Building or upgrading a metropolitan area network (MAN) or regional long-haul link requires careful fiber infrastructure planning. Unlike data center intra-rack connections where OSFP112-400G-VSR4 and QSFP56-DD-400G-VSR4 dominate, metro and long-haul networks rely on cost-effective 100G solutions such as QSFP28 100G ZR4, QSFP28 100G BIDI 40KM, QSFP28 100G BIDI 80KM, and coherent QSFP28 100G 100KM modules. This article provides a practical framework for fiber infrastructure planning—covering fiber types, connector selection, link budgeting, amplification needs, and future-proofing for 400G coherent and even 400G VSR4/DR4 extensions. Network architects will learn how to choose between BIDI and duplex solutions, when to deploy optical amplifiers, and how to migrate existing 100G long-haul links to higher speeds without re-laying fiber.
Typical metro distances range from 10km to 80km, while regional long-haul reaches 80km to 200km. For these applications, the transceiver hierarchy is:
10–40km: QSFP28 100G ER4 (40km) or QSFP28 100G BIDI 40KM (single-fiber).
40–80km: QSFP28 100G ZR4 (duplex) or QSFP28 100G BIDI 80KM (single-fiber).
80–120km: Coherent QSFP28 100G 100KM (dual fiber, often with amplification).
Beyond 120km: 100G coherent with external amplifiers or 400G ZR (future).
Fiber infrastructure decisions today must accommodate both current 100G needs and future upgrades to 400G coherent (400G ZR/ZR+), as well as potential 400G VSR4/DR4 if metro data center interconnects (DCI) become shorter.
One of the most critical planning decisions is whether to deploy fiber pairs (duplex) or utilize single fibers with BIDI technology.
Duplex uses two fibers (one transmit, one receive). Advantages:
Mature standard, widely interoperable.
Lower launch power per fiber, easier link budget.
Less sensitive to reflections compared to BIDI.
Disadvantage: Consumes twice the fiber strands. If dark fiber is leased by the pair, cost is higher.
BIDI transmits and receives on two different wavelengths over one fiber. Advantages:
Halves fiber usage—ideal when fiber is scarce or leased per strand.
Can double capacity on existing infrastructure (e.g., turn a duplex pair into two independent BIDI links).
Disadvantages:
Tighter optical budget due to wavelength separation filters.
More sensitive to back-reflection; requires cleaner connectors and often APC polish.
Must use matched pairs; less standard than ZR4.
Decision rule: If fiber lease cost per strand > $200/month, BIDI pays back quickly. If fiber is owned or plentiful, ZR4 offers simpler deployment.
All long-haul optics use single-mode fiber (SMF), typically G.652.D. However, chromatic dispersion becomes critical beyond 40km.
| Transceiver | Dispersion Tolerance (ps/nm) | Max Distance on G.652 (standard) | Need for DCM* |
|---|---|---|---|
| QSFP28 100G ER4 (40km) | ~700 ps/nm | 40km (≈680 ps/nm) | No |
| QSFP28 100G ZR4 (80km) | ~1600 ps/nm | 80km (≈1360 ps/nm) | No |
| QSFP28 100G BIDI 80KM | ~1200–1500 ps/nm (depends) | 80km (may need margin) | Sometimes (use low-dispersion fiber) |
| QSFP28 100G 100KM coherent | >40,000 ps/nm (DSP compensated) | 100–200km | No |
*DCM = Dispersion Compensation Module. For ZR4, 80km is within tolerance. For BIDI 80KM, some implementations require dispersion compensation if fiber has high dispersion coefficient. Always check datasheet.
For future 400G coherent (400G ZR), dispersion tolerance is even higher, so G.652 is fine up to 120km without DCM.
Accurate link loss calculation is mandatory. Use this formula: Total loss = (fiber length × loss/km) + connector loss × number of connections + splice loss.
Typical values:
G.652.D fiber loss: 0.22–0.25 dB/km at 1310nm (for ZR4/ER4/BIDI 40/80), 0.19–0.22 dB/km at 1550nm (for coherent 100KM).
Connector loss (LC UPC): 0.3–0.5 dB per mated pair.
Splice loss: 0.05–0.1 dB per splice.
Example for QSFP28 100G ZR4 over 75km with 6 connectors and 10 splices:
Fiber loss: 75 × 0.23 = 17.25 dB. Connectors: 6 × 0.4 = 2.4 dB. Splices: 10 × 0.1 = 1.0 dB. Total = 20.65 dB. ZR4 link budget is 21 dB. Marginal but works. Add 1dB margin: too close; consider amplifier or reduce connectors.
If loss exceeds budget, deploy an EDFA (Erbium-Doped Fiber Amplifier) or SOA (Semiconductor Optical Amplifier). For coherent 100KM modules, budgets are 25–30 dB, so amplification is often not needed until >100km.
For QSFP28 100G BIDI 80KM, the loss budget is typically 22 dB. However, reflections add effective loss; keep return loss >45dB.
