Intel E810-CQDA2 100GbE Dual-Port QSFP28 Ethernet Network Adapter

The Intel E810-CQDA2 is a dual-port 100 Gigabit Ethernet network adapter built on Intel's fourth-generation Ethernet controller architecture, delivering line-rate 100GbE connectivity via two QSFP28 ports in a full-height, half-length PCIe 4.0 x16 form factor. The adapter leverages Intel's Application Device Queues (ADQ) technology to isolate network traffic into dedicated hardware queues, reducing latency jitter for latency-sensitive workloads and improving determinism for storage, trading, and real-time analytics applications. Hardware-based RDMA over Converged Ethernet (RoCEv2) offloads memory transfer operations directly from the CPU, enabling high-throughput, low-latency data movement essential for distributed computing and storage fabrics.

The E810-CQDA2 integrates a programmable pipeline that supports Precision Time Protocol (IEEE 1588 PTPv2) hardware timestamping, enabling sub-microsecond synchronization across data-center infrastructure. The adapter includes support for SR-IOV with up to 256 Virtual Functions per port, making it well-suited for highly virtualized environments and NFV deployments. Advanced traffic management features include flow director, RSS (Receive Side Scaling), and hardware-based rate limiting, allowing fine-grained QoS control without host CPU overhead. The adapter is compatible with Intel's Dynamic Device Personalization (DDP) technology, allowing the packet processing pipeline to be reconfigized at runtime for specialized protocol offloads.

Targeted at mainstream enterprise and cloud data-center server deployments, the E810-CQDA2 (manufacturer part number E810CQDA2BLK) operates across a broad range of server platforms supporting PCIe 4.0 or PCIe 3.0 slots. It is fully supported under Linux (kernel 5.x and later via the ice driver) and Microsoft Windows Server environments, with additional support in VMware ESXi. The adapter's combination of high port density, protocol offload capabilities, and programmable pipeline makes it a broadly applicable solution for hyperconverged infrastructure, NVMe-oF storage networking, and high-performance computing interconnects.

Ideal for

• NVMe over Fabrics (NVMe-oF) storage networking, where RoCEv2 offload and low-latency ADQ queuing reduce storage access latency across disaggregated all-flash arrays
• High-frequency trading and financial services platforms requiring deterministic, sub-microsecond latency and precise IEEE 1588 PTP hardware timestamping for transaction ordering
• Hyperconverged infrastructure (HCI) deployments consolidating compute, storage, and networking traffic across dual 100GbE links with SR-IOV virtual function isolation per tenant workload
• AI and machine learning training clusters using RDMA-based collective communication libraries (e.g., NCCL) to accelerate GPU-to-GPU data transfers with minimal CPU involvement
• Network Functions Virtualization (NFV) and SD-WAN gateway servers leveraging SR-IOV and DDP pipeline programmability to offload custom protocol processing at line rate
• Large-scale virtualized server environments on VMware ESXi or KVM requiring high virtual function counts, hardware-based RSS, and reliable 100GbE uplink bandwidth per host

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.