Ubiquiti Wireless Backhaul (Point-to-Point)

This is my field guide for deploying a robust point-to-point (PtP) wireless backhaul. It covers link types, when to use each mode, channelization, alignment, and post-install optimization. The goal is a stable, low-latency link that survives real-world noise.

  • PtP Backhaul
  • WDS Bridge
  • Channel Planning
  • Alignment & Fresnel

Overview

A wireless backhaul is a Layer-2 bridge between two locations. Unlike client Wi-Fi, backhaul radios are dedicated to passing network traffic between sites, so design decisions (mode, channel, antenna, alignment) matter a lot more. The notes below are vendor-agnostic and apply well to Ubiquiti, MikroTik, EnGenius, etc.

Backhaul Link Types — When to Use What

  • WDS Bridge (PtP): Purpose-built backhaul. Radios are paired by MAC address and forward Layer-2 transparently. Use this for production PtP links; it dedicates the radio to the backhaul and avoids client traffic mixing.
  • Client Bridge: One radio acts like a client that associates to an AP. Works, but less deterministic for backhaul — better for quick temporary links, not primary transport.
  • Repeater/Extender: Same radio both talks to clients and repeats upstream — halves effective throughput and adds latency. Avoid for backhaul unless there is no alternative.
  • Mesh: Flexible multi-hop fabric. Great for resiliency, but shared spectrum and hops reduce capacity. For a single high-quality PtP, a WDS Bridge is still the clean choice.

Rule of thumb: prefer WDS Bridge for PtP backhaul so the 5 GHz radio is fully dedicated to the link.

Recommended Topology (PtP)

Site A (Root)                    Site B (Remote)
┌──────────────┐             ┌──────────────┐
│  Router/SW   │             │  Router/SW   │
└───┬─────┬────┘             └───┬─────┬────┘
    │ LAN │                         │ LAN │
    │     └───[PoE]──Radio A◄──────►Radio B──[PoE]───
    │                          5 GHz, WDS Bridge
    └─────────────────────────────────────────────── L2

Keep the backhaul radio’s channel and power independent from any local access APs.

Planning & Site Survey

  • Line of Sight (LoS): Ensure clear LoS between sites; avoid trees, buildings, and terrain ridges.
  • Fresnel Zone Clearance: Keep at least the first Fresnel zone mostly clear to prevent phase-cancellation. Raise mounts if needed and pick paths that avoid mid-link obstructions.
  • Band: Use 5 GHz for backhaul where allowed; it’s typically less congested than 2.4 GHz.
  • Mounting: Rigid mounts, proper grounding, protected cabling, and weatherproofing are non-negotiable.

Channelization & Power

  • Static Channels on PtP: Fix the backhaul on a clean, DFS-safe channel if possible; avoid sharing channels with nearby APs.
  • Channel Width: Start at 40 MHz. If the noise floor is clean and SNR is great, try 80 MHz. If interference appears, drop to 20/40 MHz for stability.
  • TX Power: Use the minimum power that achieves target SNR. Excess power increases noise and distortion.
  • Auto-Channel for Access APs: Let local client APs pick channels automatically, but keep the backhaul on its own static plan.

Physical Alignment

Perfect alignment is worth the time. Use the vendor’s alignment tools with live RSSI/SNR graphs at both ends. Make small, alternating adjustments left/right and up/down, then tighten the mounts. Re-check after thermal cycles or strong winds.

Configuration Checklist (Root & Remote)

  1. Set both radios to WDS Bridge (PtP) and pair via MAC address.
  2. Lock a static 5 GHz channel and channel width; match on both ends.
  3. Use strong WPA2/WPA3 pre-shared keys (or vendor’s secure backhaul mode).
  4. Disable services you don’t need (management SSIDs, client access) on the backhaul radios.
  5. Give radios static management IPs on your infrastructure VLAN.
  6. Enable watchdog/ping-reboot only if you’ve validated it won’t flap during maintenance.

Optimization After Turn-Up

  • Record baseline RSSI/SNR, MCS rates, throughput.
  • Perform a manual channel scan during peak hours and adjust if needed.
  • Verify latency/jitter over time; aim for stable single-digit ms on PtP.
  • Re-run an optimizer or channel planner whenever you add nearby links.

Troubleshooting & Hardening

What “Good” Looks Like

Feedback Welcome

Have tips to improve link budget, channel plans, or alignment tricks? Leave feedback. I’m iterating this playbook based on real deployments.