Fiber Optic Cabling Toronto & GTA
OS2 single-mode and OM4 multimode backbone for risers, campuses and data centres. Designed around strand count, pathway and the equipment you will buy in five years — not just the switch you are plugging in this month.
Velocity Cabling designs and installs commercial fiber optic infrastructure across Toronto and the GTA — backbone between floors, links between buildings, campus routes and data centre trunks. This page covers the decisions made before anyone pulls cable: whether fiber is the right answer, which grade, how many strands, which rating and how it will be extended later. The termination and certification side is covered on fiber splicing and termination.
When Fiber, And When Not
Fiber is not automatically better. It is better for four specific reasons, and if none of them apply, copper is cheaper and simpler.
| Trigger | Why copper fails | Typical case |
|---|---|---|
| Distance | Copper stops at 100 m for the channel | Deep warehouses, long floor plates, MDF to remote IDF |
| Between buildings | Copper creates an electrical path between structures | Campus links, outbuildings, gatehouses, yard offices |
| Electrical noise | Copper picks up interference near drives and motors | Plant floors, near VFDs, industrial process areas |
| Bandwidth headroom | Copper category caps what the cable can ever carry | Backbone links expected to outlive several switch refreshes |
The building-to-building point deserves more attention than it gets. Fiber is dielectric — it contains no metallic conductor, so it carries no electrical path between structures. Copper run between two buildings ties their grounding systems together, and any difference in ground potential, or a nearby lightning strike, has a route to travel. Fiber removes that path entirely. Even at short distances, that is usually reason enough on its own.
Single-Mode vs Multimode: The Honest Economics
The usual framing — "single-mode goes further, multimode is cheaper" — has not been accurate for years, and it leads people to the wrong answer.
The cable itself tells a different story. OS2 single-mode is frequently cheaper per metre than OM4, because it is simpler to manufacture and produced in enormous volume. Where multimode still wins is the optics: multimode short-reach transceivers remain cheaper than their single-mode equivalents, and across a data centre with hundreds of ports that difference compounds.
| OS2 single-mode | OM4 multimode | |
|---|---|---|
| Core size | 9 µm | 50 µm |
| Jacket colour | Yellow | Violet or aqua |
| 10GbE reach | 10 km and beyond | 400 m |
| 100G reach | Kilometres | Roughly 100 m, depending on standard |
| Cable cost per metre | Often lower | Often higher |
| Optics cost per port | Higher | Lower |
| Bandwidth ceiling | Effectively none at building scale | Distance-limited by modal dispersion |
| Best for | Backbone, campus, anything leaving a building, long-life links | Data centre and in-building links, many short high-speed ports |
The practical rule we work to: OS2 for anything that leaves a floor or a building, OM4 inside data centres and equipment rooms where link counts are high and distances are short. On mixed projects we install both, which costs less than it sounds because the pull is the same either way.
One thing to avoid: do not specify OM1 or OM2 for anything new. If you have 62.5 µm OM1 in a building already, treat it as legacy and plan its replacement rather than extending it.
Strand Count Is The Decision People Regret
Of everything on this page, this is the one that costs clients the most money, and it is entirely avoidable.
The glass is a small fraction of an installed fiber run. The cost sits in labour, pathway, access and disruption — and every one of those is identical whether the cable you pull contains 12 strands or 48. Pulling a second cable two years later means paying all of it again, in an occupied building, probably at night.
A single 10G link uses two strands. It is easy to convince yourself four is plenty.
Redundant paths, a second switch stack, a separate camera network, a tenant, a lab. Each wants strands.
Strands do get damaged during a building's life. Spares turn an outage into a patch change.
12 strands where 4 are needed. 24 or 48 on a main backbone is normal, not extravagant.
Unused strands only help if the next person knows they exist and which port they land on.
Underground duct and inter-building routes. Re-pulling those is the most expensive work there is.
Not sure how many strands to specify?
That is a conversation worth having before the quote, not after. We size it against what the building is likely to need, not just what is being connected this month.