400G coherent (400G ZR, 400G ZR+) is emerging for metro DCI. It uses 16QAM or QPSK modulation and can reach 80–120km without amplification. To prepare your fiber plant:
Ensure low PMD (Polarization Mode Dispersion):<0.1 ps/√km. G.652.D is fine.
Minimize connectors and splices; use fusion splicing where possible.
Document exact fiber length and loss at both 1310nm and 1550nm.
Consider deploying bidirectional amplifiers (for 400G ZR+).
Note: 400G VSR4 (OSFP112-400G-VSR4) and 400G DR4 (QSFP56-DD-400G-DR4) are not designed for long-haul; they are for data center internal use (≤500m). Do not confuse them. However, some new standards (400G LR4-10km) exist for campus links, but they are not the focus of metro planning.
For QSFP28 100G BIDI 40KM and 80KM, back-reflection is a serious issue. Standard UPC (Ultra Physical Contact) connectors have return loss of ~55dB, which may be acceptable for short BIDI links but marginal for 80km. APC (Angled Physical Contact, typically 8°) provides >65dB return loss, drastically reducing reflections. Many BIDI modules are designed for APC connectors. However, APC is not compatible with UPC; adapters or hybrid cables are needed. Plan accordingly:
If your existing plant is UPC, you have three options: change all connectors to APC (expensive), use an APC-to-UPC patch cord (degrades return loss), or accept higher reflection risk. For 80km BIDI, APC is strongly recommended.
For ZR4/ER4, UPC is sufficient.
A municipal utility owned a single dark fiber strand between two substations 72km apart. They needed 100G connectivity. Options:
Lease a second fiber ($800/month) and use QSFP28 100G ZR4 on duplex. 5-year cost = $48,000 lease + $5,000 modules = $53,000.
Use QSFP28 100G BIDI 80KM on the existing single fiber. Module cost $3,000 per pair, no extra lease. 5-year cost = $3,000.
They chose BIDI. They upgraded connectors to APC (cost $2,000). The link operated at BER<1e-15. This case illustrates the dramatic TCO advantage of BIDI in fiber-constrained environments.
Long-haul optics require proactive monitoring:
Enable DOM (Digital Optical Monitoring) for all QSFP28 100G ZR4, BIDI, and 100KM modules. Set alarms for Tx power deviation ±1dB, Rx power below sensitivity+2dB.
Perform OTDR sweeps annually to detect fiber degradation.
For BIDI links, test return loss periodically; high reflection (> -40dB) indicates connector issues.
A typical enterprise has a data center using OSFP112-400G-VSR4 and QSFP56-DD-400G-DR4 internally, and needs to connect to a remote office 60km away. The data center edge router should have a dedicated 100G port with a QSFP28 100G ZR4 (or BIDI) module. Do not try to convert a 400G DR4 port directly to long-haul. Use separate equipment. When budgeting, allocate dark fiber pairs for the long-haul link separately from data center fiber.
Yes, but reflection may cause errors. Test the link with a reflectometer. If return loss is<45dB, consider upgrading to APC or adding an optical isolator.
400G coherent ZR uses duplex SMF, same as 100G ZR4. So one pair is enough. For redundancy, plan two pairs.
Yes, but ensure the EDFA’s gain flatness covers the module’s wavelength (typically C-band). Also, input power to the EDFA must not exceed the module’s maximum receiver power.
Yes. The ZR4 will work with ER4, but the ER4 receiver may be overloaded if the ZR4 transmits at high power. Use an attenuator on the ER4 side.
Nominal 40km. With low-loss fiber (0.20 dB/km) and clean connectors, some modules reach 50km, but not guaranteed.
No. VSR4 is for ≤100m. For 2km, use 400G FR4 or 100G LR4. But 400G FR4 is expensive; better to use 100G LR4 bundled.
No. DR4’s electrical and optical design is for 500m only. The lasers are not specified for long-haul, and the DSP lacks dispersion compensation. Use proper ZR/BIDI/coherent modules.
Metro and long-haul optical infrastructure requires deliberate choices: duplex QSFP28 100G ZR4 for simplicity, single-fiber BIDI 40KM/80KM for fiber conservation, and coherent QSFP28 100G 100KM for ultra-long spans. By measuring link loss, selecting appropriate connectors (APC for BIDI), and documenting dispersion parameters, you can achieve reliable 100G links today and ease the upgrade to 400G coherent tomorrow. While data center networks embrace OSFP112-400G-VSR4 and QSFP56-DD-400G-DR4, the metro domain will evolve more slowly, making careful fiber planning a long-term asset.
Our team provides end-to-end consulting for fiber infrastructure design, including loss budgeting, BIDI vs ZR4 trade-off analysis, and 400G coherent readiness assessments. We supply all transceivers – QSFP28 100G ZR4, BIDI 40/80KM, 100KM coherent, plus data center 400G modules – with free link testing and compatibility certification. Contact us to future-proof your metro network today.
Headquarter address :Room 1603, Coolpad Building B, North District of Science and Technology Park, Nanshan District, Shenzhen,China.518057
