Full-scope ISP network engineering — from field survey through construction-ready deliverables. HLD, LLD, permit drawings, pole loading analysis, and utility coordination delivered by a MBE-certified OSP firm active in 22 states and available to deploy across all 50 U.S. states.
The two engineering failures that kill ISP builds aren't dramatic — they're quiet. Wrong fiber counts on a trunk route. An HLD that assumes splice points where the pole spacing makes that physically impossible in the field. You don't find these problems during design review. You find them when you're 60% through construction and someone notices that the splice enclosure locations don't match the LLD and the fiber count you pulled through a 2,000-foot conduit run isn't enough to serve the east-side service area.
We've seen the fiber count mistake more than we'd like to admit — on projects we didn't engineer. An HLD architect sizes a trunk based on the subscriber density projection at design time. They don't account for the second CO coming online in phase two, or the fact that the splitter architecture they've chosen forces a different split point than the physical route allows. Then the LLD team inherits that HLD and either pushes back — which costs time — or just builds to the drawing. Eighteen months later the ISP's doing a forklift upgrade on a route that's only 40% activated. That's a real scenario. We've cleaned up three of them in the past 14 months.
Field data that doesn't match design is the third failure mode, and it's more common than it should be. An address count from a GIS database that doesn't reflect actual structures. Pole heights estimated from street view instead of measured. Span lengths scaled off a county map instead of GPS-captured. When that data feeds into an LLD, the construction package doesn't work — permits get rejected, make-ready scope is wrong, crews are in the field finding out the hard way that the design doesn't match the ground. We've built our field survey process around eliminating these gaps before they make it into the engineering package.
Five phases, sequenced the way the work actually has to flow — not the way it looks on a project template.
We start with your service territory boundary, address count, and connectivity target. We pull aerial imagery, utility GIS data where available, and county parcel records. We're looking at aerial vs. underground mix, existing plant that might be reusable, and whether the proposed route corridors actually connect the way the initial scoping assumed. On BEAD projects, we also validate location fabric eligibility at this stage — catching disqualified addresses before they end up in the engineering scope. Rough route miles, pole estimates, and a risk flag list come out of this phase before a dollar of detailed design is spent.
High-level design defines everything that matters architecturally: node and hub placement, fiber count decisions on each trunk and distribution segment, GPON split ratios, splice architecture, and the system-wide topology. Don't skip the hard decisions here — fiber counts set in HLD are expensive to change in LLD, and they're nearly impossible to change after construction. We size fiber counts with headroom: the route you're building to 1,200 addresses today needs to handle 1,800 when the adjacent subdivision connects. We document the reasoning, not just the output, so the LLD team isn't guessing at design intent six months later.
LLD converts the HLD architecture into a pole-by-pole, span-by-span construction package. Every attachment point is located and dimensioned. Hardware is specified. Splice enclosure locations are placed at physically achievable points given actual span lengths and pole geometry — not theoretical ones. Strand and cable take-off is calculated from GPS-measured spans, not map estimates. The output is a construction package crews can actually work from: route design, permit drawings, strand and hardware schedules, and splice diagrams. We use AutoCAD and QGIS for the core deliverables, with Katapult data feeding the pole-specific work.
Permits run on their own timeline — one that doesn't care about your construction schedule. We start the permitting process as early as the LLD allows, which on a straightforward aerial route means permit applications go in while LLD drawings are still being finalized for non-permit sections. Municipal ROW, state DOT encroachment permits, railroad crossing applications, and environmental reviews all run as parallel tracks. Utility coordination — joint use applications, pole attachment paperwork, make-ready tracking — runs simultaneously. We've learned which utilities are 6-week reviewers and which ones take 6 months; we sequence accordingly.
Good engineering doesn't end when the construction package leaves our desk. Field conditions don't always match design — a pole gets relocated, a conduit path changes to avoid an unmarked utility, a splice point moves 400 feet because the original location turned out to be in a drainage easement. We provide field support during construction to answer RFIs, review and approve field design changes, and document deviations in real time. As-built documentation is produced from field-verified data, not from the design drawings with "as-built" stamped on them — a distinction that matters when your grant administrator or municipal franchise authority asks for closeout documentation. See our full as-built documentation process.