Backbone Architecture
TIA-568 describes a hierarchical star: a main cross-connect in the MDF, with backbone runs out to each telecom room or IDF, and horizontal copper from there to the work area. It is simple, it is what equipment expects, and it keeps fault domains small.
Two design decisions matter more than the topology diagram.
Where the IDFs go
IDF placement is really a copper decision made with fiber. Every IDF must sit within 100 metres of the furthest outlet it serves, measured along the actual cable route. On a wide floor plate that dictates how many rooms you need and therefore how many backbone runs. Get it wrong and you either exceed copper distance or build a room you did not need.
Whether you need a second path
A single backbone run is a single point of failure for an entire floor or building. Where uptime justifies it, a second route on a physically different path — different riser, different duct — turns a cut cable into a degraded service instead of an outage. The extra cost at install is far smaller than most people assume, because again the glass is not the expensive part.
Cable Ratings And Environments
Fiber cable has its own fire ratings, distinct from the CMR and CMP designations used for copper.
| Rating | Use | Notes |
|---|---|---|
| OFNP | Plenum — air-handling spaces | Required in ceiling voids used as return air. Highest rating, highest cost. |
| OFNR | Riser — vertical between floors | The usual choice for building backbone that is not in a plenum. |
| OFNG / OFN | General purpose, single floor | Not for risers or plenums. |
| Outside plant (OSP) | Buried, duct, aerial, between buildings | Built for moisture, UV and abuse. Generally not rated for extended indoor runs. |
| Indoor/outdoor | Entrance runs | Dual-rated so it can transition into the building without a separate splice point. |
The transition point catches people out. Outside plant cable typically uses a water-blocking gel and a jacket that is not rated for extended runs inside a building. Codes limit how far it can travel indoors before it must transition to listed indoor cable or terminate in an entrance enclosure — commonly cited as around 15 metres, though the specific requirement is governed by the CEC and your authority having jurisdiction. Designing the entrance point properly at survey avoids an expensive conversation at inspection. Indoor/outdoor dual-rated cable sidesteps it entirely on shorter entrances.
Pathway, Bend Radius And Slack
Fiber is not fragile, but it is unforgiving of three specific things, and all three are installation practice rather than product choice.
Pull tension
Every cable has a maximum rated pulling tension. Exceed it and you can stress the glass in ways that pass a test on the day and fail two winters later. Long or complex pulls need proper lubricant, intermediate assist points and a swivel, not more people on the rope.
Bend radius
Bends tighter than the rated minimum cause macrobend loss — light escaping the core. Typical guidance is around ten times the cable outside diameter unloaded and twenty times under tension, but follow the manufacturer's figure. Bend-insensitive single-mode to G.657 tolerates much tighter routing and is worth specifying where pathway is cramped.
Service loops
Leave slack at both ends, stored properly to bend radius. A service loop is what lets someone re-terminate after damage, move an enclosure or extend into an adjacent room without re-pulling the run. It costs a few metres of cable and saves entire projects.
Where fiber shares pathway with copper and power, containment is sized for the whole scope — see cable tray installation.
Migrating To 40G, 100G And Beyond
This is the strongest commercial argument for installing fiber properly the first time: the cable in the walls usually does not change when the speed does.
Going from 1G to 10G, or 10G to 40G and 100G, generally means changing the optics at each end. The fibre stays, provided the grade and distance support the target and the link was certified honestly at install. That is why a backbone specified for today's switch and no further is a false economy — and why the strand count and grade decisions above matter so much.
In data centres the parallel-optics question arrives with higher speeds: MPO trunk systems, and whether to standardise on base-8 or base-12 depending on the transceiver types you expect. Getting that wrong means cassettes and trunks that do not map cleanly to the optics. See data centre cabling for that side.
Where We Install Fiber
Riser backbone
MDF to floor IDFs in multi-storey buildings. OFNR or OFNP depending on the shaft, with slack at both ends and spare strands documented.
Campus and building-to-building
OS2 in duct, direct burial or aerial, with the entrance transition handled properly. Dielectric, so no electrical path between structures.
