HPE P25962-001 

Ethernet 100 GbQSFP28 Network Adapter (2‑Port

The HPE P25962‑001 is a high-performance dual‑port 100 GbE Network Adapter featuring QSFP28 connectors and the Mellanox MCX623106AS‑CDAT (ConnectX‑6) chipset. Built for data centers, cloud infrastructure, virtualization, and high-performance networking, this adapter delivers ultra-high bandwidth up to 200 Gbps total throughput. Its PCIe 4.0 x16 interface ensures efficient server communication, while advanced offload capabilities such as SR‑IOV, VXLAN/NVGRE, and NVMe‑oF deliver low-latency, scalable networking that’s ideal for enterprise and HPC applications. This compact plug-in card supports modern network workloads and integrates seamlessly with existing Ethernet fabrics—making it an excellent choice for performance-critical environments.

Features

Built on Mellanox ConnectX‑6 EN technology, this adapter offers dual QSFP56/28 ports capable of 100 Gbps each, totaling 200 Gbps bidirectional throughput. Compliant with IEEE 802.3 standards (802.3by/bj/bm/ba), it supports Ethernet protocols, RDMA over Converged Ethernet (RoCE), GPU Direct, and virtualization offloads such as SR‑IOV, VXLAN/NVGRE tunneling. Its PCI Express 4.0 x16 interface unlocks full bandwidth potential, reducing CPU overhead and enhancing application performance in cloud, virtualization, or NFV workloads. Intelligent virtualization and forwarding features enable high-density virtual machine connectivity and scalable provisioning.DOM diagnostics provide real-time monitoring of port parameters, including optical link health, temperature, and voltage. With compact passive cooling and a low-profile form factor, the card is compatible with HPE ProLiant and Apollo server platforms, delivering robust performance in dense compute clusters or storage fabrics.

Advanced Options

Engineered for demanding environments, the P25962‑001 adapter delivers reliable ultra-high-speed connectivity with low latency and robust offload performance. The ConnectX‑6 architecture enables efficient RDMA and GPU acceleration, resulting in significant CPU savings and faster data movement for virtualization and HPC tasks.