Home / Solutions / Industrial Networking & Datacom / ISP / Metro Bearer Network

DOMAIN 02 · INDUSTRIAL NETWORKING

ISP / Metro Bearer Network: One Network That Carries Broadband, Leased Lines and Wholesale Across a City

Written for local internet service providers and network operators in Africa and Latin America — the company with fibre in the ground, subscribers arriving faster than capacity, and a core that has not been redesigned since it was built. A protected metro core ring, dual-homed aggregation, a subscriber edge that handles addresses and policy per user, and every service carried as a VPN on the same bearer so that a new customer is a configuration rather than a project.

Four Things That Happen to Every Growing Operator

The network was built for the business you had. These are the four ways it stops matching the business you have now:

Subscribers grew; capacity did notEvery month brings more homes, and every home now streams. The core carries what it was built to carry, the evenings get slower, and the complaint is not about capacity — it is about you.

An enterprise leased line takes two weeks to turn upThe customer signed on Monday and asks every day after. Meanwhile each new circuit is a hand-built path across several devices, done carefully, by the one person who remembers how the last one was built.

One ring break and the whole city is offlineA road crew, a backhoe, a bridge repair. If the topology has no second direction — or if both directions share the same duct — the outage is not a fault, it is a design decision that was made years ago.

Every new service means replanning the whole networkMobile backhaul, a government contract, wholesale to a smaller operator. Each one arrives as a proposal to build another overlay, because the current network cannot carry a service it was not designed for.

Architecture: A Bearer, Not a Collection of Links

Access collects, aggregation dual-homes, the core ring protects, and every service rides a VPN on top — so growth is a configuration and a cut is a reroute:

Architecture drawn by AtlasCommTech following carrier-grade network-security design practice. Diagram labels are kept in English for engineering clarity.

Who this page is for. Local internet service providers and network operators in Africa and Latin America. Concretely: the operator with fibre already in a district or a city, a few thousand to a few tens of thousands of broadband subscribers, an enterprise leased-line business growing faster than the team that provisions it, and a core that has been extended rather than redesigned since the day it was built. Not the national incumbent with its own architecture department — that company does not need this page. If your subscriber count has doubled while your topology drawing has not changed, this is written for you.

Why us: our founder spent 13 years inside the Huawei partner ecosystem delivering carrier networks. Metro rings, subscriber edges, service provisioning at scale and the reasons rings fail are not a category we entered last year — they are the work the company grew out of.

Equipment Options

The solution is sized to your requirements and budget first — the same architecture can be delivered on several vendors' product lines. We help you choose by supply availability in your destination country, budget and your team's operating habits.

Huawei — enterprise campus, WAN and security linesMature ecosystem with a global service network.

ZTE & Wantone — comparable datacom linesPrice-performance direction; supply runs smoother in some markets.

H3C — campus and data-center linesWidely deployed campus and data-center portfolio.

Atlas industrial switches — industrial-scenario access layerOur own industrial line — compatible with any brand's core layer.

What the Design Delivers

Six properties — and the test of all six is whether next year's business fits on this year's network:

A core ring that survives a cutTraffic reverses direction around the ring fast enough that a voice call stays up. Worth stating plainly: this protects against one cut, and only if the two directions genuinely take different fibre routes.

Converged bearer: every service on one networkBroadband, enterprise leased lines, mobile backhaul and wholesale carried on the same infrastructure and kept apart by VPN and class of service — instead of a new overlay each time you sign a new kind of customer.

A subscriber edge that scales in sessionsAddressing, session handling and per-user policy at the broadband edge, sized by concurrent sessions rather than homes passed — because the two numbers diverge exactly when it matters.

Leased lines turned up in hoursA new circuit is an access port plus a service definition, not a hand-built path across the network. The sales cycle stops being limited by the provisioning cycle.

Class of service that means somethingA business circuit and an evening streaming peak share fibre without sharing fate. The contracts you sign to enterprises become deliverable rather than aspirational.

IPv6 dual-stack and address translation, plannedPublic addresses run out before subscribers do. Carrier-grade translation to keep growing on what you hold, dual-stack so IPv6 is there when asked — designed in, rather than bolted on the week you run out.

Three Sizes, One Design Logic

A bearer network is sized by subscribers, services and the fibre you actually own — here is how the tiers differ:

Numbers we design around:

Concurrent sessions at the subscriber edge — not homes passed, and not homes sold

Busy-hour capacity per ring segment, with headroom for the growth you can defend

Fibre route diversity: two ring directions in one duct is one direction wearing a costume

Scale tier	Typical operator	What the design includes
District / county ISP	One town or district · a few thousand subscribers · first enterprise customers	A small ring or a protected pair between two sites, one subscriber edge sized on concurrent sessions with room to double, enterprise handoff carried as a VPN from day one so the first leased line does not become a precedent for hand-built ones, dual upstream if two are physically available — and if only one is, we say so on the drawing rather than implying resilience you do not have.
City ISP	Metro operator · tens of thousands of subscribers · real enterprise business	Full metro core ring with aggregation sites dual-homed onto it, subscriber edge distributed so traffic does not cross the city to reach the internet, service classes that make an enterprise contract deliverable, transit plus local exchange peering, and address translation with dual-stack planned before the address block runs out.
Multi-city operator	Regional operator · several metros · wholesale and backhaul customers	Interconnected metros with a clear boundary between metro and backbone, service definitions that are identical in every city so provisioning is one skill rather than five, wholesale carried with real separation from your retail traffic, capacity planning tied to the fibre build rather than to hope, and an addressing plan that survives the next acquisition.