Warehouse and industrial
MDF to remote IDFs in deep buildings where copper cannot reach the far wall, and routes past electrically noisy plant.
Data centre and equipment rooms
OM4 and OS2 links, MPO trunk systems and structured patching built for density and airflow.
Termination, splicing and certification for all of the above is covered on fiber splicing and termination — including emergency repair when a cable is cut.
Fiber Optic Cabling: Common Questions
Four triggers. Distance beyond the 100-metre copper limit. Any link that leaves the building, because fiber is dielectric and carries no electrical path between structures. Electrically noisy environments near drives and large motors. And bandwidth headroom, where the same fibre supports 1G today and 100G later with only the optics changing. If none of those apply, copper is usually the cheaper and simpler answer.
For anything leaving a building, or any link you expect to keep for a decade, OS2 single-mode. Its bandwidth is effectively unlimited over building distances and the cable itself is often cheaper per metre than OM4. Multimode still wins on total cost inside data centres with many short links, because multimode optics are cheaper and that difference multiplies across hundreds of ports. Distance and port count decide it, not which is newer.
More than you need today, by a wide margin. The glass is a small fraction of the installed cost — the expensive parts are the labour, the pathway and the disruption, and those are identical whether the cable holds 12 strands or 48. Every client who under-provisioned regretted it; nobody has ever complained about spare strands. A 12-strand run where 4 are needed is normal practice, and 24 or 48 on a main backbone is not extravagant.
They are fire and environment ratings. OFNR is riser-rated for vertical runs between floors. OFNP is plenum-rated for air-handling spaces such as the void above a suspended ceiling used as return air. Outside plant cable is built for moisture, UV and physical abuse but is generally not rated for extended runs inside a building — codes limit how far it can travel indoors before transitioning to listed indoor cable, commonly cited as around 15 metres. Confirm the specific limit with your authority having jurisdiction.
Very often, and this is the main argument for installing it properly the first time. Upgrading 1G to 10G, or 10G to 40G and 100G, usually means changing the optics at each end rather than the cable in the walls, provided the fibre grade and distance support it. We test existing fiber and tell you honestly what it will carry before anyone budgets a replacement.
Yes. Building-to-building links are one of the strongest arguments for fiber, because it is dielectric — no copper path means no ground potential difference between structures and no surge route during a lightning event. We handle duct, direct burial and aerial routes, plus the entrance transition from armoured outside plant cable to riser-rated interior.
Every strand is certified to Tier 1 for insertion loss and Tier 2 with an OTDR trace, and you receive both along with a strand-by-strand map. The termination and certification side is covered in detail on our fiber splicing and termination page.
Toronto, Mississauga, Brampton, Vaughan, Markham, Scarborough, Etobicoke, Oakville, Hamilton, Kitchener-Waterloo, Whitby and Pickering. Crews dispatch from our Toronto base at 60 Atlantic Avenue.
Last reviewed 12 August 2026 by the Velocity Cabling install team. Cable rating and entrance transition requirements are governed by the CEC and the authority having jurisdiction.
From First Call To Certified Handover.
Five steps, same order, every job. You always know what happens next and who is on site.
Free Onsite Survey
We walk your space, document cable routes and flag every constraint. No cost, no obligation.
Written Proposal
Itemized quote covering materials, labour, timeline and compliance standard.
Pre-Install Planning
We coordinate with your PM, GC or IT team. Rack layouts and phasing locked before day one.
Expert Installation
Licensed technicians pull, dress, terminate and label every run to commercial spec.
Test & Certify
Every run Fluke tested. We don't leave until every port passes.
Toronto, The GTA & Southern Ontario.
Crews dispatch from our Toronto base and reach most GTA sites within an hour.
Fiber Sized For The Building, Not Just The Switch.
Free onsite survey within 48 hours. Strand count, grade, rating and pathway decided before anyone pulls cable.
Mon–Fri 8am–10pm · Evening & weekend installs available · Toronto, GTA & Southern Ontario