal Sites Before Going Outside The cheapest field test is the one you never have to make. Modern terrain analysis tools allow a planner to shortlist candidate sites from a desk and eliminate poor candidates before anyone puts boots on the ground. Download a DEM. USGS 1/3 arc-second DEMs (10 m resolution) are freely available for the contiguous USA via The National Map. Import into QGIS, SPLAT!, or similar. Identify candidate high points. Use the DEM hillshade layer to find hilltops, ridgelines, and tall structures within your target area. Water towers and tall buildings are often not in the DEM - add them manually. Run viewsheds. For each candidate, run a viewshed at your estimated antenna height (e.g., 15 m AGL). Note the percentage of target area covered. Rank candidates. Sort by coverage area. A single site covering >80% of target area is a strong anchor candidate. Check access and permission. The best RF site is worthless if you cannot install equipment there. Check ownership, zoning, and access before committing to a site. Field-verify top candidates. Visit only the top 2 - 3 sites. Confirm elevation, access, power availability, and backhaul options (if applicable). This workflow can substantially reduce the number of field visits required compared to ad-hoc site hunting, and produces better placements because decisions are based on quantitative terrain data rather than intuition. Coverage Radius Estimation by Terrain Type The Radio Horizon Formula The theoretical radio horizon for a single antenna at height h metres above a smooth spherical earth, accounting for standard atmospheric refraction (the 4/3-earth model), is: d (km) = 4.12 × √h_m (radio horizon, 4/3-earth refraction) For a link between two antennas at heights h₁ and h₂ the total radio horizon is: d_total (km) = 4.12 × (√h₁ + √h₂) Note: the purely geometric (optical) line-of-sight horizon uses the coefficient 3.57 × √h. The 4.12 coefficient adds the standard 4/3-earth refraction correction and is the value used for radio horizon throughout this book (consistent with the mountain-and-complex-terrain and repeater-placement pages). There is no latitude term in this formula. This is the maximum possible range on a flat, unobstructed earth - a theoretical ceiling, not a usable-coverage promise. Real terrain, vegetation, and buildings reduce this significantly. The formula provides the ceiling; terrain obstruction factors bring it down to a realistic estimate. Terrain Obstruction Factors Apply these multipliers to the radio-horizon distance to get a realistic coverage radius. The factor bands below are illustrative planning heuristics, not measured constants - no Tier-1 standard assigns these exact multipliers to named terrain classes. The underlying propagation behaviour is described in ITU-R P.526 (diffraction), P.833 (vegetation), and P.1411 (urban/suburban short-range); use those for rigorous modelling. Environment Type Obstruction Factor Effective Coverage (%) Notes Flat open (farmland, desert, water) 0.85 - 1.00 85 - 100% Near-theoretical range; ground reflections can help at low angles Gentle rolling terrain 0.65 - 0.80 65 - 80% Moderate ridge shadowing; elevated repeaters mitigate well Suburban (low-rise, gardens) 0.50 - 0.65 50 - 65% Houses and trees add 5 - 15 dB of excess loss (ITU-R P.1411 range) Dense forest / jungle 0.35 - 0.55 35 - 55% Vegetation loss ~0.2 - 0.5 dB/m through canopy at 915 MHz (ITU-R P.833 / Weissberger); saturates beyond ~14 m depth Urban (mid-rise, 3 - 8 floors) 0.30 - 0.50 30 - 50% Building diffraction dominates; rooftop-to-rooftop paths much better Dense urban (high-rise, canyons) 0.15 - 0.30 15 - 30% Multipath + shadowing severe; per-block coverage planning needed Mountainous / complex terrain 0.20 - 0.60 20 - 60% Highly variable; valleys may have near-zero coverage from a single site Worked Examples The coverage areas below are theoretical maxima derived from the radio-horizon ceiling times an obstruction factor. They are not the usable coverage you should plan around. In practice a single well-placed repeater at ~30 m AGL delivers on the order of 300 - 800 km² of reliable coverage in flat/rolling terrain, and far less in forest or urban settings - see the repeater-placement-principles and repeater-grid pages for realistic per-terrain planning densities. Treat the figures here as "maximum, not expected." Example 1 - 30 m Tower on Flat Land A