The old instinct was simple: lease extra fiber because bandwidth will grow. Today, DWDM changes that calculation. A single duplex fiber pair can support multiple wavelengths, and modern coherent optics can put hundreds of gigabits on one wavelength. The better reason to reserve extra strands is usually operational flexibility: future buildings, migrations, isolated networks, spare capacity and easier changes without reopening the dark-fiber agreement.
| Reservation | Best Fit | Main Limitation |
|---|---|---|
| 2 strands | Simple point-to-point network with DWDM option | No spare pair for migration or fiber failure |
| 4 strands | Strong default for many enterprises | Limited expansion if topology changes significantly |
| 6 to 8 strands | Growing campuses, several network functions or future facilities | Higher recurring strand cost |
| 12 strands | Long-term campus, difficult future additions, multiple buildings or strategic IRU | Can become expensive unused inventory |
Modern optical transport can place multiple high-capacity wavelengths across a single fiber pair. Current coherent platforms support 400G-class wavelengths, while newer packet-optical systems are scaling individual DWDM wavelengths considerably higher.
A business expecting capacity growth from 10G to 100G to several hundred gigabits may be able to solve that growth electronically instead of reserving ten additional dark strands.
Count the services that genuinely need their own physical strands. A business using one Ethernet backbone between two facilities may only need one active pair. Separate storage, research, security or operational networks can justify additional pairs, although VLANs and wavelengths can often provide logical separation without dedicated strands.
Topology changes the answer faster than bandwidth does. A pair running Building A to Building B cannot magically become an independent Building A to Building C route unless the physical fiber architecture supports it.
If another building, data room, plant or campus node is likely during the lease term, reserve strands and splice architecture around that physical expansion.
This is the strongest argument for moving from two strands to four. A second pair can support migration work, troubleshooting, temporary service or a replacement path if one strand becomes unusable.
It is operational redundancy inside the same route. It is not true geographic diversity.
If the organization is comfortable owning optical electronics, a small number of dark strands can support enormous capacity growth. The economics then become a comparison between additional fiber rent and DWDM equipment, optics, support and operational complexity.
Spare capacity is valuable only relative to its price. Every unused pair has a carrying cost over the full contract term.
The useful comparison is not “Would six extra strands be nice?” It is “Would we pay this much today to avoid the probable cost and difficulty of obtaining them later?”
Extra strands protect against strand-level problems and create operational flexibility. They do not protect against the failure modes that destroy the cable or route.
| Failure | Extra Strands Help? |
|---|---|
| One bad strand | Yes |
| Migration to new optics | Yes |
| Cable severed by excavation | No |
| Flooded or destroyed conduit | No |
| Bridge or route failure | No |
For true availability improvement, compare additional strands with a second geographically diverse route.
| Business Profile | Starting Point | Reasoning |
|---|---|---|
| Two offices, modest growth | 2 to 4 strands | One active pair, optionally one spare pair |
| Data-heavy enterprise using DWDM | 4 strands | Capacity grows through wavelengths, spare pair preserves flexibility |
| Growing three-building campus | 6 to 8 strands | More physical endpoints and migration options |
| Long-term healthcare, university or industrial campus | 8 to 12 strands | Long horizon, changing topology, isolated services and difficult future construction |
| Mission-critical site seeking redundancy | Two routes first | Route diversity often adds more resilience than ten extra strands in one cable |
Dark-fiber pricing is highly route-specific. Distance, market competition, construction requirements, term, maintenance and strategic value all affect price.
One currently published institutional price schedule lists a two-strand dark-fiber WAN at $1,500 per month on a 60-month term. That is $18,000 per year for the pair before considering any project-specific construction charge. It is useful as a real-world example, not a universal market benchmark.
An extra $6,000 per year of strand reservation becomes $60,000 over ten years before escalators. At that point, DWDM electronics or a better network design may be the cheaper form of future-proofing.
Reserving additional fibers is more defensible when the agreement lasts 15 or 20 years and future capacity may be difficult to obtain. Large infrastructure agreements sometimes explicitly retain substantial incremental strand counts for future use.
That does not mean an ordinary enterprise should copy a national carrier’s strategy. It does show that spare-strand value rises as the planning horizon lengthens and the cost of coming back for additional fiber increases.
Estimate a practical strand reservation and see the long-term cost of keeping the spare capacity.
