Single-mode (OS2) vs Multimode (OM4).
Single-mode OS2 carries far more distance and upgrade headroom than multimode and is the only defensible choice for any link leaving the building; the fibre itself is inexpensive and the optics dominate, which is why it increasingly wins in-building too.
Quick answer — when to choose Single-mode (OS2) or Multimode (OM4)
Quick answer
Single-mode OS2 carries far more distance and upgrade headroom than multimode and is the only defensible choice for any link leaving the building; the fibre itself is inexpensive and the optics dominate, which is why it increasingly wins in-building too. Multimode OM4 remains reasonable for short riser and floor-to-floor runs where the optics are cheaper and the speed roadmap is understood. We do not specify OM3 for new construction — its headroom runs out first. The cable is pulled once and lives for decades, so the decision is about headroom, not this year's optics.
When to use
You are specifying the backbone fibre grade for a new build or a re-cable and want it sized for the building's life rather than the current switch generation.
When not to use
The bottleneck is the horizontal copper, the Wi-Fi design or the switch uplink — the backbone grade will not fix any of those.
Option A
Single-mode (OS2)
A narrow core that carries light a long way with headroom to spare — the only sensible answer between buildings, and increasingly used inside them too.
Choose Single-mode (OS2) when
- The link leaves the building — campus, inter-block, or anything crossing open ground
- The backbone should outlive several generations of switching without being re-pulled
- The site is large enough that one fibre type everywhere simplifies spares and terminations
Option B
Multimode (OM4)
A wider core over short in-building distances, paired with optics that are typically less costly at today's speeds.
Choose Multimode (OM4) when
- Runs are short and stay inside one building — riser and floor-to-floor links
- The active equipment is already standardised on multimode optics across the estate
- The link's speed roadmap is understood and comfortably within the grade's reach
Side by side
The fields that decide it
| Dimension | Single-mode (OS2) | Multimode (OM4) |
|---|---|---|
| Practical reach | Campus distances and beyond | In-building distances; reach shortens as speed rises |
| Upgrade headroom | Carries several switching generations without re-pulling | Reach falls away at each speed step |
| Optics | Historically dearer, and the gap has narrowed | Typically less costly at today's in-building speeds |
| Termination and tolerance | Tighter alignment; cleanliness discipline is unforgiving | More forgiving to terminate and to inspect |
| Inter-building links | The defensible choice, including lightning-exposed routes | Not appropriate once the link leaves the building |
| Spares and standardisation | One grade site-wide simplifies patch, spares and training | A second grade to hold, label and keep straight |
Reach at a given speed depends on the optics, connector count and the installed loss budget — confirm every link against a measured loss budget rather than a headline distance.
Before the drawings freeze
What changes at design stage
What changes at design stage
- Separate the links that leave the building from those that do not; anything crossing open ground is a single-mode decision before anything else is discussed.
- Set the backbone against the building's life rather than the current switch generation — the fibre is pulled once and the optics change several times.
- Decide whether the site is large enough that one grade everywhere is worth more than the per-link optics saving, in spares, labelling and training.
- Establish the loss budget per link at design stage, including connector and splice count, so the grade is chosen against a number rather than a headline distance.
Infrastructure implications
- Riser routes, bend radius and pathway capacity have to be agreed with the architect before the slab closes — fibre is unforgiving about tight pulls.
- Inter-building routes need their civil work planned: duct, draw pits, entry points and the sealing where the route enters each building.
- In a lightning-prone region, an all-dielectric route between buildings avoids a metallic path the electrical design would otherwise have to bond.
- Termination space, patch management and the discipline of keeping connectors clean belong in the rack-room design, not in the commissioning week.
Getting it built
Coordination & integration
Coordination for architects & consultants
- Coordinate the inter-building route with the civil and landscape design early; a fibre route discovered after paving is an expensive conversation.
- Agree with the IT team which optics the estate standardises on, because the grade decision and the spares strategy are the same decision.
