The core is a few microns narrower on one of these — and that single dimension is the reason one runs 40 km on a laser and the other tops out around 550 m on a cheap LED.
Order the wrong fiber patch cord and a link either won't come up at all or will train at a fraction of its rated speed — which makes it tempting to treat multi-mode as simply "the cheaper, worse fiber" and single-mode as the safe default everywhere. That gets the economics backwards. The two fiber types exist because of one physical difference — how wide the glass core is — and that width determines how many paths light can take through it. Multi-mode isn't an inferior compromise; it's the cost-optimized choice for exactly the distances most data-center and campus links actually need.
Light travels down a fiber core by total internal reflection, bouncing off the core-cladding boundary at whatever angle it entered. A 9 µm single-mode core is physically too narrow to support more than one stable path, so every photon effectively travels the same distance and arrives together — a sharp pulse, even after tens of kilometers. A 50 µm multi-mode core is wide enough to support many stable paths (modes) simultaneously, each bouncing at a different angle and therefore covering a different total distance to travel the same length of cable. Those modes arrive slightly spread out in time — modal dispersion — and past a certain fiber length, that spreading blurs one bit into the next faster than any electronics can compensate for. That single geometric fact is the entire reason single-mode reaches tens of kilometers on a precisely aligned laser, while multi-mode is capped at a few hundred meters on a comparatively simple LED or VCSEL.
No — multi-mode is cost-optimized for exactly the distance range where its limitation never gets triggered. A wider core is easier and cheaper to manufacture, splice, and align, and it can be driven by an LED or VCSEL instead of a precisely wavelength-controlled laser, which is why multi-mode transceivers cost a fraction of single-mode ones. Inside a data center, on a campus backbone, or between adjacent wiring closets — runs almost always well under 550 m — that cost savings comes with zero practical downside, because the link is nowhere near multi-mode's distance ceiling. Specifying single-mode fiber for those short runs doesn't buy anything; it just pays for laser-grade precision the link will never need. Single-mode earns its higher cost only once the run gets long enough that modal dispersion would actually have mattered — inter-building, inter-site, or carrier-distance links.
Explains why single-mode and multi-mode fiber exist as genuinely different technologies rather than a good/bad pair — a narrow 9 µm single-mode core supports exactly one light path for long, laser-driven reach, while a wider 50/62.5 µm multi-mode core supports several simultaneous paths (modal dispersion) that cap distance but allow cheap LED/VCSEL transceivers — using cross-section and longitudinal diagrams of both.
Multi-mode fiber shows up in shorter, cheaper links and single-mode shows up in longer, pricier ones, so it is an easy (and mostly harmless) shortcut to rank them as "budget" versus "premium." That framing breaks down the moment someone applies it in the wrong direction — over-specifying single-mode for a rack-to-rack data-center link wastes money on laser-grade optics the run will never need, while trying to stretch multi-mode across an inter-building or campus backbone run risks exceeding its dispersion-limited distance and producing an unreliable or non-functional link.
Both fiber types are glass strands surrounded by cladding of a slightly lower refractive index, guiding light down the core by total internal reflection. The difference is core diameter. Single-mode fiber (OS1/OS2) has a core around 9 microns — narrow enough that the fiber's numerical aperture supports only a single stable transmission mode (light path), driven by a precisely wavelength-controlled laser (typically a DFB or EML at 1310/1550 nm).
Multi-mode fiber (OM1–OM5) has a much larger core, 50 or 62.5 microns, wide enough to support many simultaneous modes bouncing at different angles. Because those modes cover different physical path lengths to traverse the same fiber length, they arrive spread out in time — modal dispersion — which is what ultimately limits multi-mode's reach. Multi-mode is typically driven by cheaper LED or VCSEL sources at 850/1300 nm.
Structured cabling standards (TIA-568) specify multi-mode fiber classes (OM3, OM4, OM5) for horizontal and intra-building backbone runs precisely because those distances — typically well under the ~550 m OM4 limit at 10 Gbps — never come close to triggering modal dispersion limits, so there is no reason to pay for single-mode optics. Single-mode becomes the right (and often only viable) choice for inter-building campus backbones, metro links, and any carrier or long-haul connection, where distance alone rules out multi-mode regardless of budget. Getting this choice wrong at design time is expensive to fix later: fiber types are not interchangeable at the transceiver level (mismatched SFPs will not link up reliably, if at all), so a cabling plant built with the wrong fiber type often has to be re-pulled rather than simply re-terminated.
Not reliably. A single-mode laser launched into a multi-mode core will excite far more modes than the link was designed for, causing excess dispersion and unpredictable errors; a multi-mode LED/VCSEL launched into single-mode fiber under-fills the tiny core and produces a very weak, unreliable signal. Fiber type and transceiver type need to match.
It is an industry jacket-color convention (not a physical requirement) so installers and technicians can visually distinguish fiber types at a glance during installation and troubleshooting — yellow for single-mode (OS1/OS2), orange for older multi-mode (OM1/OM2), aqua for laser-optimized multi-mode (OM3/OM4), and lime green for OM5.
Yes, but at shorter distances than single-mode and usually using parallel-optics transceivers (like SR4) that spread the signal across multiple fiber pairs inside one MPO cable rather than a single pair, specifically to work around modal dispersion at higher speeds.
Not necessarily "safer" — it is strictly more capable on distance, but the optics cost more, connector alignment tolerances are tighter (a poor termination hurts a single-mode link more than a multi-mode one), and for genuinely short runs there is no performance benefit at all, only added cost.
They are laser-optimized multi-mode grades with progressively better bandwidth-distance products: OM3 typically supports 10G to about 300 m, OM4 extends that to roughly 550 m (and shorter reaches at 40/100G), and OM5 (wideband multi-mode) adds support for short-wavelength division multiplexing to carry more channels over fewer fiber pairs.
Try our Enterprise IT Networks Studio
More calculators, simulators, and guides for this discipline.