QSFP+, QSFP28, QSFP-DD and OSFP
A practical engineering guide to high-density pluggable form factors used in 40G, 100G, 200G, 400G, 800G and emerging 1.6T network architectures.
Executive Summary
QSFP+, QSFP28, QSFP-DD and OSFP are high-density pluggable form factors designed for multi-lane optical and copper interconnects. They are commonly used in data centers, enterprise core networks, service provider infrastructure, AI clusters and high-performance computing environments. In simple terms: QSFP+ is mainly associated with 40G, QSFP28 with 100G, QSFP-DD with 200G/400G/800G evolution, and OSFP with high-power, high-performance 400G/800G and next-generation AI interconnects.
Key Takeaways
- QSFP+ typically means 4x10G electrical lanes and 40G-class operation.
- QSFP28 typically means 4x25G lanes and 100G-class operation.
- QSFP-DD doubles density with an 8-lane interface and supports 200G, 400G and 800G generations.
- OSFP is physically larger and optimized for thermal headroom in 400G/800G and AI data-center links.
- Form factor selection is not only about speed; it also affects power, airflow, cable management and switch faceplate density.
- Breakout support is architecture-dependent: common examples include 40G to 4x10G, 100G to 4x25G, and 400G to 4x100G.
- QSFP-DD ports are generally designed to preserve backward compatibility with QSFP modules, but platform support must be verified.
- OSFP and QSFP-DD are competing high-speed ecosystems; compatibility depends on the host cage, electrical interface, module type and vendor implementation.
1. Why high-density pluggables matter
As network bandwidth increased from 10G and 40G toward 100G, 400G and 800G, equipment vendors needed compact front-panel interfaces that could deliver more capacity per rack unit. High-density pluggables solved this problem by combining multiple electrical and optical lanes into one serviceable module.
For engineers, the key advantage is density. A switch can provide many high-speed ports in a compact form factor while still allowing field replacement of optics, DACs, AOCs and breakout cables. The trade-off is that higher density increases the importance of thermal design, signal integrity and platform-specific compatibility.
2. QSFP+: the 40G generation
QSFP+ stands for Quad Small Form-Factor Pluggable Plus. It is most commonly associated with 40GbE applications based on four 10G-class lanes. Common deployments include 40GBASE-SR4 over multimode fiber, 40GBASE-LR4 over single-mode fiber and 40G DAC or AOC links.
QSFP+ is still present in many brownfield data centers, enterprise aggregation layers and storage environments. It is mature, cost-effective and useful where 40G capacity is sufficient or where the existing switching infrastructure was designed around 4x10G lane architecture.
3. QSFP28: the 100G workhorse
QSFP28 keeps the same general QSFP-style footprint but raises the lane rate to the 25G class, enabling 100GbE with four lanes. It became one of the most important form factors for 100G data center and enterprise core networks.
Typical QSFP28 variants include 100G SR4 for short-reach multimode fiber, CWDM4 and LR4 for single-mode fiber, PSM4 for parallel single-mode architectures, and DAC/AOC solutions for shorter distances. QSFP28 is also widely used for 100G to 4x25G breakout, making it a flexible option for server access and leaf-spine designs.
4. QSFP-DD: double-density evolution
QSFP-DD means Quad Small Form-Factor Pluggable Double Density. It expands the QSFP concept from four lanes to eight lanes, allowing higher aggregate bandwidth in a compact front-panel footprint. Depending on lane speed and modulation, QSFP-DD supports multiple generations, including 200G, 400G and 800G implementations.
A major value of QSFP-DD is density. It allows switch vendors to place many 400G or 800G ports on a faceplate while preserving a familiar operational model. Many QSFP-DD cages are designed to accept earlier QSFP modules mechanically, but the actual operating modes depend on the switch platform, firmware and port configuration.
5. OSFP: thermal headroom for 400G and 800G
OSFP stands for Octal Small Form-Factor Pluggable. It is an 8-lane form factor designed with more physical volume and thermal capacity than QSFP-DD. That extra thermal headroom is one reason OSFP is common in high-performance 400G and 800G environments, especially AI, HPC and dense data-center interconnect applications.