Equipment Roles (Categories, Not Models)

The solution is built from these product categories — the brand is chosen with you at design stage. Exact models depend on your subscriber count, capacity per ring segment, service mix and destination country — so we spec models after your requirements list, not before.

Role	What it does
Metro core routers	Carry the ring and the protection behaviour that makes a cut a reroute instead of an outage. Sized on busy-hour capacity per segment with headroom — this is the layer where under-buying is expensive later and over-buying is expensive now.
Aggregation routers / switches	Collect access sites onto the ring, dual-homed so no aggregation site depends on one path. The layer where most real-world resilience is either created or quietly given away.
Subscriber edge (broadband network gateway)	Terminates broadband sessions, hands out addresses, applies per-user policy and does carrier-grade address translation. Sized by concurrent sessions and translation capacity — the two numbers that decide whether growth is a good month or an outage.
Optical access & enterprise handoff	Where homes and business premises meet the network. We supply the network-side equipment categories and are direct about which parts — the outside plant, the civil works — you should be buying locally instead.
Management & monitoring platform	Provisioning, capacity trends and fault localisation. The difference between knowing a ring segment is filling up next quarter and discovering it during an evening peak.
Optics, spares and the boring inventory	Transceivers, patch cords, spare line cards. Unglamorous, routinely forgotten at budget stage, and the reason a two-hour repair sometimes becomes a two-day one. We list them.

Send us your requirements list — subscriber count, busy-hour capacity, service mix, fibre routes you actually own — and the model list follows. That order keeps the design honest.

Design Notes & Honest Limits

Read this before you commit:

The core cost of a bearer network is fibre and facilities, not equipment. Cable, ducts, poles, permits, power and sites dominate the investment; the routers are a modest share of the total. Two conclusions follow, and we will state both even though we sell the routers: do not over-economise on the equipment — the saving is small and the ceiling it creates is not; and do not gold-plate it either — money that buys another kilometre of fibre usually buys more network than money that buys a bigger box.
A ring protects against one cut at a time, and only if the fibre routes are genuinely different. Two directions sharing a duct, a bridge or a trench are one direction wearing a costume, and no protocol will discover that for you at the moment of the cut. We ask for your actual fibre routes before we draw the ring, and if the diversity is not there we write it on the drawing rather than around it.
Availability figures are design targets, not promises we can make on your behalf. What your subscribers experience depends on power, on physical route diversity, on how quickly a team reaches a site at night and on the discipline of your change process. The equipment can support a high target; the outcome is a system property, and most of that system is not something we sell.
The subscriber edge is sized by concurrent sessions and translation capacity, not by homes passed. These numbers separate exactly when growth accelerates — which is the worst possible moment to discover the gap. We size on the session figure with headroom and show you the arithmetic.
Licensing policy, feature availability and product availability differ by brand and by destination country, and for carrier equipment the software licence structure can matter as much as the hardware price. We check and confirm both for your country at the design stage, before you commit.

FAQ

Who is this designed for?

Local internet service providers and network operators in Africa and Latin America — the company that already has fibre in a district or a city, a few thousand to a few tens of thousands of broadband subscribers, a growing enterprise leased-line business, and a network that grew by addition rather than by design. Not tier-one national carriers with their own architecture teams. If you are the operator whose subscriber base has doubled while the core has not changed since it was built, this page is written for you.

Why does a leased line take two weeks to turn up today?

Almost always because each new customer is a network change rather than a configuration. When services are carried by hand-built paths, adding one means touching several devices, checking that nothing else breaks, and finding an engineer who remembers how the last one was done. When services ride VPNs on a bearer that was designed to carry them, adding a customer is a configuration on two devices — the access port and the service definition. Same fibre, same team; the difference is whether the network was designed for the business you actually sell.

Can a ring really protect against a fibre cut?

Against one cut, yes — that is exactly what it is for. Traffic reverses direction around the ring and the interruption is short enough that a voice call survives it. Two important honesty notes. First, a ring protects against one cut at a time; if your two ring directions share a duct or cross the same bridge, one backhoe takes both, and no protocol will help. Second, ring protection covers the fibre, not the power: if the site loses power, the site is down regardless. Ring design is fibre-route design first and configuration second.

Should we spend our budget on higher-end routers or on more fibre?

On fibre and facilities, nearly always — and we will say so even though we sell the routers. In a metro bearer network the dominant costs are cable, ducts, poles, permits, power and the sites themselves. Equipment is a real but modest share of total investment, and the money saved by under-specifying it is small next to what it costs you when the box cannot carry the subscribers you signed. Do not over-economise on the one line item that decides whether the network scales. Do not gold-plate it either — buy for the growth you can defend, on the network you can actually build.

Do we need to worry about IPv6 and address exhaustion?

Yes, and it is cheaper to plan now than to retrofit. Most growing operators reach the point where public addresses run out before subscribers do, and the usual answer is a combination: carrier-grade address translation at the subscriber edge to keep growing on what you have, plus dual-stack so IPv6 is available to customers who ask and to services that increasingly assume it. Address translation costs you session capacity and some logging obligations; we size for both rather than discovering them later.

Send us your subscriber count, busy-hour capacity and the fibre routes you own

An engineer replies with a ring and aggregation outline, a subscriber-edge sizing and the equipment-category list — plus a straight answer on where your money is better spent than on boxes. Send us your requirements list; the model list follows.

WhatsApp an engineer →