High-Level Design (HLD)

High-Level Design (HLD) sets the architecture of a fibre network before detailed routing begins. Network type, splitter ratios, hub and exchange locations, broad coverage areas — every input that constrains the LLD phase. A well-built HLD turns a project into a defensible plan; a thin one stores up rework.

We deliver HLDs for FTTH and FTTP rollouts across UK and US markets, working from buildable assumptions rather than glossy concept work.

What HLD actually decides

Six decisions get locked in HLD that LLD then implements:

1. Network architecture. PON vs active Ethernet vs hybrid. Each has cost and operational implications that ripple through the project lifecycle.

2. PON technology choice (where applicable). GPON, XGS-PON, or XGS-capable infrastructure designed for upgrade. The choice trades upfront cost against future-proofing.

3. Split ratio. 1:32 is the default for residential FTTH in most geographies; 1:64 increases premise density per fibre at the cost of optical budget. The right answer depends on network depth and the upgrade horizon.

4. Hub and aggregation locations. Where the OLT sits, where the splitters cabinet locates, where mid-span splice closures land. Drives both build cost and operational maintenance.

5. Coverage area boundaries. What’s in and out of phase 1, what’s deferred, what edge premises will require disproportionate cost to reach.

6. Backbone routing logic. Whether the backbone follows existing duct (PIA in the UK), aerial paths, new ducts, or a mix. Drives permitting and construction timeline.

What’s in an HLD deliverable

Across UK and US projects, an HLD pack typically contains:

Topology diagrams — schematic and geographic view of the network architecture
Splitter architecture — split ratios, splitter cabinet locations, fibre allocation logic
Coverage maps — premises in scope, density analysis, build phasing
Hub locations — OLT and aggregation point siting with rationale
Backbone route options — high-level route choices with comparative trade-offs
Indicative BOM — at the spec level (cable types, equipment classes), not part numbers
Preliminary cost model — capex per premise across the chosen topology
Risk register and assumptions log — what’s assumed, what needs confirming in LLD

Standards and operator alignment

Our HLDs are produced to the standards of the operator or funder the design will run under:

UK: BT, Openreach, Virgin Media, MBNL standards as required
US: RUS loan design standards for rural fibre, BEAD compliance for federally-funded projects, individual ILEC and altnet standards
Australia: NBN-aligned designs where applicable
EU: market-by-market — French (Arcep), German (BNetzA), Nordic standards

For any project taking RUS or BEAD funding, the HLD output is structured to feed straight into the funder approval workflow without rework.

Inputs we need to start

A productive HLD engagement starts with:

Coverage area definition — boundary, premise count, optionally premise mix
Commercial parameters — target take-rate, build budget envelope, planned launch date
Topology preferences — if the operator has standardised on a specific architecture
Existing infrastructure — what’s reusable (existing fibre, ducts, hubs)
Regulatory context — funding source, operator standards, permit jurisdictions

Where some of these are still being decided, the HLD becomes a shaping exercise rather than a fixed-scope deliverable — we run it as an iterative engagement rather than a one-shot design.

Output formats

HLDs ship as:

PDF report with maps, diagrams, BOM, cost model
GIS files (QGIS, ArcGIS, or GeoJSON) for the spatial elements
CAD output where the downstream team requires DWG (typically for permit-authority pre-submissions)
Structured data (CSV, Excel) for cost model and BOM where that’s the client’s preferred format

Common pitfalls in outsourced HLD work

HLD that’s actually a desktop study. A “Complete Guide to Fiber Optic Network Design” is not an HLD. An HLD has to make decisions, not summarise them. We commit to topology, split ratio, hub locations — with rationale — rather than presenting options in perpetuity.

Topology that doesn’t survive LLD. An HLD that doesn’t account for real geographic constraints (ROW, terrain, easement availability) produces splitter cabinet locations that can’t actually be built. We sanity-check at HLD against likely LLD constraints rather than discovering issues months later.

Cost models without sensitivity. A point-estimate cost model is reassuring and wrong. We produce ranges with explicit drivers — make-ready required vs not, ROW cost variability, take-rate impact on backhaul sizing.

Typical timelines

Small coverage area (under 5,000 premises) — 3-4 weeks
Mid-sized rollout (5,000-25,000 premises) — 4-8 weeks
County-scale or large altnet (25,000+ premises) — 8-16 weeks, typically phased

These assume access to coverage data and operator standards. Where these need to be assembled or clarified, add 2-3 weeks at the front.

How we deliver

HLD work runs onshore-led with offshore production support. The architectural decisions — topology, split ratio, hub siting — sit with senior engineers who’ve designed FTTH networks at scale. Production output (drawings, maps, BOM build-out) runs through our offshore team, with onshore review before any deliverable goes back to the client.

Talk to us about an HLD

Tell us the coverage, the timeline, and any operator or funding constraints. We’ll come back with scope, price range, and approach. For projects under 10,000 premises, we usually scope within two business days; for larger programmes, a scoping call works better.