- Confirm the labelling and as-built record requirement at design stage so the backbone is maintainable by someone who did not install it.
- Where the route shares containment with power, agree separation with the electrical consultant rather than resolving it on site.
Integration considerations
- The grade decision and the optics decision are one decision; specifying fibre without confirming the transceivers the estate holds creates an unusable link.
- Patch leads must match the grade at both ends — a mismatched jumper is among the most common and least obvious backbone faults.
- Test results and the as-built loss budget belong in the handover pack, because the link that was never certified is the one that fails later.
- Where the backbone carries life-safety or building-system traffic, its resilience and failover behaviour belong in the wider network design.
Typical project fit
Where each is the right answer
Decision summary
- Campus with several blocks and an open-ground route between them
- Single-mode OS2 for every inter-building link, with an all-dielectric route and the civil work planned into the site design.
- Single office building with short riser links and standardised optics
- Multimode OM4 in the riser is reasonable where the speed roadmap is understood and the estate already holds those optics.
- Large site that wants one grade, one spares kit and one training standard
- Single-mode throughout — the per-link optics saving rarely survives the cost of running two grades across a large estate.
From the field
Common mistakes
- Choosing the grade from a headline distance figure rather than a measured loss budget including every connector and splice.
- Running multimode between buildings because a short link looked well inside its reach, then finding the route cannot carry the next speed step without a second civil exercise.
- Specifying OM3 for new construction, which spends its headroom on the first speed step.
- Holding two fibre grades across a large estate without the labelling and spares discipline to keep them straight.
- Pulling a backbone with no spare cores, so the first additional link means a second civil exercise.
TechnoGuru engineering perspective
How we specify it
We treat the backbone as a twenty-five-year decision and the optics as a five-year one. Anything leaving the building is single-mode, planned with its duct, draw pits and entry sealing in the civil design and kept all-dielectric where lightning exposure is real. Inside a single building, multimode in the riser is still reasonable where the speed roadmap is understood, but on any site big enough to need spares and a training standard, one grade throughout usually wins. We do not specify OM3 for new work, we size every link against a measured loss budget, and the certification results go into the handover pack.
Where this connects
Services, sectors, tools & proof
Relevant services
Relevant tools & calculators
Related comparisons
Further reading
Selected proof
Owner-approved, publicly-verifiable project references. No quantities, pricing, layouts or restricted detail are shown.
· How we approach this on a live brief
On a real networking & it brief the choice between Single-mode (OS2) and Multimode (OM4) usually turns on a few of the fields above, not every row in the table. Share the drawings, the BOQ or the constraints and a project lead will name the one we would specify for the building — with the engineering reason set out in writing.
· Frequently asked
Single-mode vs multimode fibre backbone —
what people ask first.
Is single-mode overkill inside one building?
Less often than it used to be. The fibre itself is inexpensive and the optics dominate, so the case for multimode rests on the transceivers the estate already holds. On a site large enough to need spares, labelling and a training standard, running one grade throughout usually costs less in practice than keeping two straight.
Why not specify OM3?
Because its headroom runs out first. OM3 supports the speeds most buildings run today, but its reach falls away at each step up, and a backbone is pulled once and lives for decades. For new construction we specify OM4 in-building where multimode is appropriate, and OS2 wherever the link leaves the building.
What decides the grade on a long in-building run?
A measured loss budget, not a headline distance. Every connector, splice and patch point adds loss, and the reach quoted for a given speed assumes an ideal channel. We build the loss budget per link at design stage, including the connector count the pathway actually forces, and choose the grade against that number.
What do you need to specify a backbone?
The building or campus layout with the routes between blocks, riser positions and rack-room locations, the speed roadmap the estate expects, the optics already standardised on, and any lightning or civil constraint on the external route. From those we advise the grade per link and route the commercial question to a written estimate after drawings.
· Begin
Review a backbone specification
after a look at the drawings.
Share drawings, a BOQ, a project brief or site information and a project lead will return an engineering reading — not a sales pitch — within two working days.