Specifications
EXP100-32X
The Edgecore EXP100-32X is a high-performance, low-power, low-latency, fully-programmable SDN data center switch designed for data center, cloud, and AI networks.
Key Features
- Supports 3.2 Tbps full-duplex throughput
- Supports 25G NRZ SerDes with a retimer-less design
- The 32 x 100G QSFP28 ports can breakout to 2 x 50G, 4 x 25G, or 4 x 10G
- UEC-ready programmable congestion measurement and signaling to facilitate dynamic load balancing and fast failure detection and recovery
- UEC-ready weighted congestion-aware dynamic load balancing for reducing tail latency
- Support front to back airflow; PUSs and fans redundancy
- Preloaded with ONIE, supports Xsight OS and Community SONiC
- Fully Programmable
3.2T Data Center Switch
32 x 100G QSFP28 ports, each can breakout to 4 x 25G, or 4 x 10G.
High performance, low power, low latency, fully programmable SDN data center switch designed for data center, cloud, and AI networks.
| Parameter | Specification | Parameter | Specification |
|---|---|---|---|
Switch Ports | 32 x QSFP28 | Timing and Sync | Optional |
Switch Silicon | Xsight Labs X2 | Fans | 5+1 hot-swappable redundant fans front to back |
Processor | Intel® Xeon® D-1713NTE 4-Core 2.2GHz | Storage Temperature | -40°C – 70°C (-40°F – 158°F) |
BMC | Optional | Operating Temperature | 0°C – 40°C (32°F – 104°F) |
Memory | 32GB DDR4 SO-DIMM with ECC | PSU | 2 redundant, load-sharing, hot-swappable AC PSUs |
Storage | 1TB m.2 2280 NVMe SSD | Power Draw | Maximum 863W |
Latency | 755 ns | ||
Rack | 1 RU form factor |
Connectivity examples

Compatible
Software Option
EXP-OS

Compatible
Transceivers
- 100G SR4
- 100G LR4
- 100G DR1
- 100G FR1
- 100G LR1

Compatible
Cables
- 100G DAC
- 100G AOC
Applications
Deployment scenarios for this module
Data Center & Enterprise Networks
High-speed connectivity for leaf-spine architectures, server aggregation, and backbone upgrades supporting growing east-west traffic.
Telecom & Service Provider Transport
Robust optical connectivity for aggregation, metro, and long-haul transport across carrier and service provider networks.
Cloud, AI & Hyperscale Infrastructure
Reliable, low-latency links designed for high-performance computing, AI workloads, and scalable cloud environments.
Overview
Leaf or ToR switch
The EXP100-32X is a 32-QSFP28-port switch in a 1RU form factor. Each of the QSFP28 ports can breakout to 2 x 50G, 4 x 25G, or 4 x 10G, optimizing compute racks by enabling 25Gbps-per-lane server links and 100G uplink optics. The device is an ideal leaf or ToR switch for cloud data center operators that are driving the rapid adoption of 100Gbps optical modules while also upgrading compute server NICs to 100Gbps speeds. aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
Xsight Instruction Set Architecture
The EXP100-32X also supports Xsight Instruction Set Architecture (ISA) for users to define flexible packet processing rules,routing protocols, and network policies. The programmability is also an important underlying capability for many of the advanced features, including queue management, congestion control, and telemetry. The programmability handles newly-defined notification mechanisms, such as the Ultra Ethernet Transport (UET) protocol, packet trimming, and back-to-sender signaling.
Dynamic Load Balancing (DLB)
Optimized for interconnecting AI/ML, storage, and compute clusters in data centers, the EXP100-32X supports congestion-aware routing, which comprises measuring and signaling congestion as well as load balancing across the network fabric. For a given fabric link, the EXP100-32X can measure flow rate, latency, and queue level. When congestion is detected, it can signal using Explicit Congestion Notification (ECN), Congestion Signaling (CSIG), or other customer-defined mechanisms.
Warranty & Compatibility
Zero Risk
Using our transceivers does not void the warranty of your network equipment. Global OEM policies and consumer protection laws (like the Magnuson-Moss Warranty Act) prevent manufacturers from denying support based solely on the use of third-party components.
Technical Support & RMA
Have an issue? We provide full technical support and hassle-free RMA on all eligible transceivers. Our team is ready to help with any technical questions about compatibility or deployment. contact our technical support: support@ecin.ca
Q&A
Commonly asked questions
No. Under the Magnuson-Moss Warranty Act and similar global consumer protection laws, OEMs cannot void your equipment warranty simply because you use a third-party transceiver. Your equipment warranty remains intact.
Yes. Our modules can be coded for major OEM platforms (Cisco, Juniper, Arista, Nvidia, etc.) and support open networking environments including SONiC.
A compatible generic version can be provided and coded before delivery.
Common 400G QSFP-DD optic types include DR4 for parallel single-mode fiber up to 500m, FR4 for duplex single-mode fiber up to ~2km, LR4 for duplex single-mode fiber up to ~10km, short-reach multimode options such as SR8 / SR4.2, and extended/coherent options such as ZR / ZR+ for longer DCI and metro applications. The attached datasheet specifically covers the EN-QDD-DR4, a 400GBASE-DR4 QSFP-DD module using 4 parallel optical lanes, 1310nm wavelength, MTP/MPO-12 APC connector, and up to 500m transmission over single-mode fiber with FEC.
Choose the 800G QSFP-DD transceiver according to the required distance and installed fiber infrastructure. DR8 is best for data-center links up to 500 m and uses parallel single-mode fiber with an MPO-16/APC connector. 2×400G FR4 supports distances up to 2 km over two duplex single-mode fiber pairs, making it suitable for aggregation and short data-center interconnects. 2×400G LR4 extends the reach to 10 km over duplex single-mode fiber for campus, inter-building, and metro-access applications. FR4 and LR4 connector configurations may use dual LC or dual CS interfaces, depending on the module. Always verify the host platform’s breakout modes, connector type, FEC configuration, power budget, and QSFP-DD800 compatibility before deployment.
Yes. Our modules are coded for the specific platform you select at purchase. They are recognized natively by the switch OS without any additional configuration or commands.
No, E.C.I. NETWORKS 400G transceivers are designed to meet applicable IEEE and MSA specifications, deliver the required 400G data rate and optical reach for their optic class, and support standards-based monitoring such as DDM/DOM. Actual link performance depends on choosing the right 400G type, such as DR4, FR4, LR4, SR8, or ZR/ZR+, and matching it to the correct cabling, optical budget, switch compatibility, and thermal environment.
Yes. We offer full support, RMA service, and warranty coverage on all modules.
Resources
Downloads and documentation