OSFP is not physically the same as QSFP-DD. Engineers should not treat OSFP and QSFP-DD as interchangeable module formats. The switch cage, heat sink design, airflow model, power class, and vendor implementation determine which module family can be used.
6. Lanes, signaling and breakout behavior
The lane architecture is central to understanding these form factors. QSFP+ typically aggregates four 10G-class lanes. QSFP28 aggregates four 25G-class lanes. QSFP-DD and OSFP use eight lanes and can scale by increasing per-lane speed and modulation complexity.
Breakout support depends on how the switch ASIC and port group are designed. Common breakout examples include 40G QSFP+ to 4x10G SFP+, 100G QSFP28 to 4x25G SFP28, 400G QSFP-DD to 4x100G QSFP28, and 800G OSFP to 2x400G or 8x100G, depending on the hardware and cable assembly.
7. Optical modules vs DAC, AOC, ACC and AEC
High-speed pluggable ports can use optical transceivers, passive DAC, active copper cable (ACC), active electrical cable (AEC) or active optical cable (AOC), depending on reach and system design. Passive DAC is cost-effective for very short connections, but it becomes less practical as reach and speed increase.
At 400G and 800G, cable choice becomes an architectural decision. Engineers must consider bend radius, airflow, power consumption, insertion loss, latency, retimer requirements and serviceability. In AI clusters, short-distance DAC/AEC and longer-distance AOC/optical options may coexist in the same deployment.
8. Compatibility and coding considerations
High-speed optics and cables should be selected for both physical compatibility and host recognition. A module may have the correct connector and speed class but still fail to operate if the switch or NIC requires a specific coding profile, firmware support or port-mode configuration.
For multi-vendor environments, engineers should validate the exact switch model, OS version, port speed, FEC mode, breakout configuration and module EEPROM coding before large-scale deployment. This is especially important for 400G and 800G platforms where thermal and power limits are tighter.
9. Where each form factor is typically used
QSFP+ is usually found in 40G legacy and aggregation networks. QSFP28 is the practical standard for 100G enterprise core and data-center deployments. QSFP-DD is widely used where high-density 400G and 800G switching is required. OSFP is frequently selected in AI and HPC environments where high power, airflow, and thermal headroom are critical design factors.
10. How engineers should select the right form factor
Start with the switch or NIC port specification. Identify whether the port is QSFP+, QSFP28, QSFP-DD or OSFP, then confirm the supported speeds, FEC modes and breakout options. Next, choose the media type based on reach, density and operational constraints.
For new AI or high-density data-center deployments, pay special attention to thermal margins and cable management. For brownfield enterprise environments, backward compatibility and migration strategy may matter more than maximum speed. In all cases, lab validation is recommended before volume rollout.
Engineering Selection Checklist
- Confirm host cage type: QSFP+, QSFP28, QSFP-DD or OSFP.
- Verify supported speeds: 40G, 100G, 200G, 400G, 800G or emerging 1.6T modes.
- Check lane architecture: 4-lane vs 8-lane operation and supported modulation.
- Validate media choice: passive DAC, active cable, AOC or optical transceiver.
- Review breakout support: port groups, FEC mode and platform configuration.
- Evaluate thermal limits: especially for 400G/800G modules and dense AI switches.
- Confirm coding compatibility: host vendor, firmware version and EEPROM profile.
- Test before rollout: verify link-up, error counters, DOM telemetry, FEC behavior and stability under load.
Bottom Line
QSFP+, QSFP28, QSFP-DD and OSFP are not simply different names for the same connector. They represent different generations of lane architecture, speed, density and thermal design. QSFP+ is the 40G generation, QSFP28 is the 100G workhorse, QSFP-DD is a high-density 400G/800G evolution path, and OSFP is a thermally robust form factor widely used in demanding AI and HPC interconnect environments.
For support selecting QSFP, QSFP-DD, OSFP, DAC, AOC, AEC or breakout solutions for enterprise, data center or AI network platforms, contact ATL Optics for a compatibility review.