FTTH

Fiber to the home

The FTTH skill turns Nexma into a fiber-network design platform. It models the full passive optical network from the central office to the home, validates optical budgets across cascaded splitters, and enforces the color-code and bend-radius rules a production planner lives by. FTTH was Nexma's first vertical and remains the most mature skill in the catalog.

What this skill models

The ontology captures a real PON hierarchy with the entity types, link types, and constraints a fiber engineer expects.

Entity types. OLT, Cabinet, Splitter, Closure, Drop, OpticalNetworkUnit, Duct, FiberCable, Splice, and Strand. Each carries typed properties — port count, splitter ratio, fiber count, connector type.
Link types. The hierarchy runs OLT to Cabinet to Splitter to Drop to ONU. Side links capture structure too: a FiberCable contains many Strands, and a Splice joins two Strands.
Key constraints. Splitter ratios are restricted to standard values (1:8, 1:16, 1:32, 1:64); end-to-end optical loss must stay within budget; cable bends must respect a minimum radius; fiber assignments follow TIA-598-D color order; cascade depth is capped.

The ontology also ships the constants a planner would otherwise look up: TIA-598-D color codes for 12-fiber and 24-fiber buffer tubes, standard splitter insertion-loss tables, and dB budgets for common PON architectures.

What Jax can do

Jax reasons in fiber terms because the Skill binds its capabilities to this ontology. You describe intent in plain language; Jax designs and validates against the constraints.

Generate networks. "Plan a feeder and distribution network for every household in this polygon. Reuse existing pole runs. Splitter ratio 1:32." Jax derives demand from OSM building footprints if you have no household layer.
Place cabinets under a street-network constraint and a maximum-feeder-length rule.
Size splitters — the 1:32 versus 1:64 trade-off for a target take rate.
Validate the optical budget across a multi-stage cascade, surfacing any over-budget path.
Refurbish in place — "swap this 1:8 for a 1:16 and re-route" — without breaking a downstream constraint.

Heavy optimization (cabinet siting, feeder layout) runs through the Nexma MathEngine; the result is written straight to the world model, so the map and tabs update the moment Jax finishes.

Tip: Give Jax the take rate and the splitter policy up front. Optical budget is a function of cascade depth and split ratio, so fixing those two first makes every downstream validation deterministic.

Constraints and standards

The skill enforces the standards a fiber design has to satisfy to be buildable.

Concern	Rule enforced	Standard
Split ratios	1:8, 1:16, 1:32, 1:64 only	PON architecture practice
Optical budget	End-to-end loss within dB budget	ITU-T G.984 / G.9807
Fiber color order	Strand assignment follows buffer-tube order	TIA-598-D
Bend radius	Cables respect minimum bend radius	Manufacturer spec
Cascade depth	Capped per architecture	Loss-budget derived

Export targets include GeoJSON, KML, and vendor formats such as NetAdmin, VETRO, and IQGeo from Project to Export.

Out of scope today: detailed financial modeling beyond high-level cost estimation, as-built reconciliation against legacy GIS (roadmap), and active-equipment provisioning (handled by NMS systems). The ontology is open — most production deployments extend it with their own constraints and vendor catalogs.

Where to go next

FTTH network tutorial — design a network end to end.
Skills overview — the full catalog.
The ontology — how entity types and constraints are defined.
Nexma MathEngine — the optimization layer behind cabinet siting.