community group installs a repeater at the top of a 30 m self-supporting tower in flat agricultural land. The mobile clients they serve have antennas at 1.5 m AGL. Radio horizon (tower): 4.12 × √30 = 4.12 × 5.48 = 22.6 km Radio horizon (client): 4.12 × √1.5 = 4.12 × 1.22 = 5.0 km Total radio horizon: 22.6 + 5.0 = 27.6 km Obstruction factor (flat open): 0.90 Theoretical coverage radius: 27.6 × 0.90 ≈ 24.8 km Theoretical coverage area (ceiling): π × 24.8² ≈ 1,930 km² That ~1,930 km² is the radio-horizon ceiling, not a usable-coverage guarantee. Realistic usable coverage from a single well-sited rural tower is far lower - typically on the order of 300 - 800 km² once real-world fading, foliage, and link-margin requirements are accounted for. Plan around the usable figure, not the ceiling. Example 2 - 10 m Mast in Suburbs A volunteer mounts a repeater on a 10 m mast attached to their house in a typical American suburb (one-storey houses, trees). Mobile clients at 1.5 m. Radio horizon (mast): 4.12 × √10 = 4.12 × 3.16 = 13.0 km Radio horizon (client): 4.12 × √1.5 = 4.12 × 1.22 = 5.0 km Total radio horizon: 13.0 + 5.0 = 18.0 km Obstruction factor (suburban): 0.55 Theoretical coverage radius: 18.0 × 0.55 ≈ 9.9 km Theoretical coverage area (ceiling): π × 9.9² ≈ 308 km² This is a solid community repeater covering roughly the footprint of a small city, though the 308 km² figure is again a ceiling - expect meaningfully less usable coverage. Pockets of shadow behind larger buildings will exist. A wardriving survey is recommended after installation to confirm actual coverage. Example 3 - Rooftop in Dense Urban A repeater is placed on the flat roof of a 7-storey (21 m) apartment building in a dense city. Surrounding buildings average 6 storeys (18 m). Mobile clients at street level (1.5 m). Radio horizon (roof): 4.12 × √21 = 4.12 × 4.58 = 18.9 km Radio horizon (client): 4.12 × √1.5 = 4.12 × 1.22 = 5.0 km Total radio horizon: 18.9 + 5.0 = 23.9 km Obstruction factor (dense urban): 0.25 Theoretical coverage radius: 23.9 × 0.25 ≈ 6.0 km Theoretical coverage area (ceiling): π × 6.0² ≈ 113 km² Only 25% of the theoretical range is realised because the repeater barely clears the surrounding roofline. Raising the antenna by 3 additional storeys (to 30 m) lifts the horizon to ~27.6 km and, at an improved ~0.35 obstruction factor, the ceiling rises to a radius of ~9.7 km - roughly π × 9.7² ≈ 295 km² (still a theoretical maximum, with usable coverage well below that). In dense urban networks, rooftop height relative to surroundings matters enormously. Coverage Overlap: The 20 - 30% Rule When planning adjacent repeaters, their coverage footprints should overlap by 20 - 30% of the coverage radius. This is a rule of thumb borrowed from cellular and WLAN cell-planning practice (which commonly calls for 15 - 30% overlap). It ensures: No gap corridor between repeaters where nodes lose connectivity Sufficient signal margin at the cell edge for reliable forwarding (not just barely detectable signals) Redundancy: a node in the overlap zone can reach two repeaters If repeater A has a coverage radius of 8 km and repeater B also has 8 km, place the two sites about 12 - 13 km apart to keep ~20 - 30% overlap (a single shared lens-shaped zone roughly 3 - 4 km wide in the middle - not "on each side"). At exactly 16 km the two cells are tangent: they just touch, with zero overlap margin and zero gap. Only beyond ~16 - 18 km does a real gap corridor open up where nodes may lose both repeaters. Link Budget Margin: Target >10 dB Coverage radius estimates are not hard boundaries - they define the distance at which the link margin drops to zero. In practice, a link operating at exactly the sensitivity floor is unreliable. Fading, multipath, vegetation sway, and atmospheric changes will cause it to fail intermittently. Target at least 10 dB of margin at the cell edge for reliable operation. This means planning for a coverage radius at which the received signal is 10 dB above the receiver sensitivity floor. The 10 dB figure is a standard engineering convention (see any link-budget reference); the margin you actually need scales with the environment - budget more in heavy-fading or foliage-dense paths. For Meshtastic on SF12 / BW125 (the Long Slow preset), receiver sensitivity is approximately −137 dBm with the SX1262's Rx Boosted gain enabled (standard gain is ~3 - 4 dB worse). Note that LongFast (SF11 / BW250) has a sensitivity closer to −131 dBm, not −137. A link budget target of −127 dBm at the coverage boundary gives 10 dB of margin relative to the SF12/BW125 floor. Links measured below −127 dBm at normal operating distances should be treated as marginal and either reinforced with a relay node or deprioritised until a better repeater site is available. Quick field check: If a node reports RSSI < −125 dBm or SNR < −10 dB when communicating with its nearest repeater, that link is at or below the 10 dB margin boundary. Plan to add a relay or move the repeater closer. The Repeater Grid Approach for Urban Coverage Why a Grid Approach? Ad-hoc repeater placement - putting nodes wherever a willing host can be found - produces uneven coverage with clusters of overlapping repeaters in some areas and dead zones in others. A systematic grid approach starts from coverage requirements and works backward to site requirements, ensuring uniform coverage and efficient use of channel capacity. Target Repeater Density by Area Type The figures below are planning heuristics, not guarantees: actual spacing and density depend heavily on antenna height and local building/terrain density, so treat them as rough starting points and confirm with a real coverage survey. Area Type Target Repeater Spacing Approx. Density Rationale Urban core (high-rise, dense) 1.0 - 2.0 km 1 per 1 - 3 km² Heavy building obstruction limits each repeater to a small footprint Suburban (low-rise, residential) 3.0 - 5.0 km 1 per 7 - 20 km² Moderate obstruction; rooftop repeaters can reach 3 - 8 km reliably Exurban / light industrial 5.0 - 8.0 km 1 per 20 - 50 km² Low building density; tall structures (grain elevators, water towers) excellent Rural (farmland, grassland) 8.0 - 15.0 km 1 per 50 - 175 km² Near line-of-sight; hilltop or tower sites dominate Wilderness / remote 15.0 - 30.0 km+ 1 per 175 - 700 km² Solar-powered mountain-top repeaters; realistic in open/semi-open terrain only These spacings are heuristics that assume repeaters are elevated (10 - 30 m AGL for urban, 30+ m for rural) and achieve 10 dB of link margin at the target spacing (cross-reference your own link-budget margin calculations). For low ground-mounted installations, spacing scales roughly with the √h radio-horizon relationship rather than linearly; as a rough rule of thumb, reduce spacing by 30 - 40% (or recompute from the horizon scaling for your specific heights). Step 1 - Identify Anchor Sites Anchor sites are high, prominent structures that provide disproportionately large coverage footprints. Finding and securing anchor sites is the most important work in urban coverage planning. Common anchor site types: Water towers: Heights vary widely (some standpipes are under 20 m, some elevated tanks exceed 50 m; a common range is roughly 25 - 50 m AGL — consult AWWA tank standards if you need precise figures). Often publicly owned, with existing antenna mounts. Utility companies are sometimes receptive to emergency communications partnerships. Ideal anchor nodes. Tall commercial buildings (10+ storeys): Roof access is harder to obtain but a single mid-rise rooftop repeater in a dense urban core can match 5 - 10 lower installations. Target telecommunications companies, building management firms, or property owners sympathetic to community projects. Hilltops and ridgelines: In cities built on rolling terrain (Pittsburgh, San Francisco, Seattle), hilltops within or adjacent to urban areas are the highest-value sites. Even a 5 - 10 m height advantage over surrounding terrain dramatically extends range. Radio / TV transmission towers: Licensed broadcasters and tower companies sometimes offer shared mounting space. The collocation fee may be justified by the coverage gained. Church steeples and clock towers: Often the tallest structures in older residential neighborhoods. Many faith communities are receptive to community emergency communications projects. Map all potential anchor sites in a GIS tool. Run viewshed analyses from each. Rank by coverage area. Secure the top 3 - 5 sites before planning secondary fill repeaters. Step 2 - Fill Planning Once anchor sites are installed and their actual coverage verified (wardriving survey or community signal reports), overlay the confirmed coverage zones on your planning map. Grid cells with no coverage, or with RSSI below −120 dBm, are fill targets. Fill repeaters do not