The scope varies by project phase and build type, but a complete ISP engineering engagement includes these deliverables. We don't abbreviate this list — here's what you're actually getting.
Field data collection and photogrammetry. Our primary platform for aerial pole surveys — GPS-tagged attachment heights, span measurements, and pole inventory that feed directly into O-Calc Pro and AutoCAD workflows.
Mobile field collection for projects where Katapult's full workflow isn't required. Faster setup for smaller pole counts where turnaround speed matters more than full GIS integration.
Open-source GIS platform for route design, service territory mapping, address-level coverage analysis, and BEAD location fabric validation. We use QGIS throughout the HLD and LLD phases.
Construction drawing production — plan and profile sheets, permit drawings, splice diagrams, and pole attachment details. Every permit-ready drawing we produce comes out of AutoCAD.
NESC pole loading analysis for fiber attachment applications. Our standard tool for CLEC and ISP joint use work — integrates with Katapult field data and is accepted by most utility pole owners for attachment applications.
This is the part of ISP network engineering that doesn't show up in vendor pitch decks — and it's where most budget overruns actually originate.
HLD is a planning document. It's supposed to make assumptions — that's the point. You can't fully field-verify a route before you've designed it. But there's a difference between documented assumptions and undocumented ones. When HLD places a hub node at a road intersection because it looks good on a map, without verifying that there's ROW available, conduit capacity, or a power source within 300 feet — that's an undocumented assumption that's going to cost somebody money in LLD. We've seen hub relocations that required rerouting 1.3 miles of distribution fiber because the HLD assumed a node location that turned out to be on private property with no ROW access.
Fiber count decisions made at HLD are the single biggest lever you have on total project cost — and the single biggest source of expensive surprises. A 144-count trunk that should have been 288-count adds cost in the form of a future splice-in and lash. A 288-count trunk where 144 would have been sufficient adds cost now. The right answer depends on: actual address density by route segment, your phased buildout plan, the GPON architecture you've committed to, and whether you're planning any mid-route node connections. HLD has to answer all of these before the fiber count goes on paper. We document every fiber count decision and the reasoning behind it, because those decisions get revisited — and having the reasoning documented saves arguments later.
Splice architecture is where HLD and LLD misalignment shows up most clearly in the field. HLD places splice points at logical network locations. LLD has to place them at actual pole locations with sufficient slack storage, appropriate pull distance from the previous splice, and accessible hardware mounting. When HLD splice locations don't survive the LLD field-verification process, you're either adding splice points — which adds cost and optical loss — or you're accepting a design that doesn't work physically. We run preliminary field feasibility on HLD splice placements before they're finalized, so the locations that go into HLD are ones LLD can actually execute. Read more in our guide on FTTH HLD design mistakes.
The engineering approach changes meaningfully depending on your technology platform — and mismatching the engineering to the technology is an expensive mistake.
GPON (Gigabit Passive Optical Network) is the dominant platform for greenfield FTTH in the U.S. — it's what most ISPs building new fiber infrastructure are deploying. The engineering decisions that matter most in GPON design are: split ratio (1:32 or 1:64, each with different power budget implications), whether you're doing centralized or distributed splitting, and where the first split and second split points land in the physical network. A 6.2 dB optical power budget margin looks fine on paper until someone puts the second split point 8,000 feet from the first split on a 48-fiber distribution cable — then you're doing optical loss calculations at the ONU and hoping the field splices are clean.
EPON is less common for new builds but still present in markets where the ISP is deploying equipment from vendors whose ecosystem favors it. The engineering fundamentals are similar, but the wavelength plan and power budget numbers are different. We've worked with both and design to the actual platform specs, not a generic passive optical template.
Active Ethernet deployments — where every subscriber gets a dedicated fiber all the way to the CO — simplify the splitting architecture but drive up fiber count dramatically. A 500-address deployment on active ethernet needs a fundamentally different route design than the same deployment on 1:32 GPON. The trade-off is real: active ethernet is less engineering-complex per subscriber, but the fiber take-off is much higher. We model both scenarios when clients are still evaluating technology platform and provide the engineering cost and route design implications for each.
