Single-wavelength 100G technology is quietly transforming the data center interconnect landscape with a simpler architecture and lower costs.
In the evolution of high-speed data center interconnects, 100G optical modules have become an indispensable core component. Traditional 100G solutions, such as SR4 and LR4, rely on multi-channel or multi-wavelength parallel transmission. This complexity leads to high power consumption and high cost challenges.
Single-lambda (single-wavelength) solutions based on PAM4 modulation technology—100GBASE-DR1, 100GBASE-FR1, and 100GBASE-LR1—achieve 100G transmission using only a single wavelength, significantly simplifying the optical network architecture.
PAM4 Modulation and Single-Wavelength Transmission Principles
The core breakthrough of single-wavelength 100G technology lies in its use of PAM4 (four-level pulse amplitude modulation) technology. Compared to the two-level transmission (0/1, 1 bit per symbol) of traditional NRZ (Non-Return-to-Zero) encoding, PAM4 transmits two bits of data using four different voltage levels (corresponding to 00/01/10/11), doubling data throughput at the same baud rate.
This means that single-wavelength 100G modules no longer require four different wavelengths like LR4 or four independent optical channels like SR4.
The implementation of PAM4 technology requires a fully coordinated design of the entire link. The transmitter uses an electro-absorption modulated laser (EML) and a PAM4 driver chip to generate a four-level optical signal, combined with pre-emphasis technology to compensate for high-frequency losses.
The receiver uses an adaptive equalization filter to dynamically eliminate inter-symbol interference caused by fiber dispersion, and utilizes strong forward error correction (FEC) to address PAM4’s noise immunity disadvantages.
This technological breakthrough enables QSFP-100G-DR-S, QSFP-100G-FR-S, and QSFP-100G-LR-S to achieve 100G speeds without increasing the number of wavelengths, providing a more streamlined and efficient solution for data center interconnects.
Detailed Technical Parameters of DR1, FR1, and LR1
Although all belong to the single-wavelength 100G family, QSFP28 100G DR1, QSFP28 100G FR1, and QSFP28 100G LR1 each have their own technical features designed to meet different transmission requirements. Below is a comparison of the key parameters of the three types of modules:
| Parameter | DR1 | FR1 | LR1 |
| Transmission Distance | 500m | 2km | 10km |
| Wavelength | 1310nm | 1310nm | 1310nm |
| Modulation Format | PAM4 | PAM4 | PAM4 |
| Fiber Type | SMF | SMF | SMF |
| Typical Power Consumption | ≤4W | ≤4W | ≤4.5W |
| Interface Type | Duplex LC | Duplex LC | Duplex LC |
From a technical implementation perspective, all three types of modules utilize single-mode fiber and PAM4 modulation, but the main differences lie in transmission distance and corresponding optical design.
DR1 is designed for short-distance interconnection, suitable for intra-rack connections in data centers; FR1 covers intra-campus connections in data centers; and LR1 meets longer-distance interconnection needs, up to 10 kilometers.
In terms of power consumption, single-wavelength 100G modules offer significant advantages over traditional LR4 solutions. For example, Intel Silicon Photonics’ 100G DR/FR/LR QSFP28 optical modules have a maximum power consumption of 4.5W, while QSFPTEK FR1 modules consume only 4.3W. This translates to significant energy savings in high-density deployments.
How to Choose Based on Your Application Scenario?
The choice between DR1, FR1, or LR1 depends on three key factors: transmission distance, network architecture, and total cost of ownership.
DR1 (500m) is primarily used for internal data center connections, particularly inter-cabinet interconnection in leaf-spine architectures. For connecting rows of racks within large data center computer rooms, DR1’s 500m transmission distance is sufficient for most scenarios. Its value lies in its balance between cost and performance, making it suitable for high-density rack environments.
FR1 (2km) is an ideal choice for data center campus interconnection. Fiber optic connections between buildings within a campus typically fall within this distance range. FR1 modules, such as the HW QSFP-100G-FR1, feature a built-in Broadcom chip that converts 4x25G NRZ electrical signals into 1x100G PAM-4 optical signals. They are suitable for a variety of applications, including data centers, 100G Ethernet, and 400G to 4x100G breakout.
LR1 (10km) targets longer-reach applications, such as metropolitan area network connectivity or inter-data center campus interconnection. For scenarios requiring 10km transmission, LR1 provides a single-wavelength solution, avoiding the complexity of the traditional LR4 four-wavelength solution. LR1 is also a suitable choice for telecommunications applications such as 5G fronthaul.
In terms of network evolution compatibility, single-wavelength 100G modules are seamlessly compatible with future 400G networks. For example, FR1 modules can interoperate with 400G DR4+ optical modules in 4x100GbE breakout applications, protecting investments and lowering the barrier to entry for 400G evolution.
The Role of Single-Wavelength 100G in Data Center Development
Single-wavelength 100G technology is not only a solution for current high-speed interconnection but also a bridge for future network evolution. As 400G/800G technologies become more widespread, single-wavelength 100G, as a foundational technology, will continue to play a vital role.
In terms of power consumption optimization, the industry is developing more advanced solutions. Luxshare Technology’s LPO (Linear Drive Pluggable Optics) and LRO (Low Power Optics) architectures further reduce optical module power consumption.
The LPO architecture eliminates the DSP on both the transmit and receive sides, offering the lowest power consumption and cost. LRO, on the other hand, retains a traditional DSP solution on the transmit side but eliminates the DSP on the receive side, achieving a balance between power consumption and compatibility.
Another trend in single-wavelength 100G technology is its integration with new packaging options such as CPO (Co-Packaged Optics) and NPO (Near-Packaged Optics). These technologies integrate the optical engine more closely with the switch ASIC, reducing the number of electrical interfaces and further reducing power consumption and latency.
For future network architectures, the simplified design of single-wavelength 100G modules makes them easier to maintain and manage. With the development of tunable single-wavelength modules, network operators can further reduce spare parts inventory and increase network flexibility.
Conclusion
As data centers evolve toward 400G/800G, the importance of single-wavelength 100G technology will become increasingly prominent. Network architects face a choice not only to meet current needs but also to pave the way for future network evolution.
Compared to traditional LR4 solutions, single-wavelength solutions can reduce data center interconnect deployment complexity and costs, offering advantages in terms of technology lifecycle and total cost of ownership.
Whether choosing DR1, FR1, or LR1, the key is to comprehensively consider actual transmission requirements, fiber resource availability, and future evolution paths. In today’s AI-driven explosion of computing power, this choice will directly impact the competitiveness and scalability of data centers in the coming years.