need to be as elevated as anchor sites - they only need to bridge a specific gap. A 5 m residential rooftop install may be sufficient to cover a neighborhood dead zone if it is positioned on the line of sight between two anchor sites. Prioritise fill sites that: Cover the largest dead zone area with the smallest number of new nodes Can hear at least two anchor-tier repeaters (for redundancy) Are accessible for maintenance Step 3 - Documenting Coverage Gaps Maintain a living coverage map updated after every wardriving run or community signal report. A shared GIS layer (e.g., a Google MyMaps or an QGIS project shared via cloud storage) accessible to all network administrators is the best tool for this. For each coverage gap, record: Geographic boundary of the gap (polygon on the map) Estimated area (km²) and population affected Identified candidate fill site(s), if any Status: open / candidate identified / site secured / fill node deployed Last confirmed date of the gap (wardriving date or community report date) A structured gap log prevents the common failure mode of deploying fill nodes in areas that are already well-covered while neglecting persistent dead zones that lack a vocal advocate. Worked Example: Planning a Mid-Sized City Grid Target: A city of 80,000 people covering approximately 120 km², mostly suburban with a 2 km² dense downtown core. Downtown core (2 km²): Target 1 - 2 km spacing → need 1 - 2 anchor repeaters. Identified a 12-storey bank building and a water tower 1.3 km apart. Both anchor sites secured. Downtown coverage achieved with 2 nodes (2 nodes / 2 km² = 1 per 1 km², within the 1 - 3 km² urban-core target). Suburban ring (118 km²): Target 4 km spacing. Hexagonal packing at 4 km site spacing (each site covering a hex of roughly 2 km radius, ≈ 2.6 × 2² ≈ 10 km² per site) gives 118 / 10 ≈ 11 - 12 sites, not 7 - 8 — a triangular/hex grid covers less area per site than a naive square estimate. Identified 9 water towers and 3 church steeples evenly distributed across the suburban ring. All 12 sites secured. Average spacing: ~3.9 km (12 nodes / 118 km² = 1 per ~9.8 km², within the 7 - 20 km² suburban target). Total anchor infrastructure: 14 nodes for 120 km². Report density per zone rather than as a blended average: downtown 1 per 1 km² and suburban 1 per ~9.8 km², both meeting their zone targets. (A single blended "1 per ~8.6 km²" figure would conceal that each zone is independently compliant.) Post-wardriving fill: Survey revealed 3 dead zones in valley neighborhoods. 3 fill repeaters added on rooftops in those valleys. Total: 17 nodes for complete city coverage. Compare this to an ad-hoc approach: a typical volunteer-driven deployment in a city this size might have 40 - 80 ground-level nodes with significant overlap in connected areas and persistent dead zones in underserved neighborhoods. The grid approach delivers better coverage with far fewer nodes and less channel congestion. Designing for Redundancy Why Single-Path Mesh Is Fragile A tree-topology mesh - where each node has exactly one path back to the network core - is the natural shape that forms when coverage is barely adequate. In a tree topology, the failure of any interior node partitions the network. Nodes "below" the failed repeater become an isolated island: they can hear each other but cannot reach the wider network. For a community mesh serving emergency communications, this is unacceptable. The very events that trigger heavy mesh use (storms, earthquakes, infrastructure failures) are also the events most likely to take repeaters offline through power loss, physical damage, or access loss. Per-Zone (+1) Redundancy Note on terminology: in classic systems engineering, N+1 redundancy means N components are needed to carry the load plus one shared spare for the whole system. A mesh network needs something stronger than that — a single shared spare for the entire network does not help if the spare cannot reach the area that lost its repeater. What we want for a mesh is per-coverage-zone redundancy: every coverage area should have a backup path, not just the network as a whole. In practice, this translates to: Every node in the network should be able to reach the network core via at least two independent paths through different physical repeaters. A "network core" is the set of nodes with Internet gateway access or the central coordination point. In a city mesh, the