ISPs upgrading from hybrid fiber-coaxial (HFC) infrastructure to deeper fiber face a different set of engineering problems — none of which are straightforward. You're not starting from a clean sheet; you're working within existing node service areas, existing conduit routes, and existing ROW relationships. The engineering question is: where does the fiber go, and what happens to the HFC plant that's left? Node+0 deployments push fiber deeper but preserve the coax last mile. FTTH replacements eliminate the HFC plant entirely. Each has a different engineering scope — and a different permitting and utility coordination footprint. We've done both and can scope the engineering for wherever your hybrid upgrade strategy lands.
| Factor | Outsourced (Draftech) | In-House Team |
|---|---|---|
| Speed to market | Engineering starts within 1–2 weeks of contract. No hiring cycle. | 6–12 months to hire, onboard, and get engineers project-ready |
| Cost per route mile | $1,100–$1,800/mile for full scope — no benefits, overhead, or idle-time cost | Fully-loaded engineer at $87/hr + overhead; efficient only at 200+ miles/year continuous volume |
| Scalability | Scale up for BEAD-funded surge, scale down between projects. No fixed headcount. | Fixed team capacity — either understaffed during peaks or paying for idle capacity in troughs |
| BEAD compliance | Familiar with NTIA requirements, location fabric validation, and grant closeout documentation | Depends entirely on whether you can hire engineers who've done BEAD work before |
| Geographic reach | Active in 22 states, deployable across all 50 U.S. states. Multi-state builds are straightforward. | Effective only in markets your team knows. Multi-state projects require significant additional research. |
Here's what you're buying when you engage a full-scope ISP network engineering firm — and roughly what each element costs on a 1,000-address aerial FTTH project.
| Deliverable | Typical Scope | Cost Range |
|---|---|---|
| Field survey & data collection | GPS walk, pole inventory, attachment heights, span measurements | $12,000–$22,000 (1,000 addresses, aerial) |
| HLD network architecture | Topology, fiber counts, node/hub placement, split architecture | $8,000–$16,000 |
| LLD & construction drawings | Pole-by-pole design, splice diagrams, hardware schedules | $18,000–$38,000 |
| Permit drawings & submissions | Municipality, DOT, railroad — varies by jurisdiction count | $6,000–$20,000 |
| Pole loading analysis | O-Calc Pro analysis per pole, make-ready design for failures | $85–$120/pole for analysis; make-ready design extra |
| Utility coordination | Joint use applications, make-ready tracking, multi-owner coordination | $4,000–$12,000 depending on utility owner count |
| As-built documentation | Field-verified as-builts, BEAD closeout package if applicable | $8,000–$18,000 |
Total engineering cost for a 1,000-address aerial FTTH build — full scope, survey through as-builts — typically runs $1,340–$1,800 per route mile. Underground projects run higher due to permitting complexity. The numbers above are benchmarks; we quote per-project based on actual scope and terrain.
FREE FIRST PROJECT
Active in 22 states. First 20,000 LF project free — no commitment. Aerial or underground, anywhere in our active service territory.
Claim Free Design →ISP network engineering covers the full pre-construction and construction-support scope: field survey, route design, permit drawings, pole loading analysis, utility coordination, and construction-ready deliverables. For FTTH builds, that also means GPON architecture, splitter placement, fiber count decisions, splice point design, and as-built documentation after construction. The specific scope depends on whether you're starting from a clean-sheet HLD or picking up an existing design partway through — we can take handoffs at any phase.
A complete ISP network engineering package — from discovery kickoff to construction-ready deliverables — typically runs 10 to 22 weeks. Field survey is usually 2–4 weeks for a 1,000-address deployment. HLD is another 2–3 weeks. LLD and construction drawings are 4–8 weeks. Permitting often sets the critical path — a simple municipal ROW permit might clear in 3 weeks; a state DOT or railroad crossing can take 6 months. We start permitting as early as the LLD allows and run parallel tracks so you're not waiting on permits after engineering is complete.