core might be 3 - 5 well-connected anchor repeaters. To verify per-zone redundancy for a given node: Identify all repeaters that node can directly reach (RSSI > −120 dBm). For each of those repeaters, confirm it has at least one other path back to core. If the node can only reach a single repeater, it has no redundancy. If that repeater fails, the node is isolated. Ring Topology vs. Tree Topology Tree Topology Nodes connect to the nearest repeater, which connects to the nearest anchor, which connects to core. This forms a tree. Advantages: simple to plan and understand. Disadvantages: any broken branch isolates all nodes below it. Single points of failure are everywhere. Core ├── Anchor A │ ├── Repeater A1 │ │ └── Client nodes (isolated if A1 fails) │ └── Repeater A2 └── Anchor B └── Repeater B1 └── Client nodes (isolated if B1 or Anchor B fails) Ring Topology Anchor repeaters are interconnected in a ring so that each anchor has two paths back to core. Fill repeaters connect to two anchors where physically possible. This creates a lattice rather than a tree. Core ── Anchor A ── Anchor B ── Anchor C ── Core \ | / Repeater Fill Repeater (hears A (hears (hears B and B) B and C) and C) Ring topology requires more careful planning and more anchor sites (each anchor must be within radio range of two others), but it eliminates the single points of failure that make tree topologies fragile. Recommendation: Design anchor-tier repeaters in a ring or lattice. Fill-tier repeaters can remain in a simplified tree to anchor, but each fill node should reach at least two anchors where terrain permits. Identifying Single Points of Failure with Path Analysis A single point of failure (SPOF) is any node whose failure disconnects part of the network. Identify SPOFs through path analysis: Draw the network graph. Each repeater is a node. Each radio link between repeaters is an edge. Include only reliable links — for example, those with about 15 dB of margin above the receiver's sensitivity. With a typical LongFast/LongSlow sensitivity near −130 dBm, that means including only links measured at roughly −115 dBm or stronger. (Adjust the cutoff to your preset's actual sensitivity floor.) Find the bridge nodes (cut vertices). A cut vertex is any node whose removal disconnects the graph. You can find these visually: any node that is the only link between two clusters is a single point of failure. If you have your topology in software, a graph tool (such as Gephi, or Python's networkx.articulation_points()) will list them automatically — but for most community meshes, eyeballing the map for any node that is the sole bridge between two groups is enough. Prioritise SPOF mitigation. For each SPOF identified, either add a redundant link (find a fill repeater position that bypasses the SPOF) or ensure the SPOF node has UPS backup power, weatherproof housing, and remote monitoring. Testing Redundancy by Taking a Node Offline Theoretical redundancy analysis should be validated with live tests. The procedure is simple: Notify operators. Announce a planned maintenance window (e.g., "Node X will be taken offline for 30 minutes on Saturday 14:00 UTC for redundancy testing"). Take the target node offline by powering it down or disconnecting its antenna. Measure impact. Using a network map (MeshMapper, Meshtastic node list, or MeshCore admin panel), observe which nodes lose connectivity. Nodes that disappear from the map are isolated - this is your actual failure impact, which may differ from the theoretical prediction. Document the partition. Record which nodes were isolated and for how long they would be unreachable in a real failure event. Restore the node and plan remediation for any isolated segments found. Perform redundancy tests at least once per year, and after any significant change to the network topology (adding or removing anchor repeaters, significant coverage expansion). Practical Redundancy Checklist Every anchor repeater can reach at least two other anchors directly. Every fill repeater can reach at least two anchors directly. No anchor repeater is a single point of failure for more than one fill repeater. All anchor repeaters have UPS or generator backup covering at least 72 hours. Network graph has been drawn and cut vertices identified. Redundancy live-test performed in the last 12 months. Failure impact documented: "If node X fails, Y nodes lose connectivity."