HLD defines the network architecture: node placement, fiber counts, GPON split ratios, hub locations, and route corridors. It's a strategic planning document — it tells you what the network will look like and roughly where it'll go. LLD converts that architecture into a pole-by-pole, span-by-span construction package: exact attachment points, hardware specifications, strand lengths, splice locations, and permit drawings. The problem is that most ISP projects treat these as purely sequential phases — LLD often reveals that HLD assumptions don't hold in the field, which means expensive rework. We run preliminary field feasibility checks during HLD to catch those conflicts early. For a deeper explanation, read our guide on FTTH HLD design mistakes.
Full-scope ISP network engineering — field survey through construction-ready deliverables — typically runs $1,100–$1,800 per route mile for aerial FTTH. Underground routes run higher due to permit complexity and plan preparation. Pole loading analysis adds $85–$120 per pole. Rural BEAD projects on long, lightly-loaded routes come in at the lower end of the range; dense suburban builds with multiple utility owners and complex permitting run toward the top. We quote per-project based on actual scope and terrain — contact us with your route miles and address count for a project-specific estimate.
Yes. We've engineered BEAD-eligible deployments across multiple states and understand the documentation requirements — location fabric coverage mapping, cost reasonableness thresholds, and the BEAD engineering requirements for construction-ready deliverables that satisfy grant closeout. BEAD adds administrative overhead: location validation, challenge response documentation, and fiber as-built documentation tied to specific address-level coverage. We build those requirements into the scope from the start so grant administrators aren't chasing documentation after construction ends.
Yes. Field conditions don't always match design — a pole gets relocated, a conduit path shifts to avoid an unmarked utility, a splice enclosure location moves because the original spot turned out to be in a drainage easement. We provide construction-phase support to answer RFIs, review and approve field design changes, and document deviations in real time. As-built documentation is produced from field-verified data, not from the construction drawings with "as-built" stamped on top. That distinction matters for grant closeout and for the next engineer who touches the network.
Draftech is active in 22 states and available to deploy across all 50 U.S. states for ISP network engineering. Our highest-volume markets include Florida, Texas, Georgia, North Carolina, Ohio, Virginia, Pennsylvania, and California. We have established field crews, permit relationships, and utility contacts across these markets — which compresses startup time on new projects significantly. We've covered 44,000+ miles of OSP design and passed 2.6M+ addresses across our active project history.
Multi-owner territories are common — a 15-mile aerial route might cross three different electric utilities, a telephone company's plant, and two co-op service areas. Each has its own joint use process, its own attachment application forms, its own make-ready timelines, and its own quirks. We inventory the pole ownership by section at the start of the project, sequence the attachment applications so they go to each owner in priority order, and track the status of each independently. We don't let one slow utility hold up the sections where the other owners have already cleared. It's a coordination problem, and we've had 187 days to learn how to manage it on large multi-owner builds.
ARE YOU AN ISP NETWORK ENGINEERING FIRM?
This page describes the service we deliver to clients. If you provide ISP network engineering — HLD, LLD, OSP design, or field survey — and you're looking for a consistent subcontract pipeline, we have ongoing capacity needs across 22 active states.
Tell us your route miles, address count, and build type — aerial, underground, or mixed. We'll scope the project, flag the risk items, and give you a realistic timeline. And if you're not sure where to start, our first 20,000 LF is free — no commitment, no sales process before we deliver the work.
Contact Our Engineering TeamOr email directly: info@draftech.com — we reply within one business day. | Claim your free 20,000 LF design →
SERVICE AREAS
Active in 22 states and deployable across all 50 U.S. states — including our highest-volume BEAD markets:
View all service areas →Draftech International, LLC is an MBE-certified OSP engineering firm providing ISP network engineering services across all 50 U.S. states — from rural co-op BEAD deployments to metro FTTH builds for Tier-1 carriers. 15280 NW 79th CT, Suite 102, Miami Lakes, FL 33016. Contact our engineering team to discuss your project.