ConnectX-7 MCX75310AAS-HEAT Dual-Protocol 400Gb/s Smart Adapter Card

The NVIDIA ConnectX-7 MCX75310AAS-HEAT is a single-port 400Gb/s InfiniBand and Ethernet smart adapter engineered for the most demanding AI, scientific computing, and hyperscale cloud data center workloads. Featuring NVIDIA In-Network Computing engines, inline hardware security (IPsec, TLS, MACsec), and PCIe Gen 5.0 host connectivity, it delivers ultra-low latency, extreme throughput, and scalable acceleration from edge to core data centers.

• Ports1x QSFP112 (400Gb/s)
• Max Speed400Gb/s
• Host InterfacePCIe Gen 5.0 x16
• Form FactorPCIe HHHL

The NVIDIA ConnectX-7 MCX75310AAS-HEAT belongs to the ConnectX-7 family of Remote Direct Memory Access (RDMA) network adapters supporting InfiniBand and Ethernet protocols at speeds up to 400Gb/s. It enables smart, scalable, and feature-rich networking solutions that address traditional enterprise needs as well as the world's most demanding AI, scientific computing, and hyperscale cloud workloads.

ConnectX-7 provides a broad set of software-defined, hardware-accelerated networking, storage, and security capabilities, empowering agile and high-performance solutions from edge to core data centers to clouds. Its In-Network Computing engines, ultra-low latency, and extreme throughput deliver the acceleration, scalability, and feature-rich technology needed for modern data-driven scientific computing.

• Up to 400Gb/s InfiniBand (NDR) and Ethernet on a single port
• NVIDIA In-Network Computing engines with SHARP support and collective operations offloads
• Inline hardware security: IPsec, TLS, and MACsec encryption/decryption (AES-GCM 128/256-bit)
• GPUDirect RDMA and GPUDirect Storage for direct GPU and storage data paths
• PCIe Gen 5.0 with up to x32 lanes and multi-host support
• Advanced timing: IEEE 1588v2 PTP with 12ns accuracy, SyncE, configurable PPS
• Storage acceleration: NVMe-oF, NVMe/TCP, T10-DIF, SRP, iSER, NFS over RDMA, SMB Direct
• Platform security: secure boot with hardware root-of-trust, secure firmware update, device attestation

ConnectX-7 offloads collective operations and supports NVIDIA Scalable Hierarchical Aggregation and Reduction Protocol (SHARP), reducing CPU utilization and accelerating MPI, AI, and HPC workloads. The adapter includes in-network on-board memory and rendezvous protocol offload.

The adapter implements RDMA with send/receive semantics, hardware-based reliable transport, extended reliable connected transport (XRC), and dynamically connected transport (DCT). It supports adaptive routing, enhanced atomic operations, advanced memory mapping (UMR), on-demand paging (ODP), and registration-free RDMA memory access.

ConnectX-7 enables high-performance data storage by leveraging RDMA/RoCE, GPUDirect Storage, and hardware-based NVMe-oF offload engines. It supports block-level encryption (XTS-AES 256/512-bit key), NVMe over TCP, T10-DIF signature handover, SRP, iSER, NFS over RDMA, and SMB Direct.

Hardware engines offload and accelerate security with inline encryption/decryption of TLS, IPsec (over RoCE), and MACsec using AES-GCM 128/256-bit keys. Platform security includes secure boot with hardware root-of-trust, secure firmware update, flash encryption, and device attestation.

NVIDIA ASAP2 technology accelerates SDN and VNF workloads at line rate with no CPU penalty, supporting SR-IOV, VirtIO acceleration, overlay network acceleration (VXLAN, GENEVE, NVGRE), connection tracking (L4 firewall), flow mirroring/sampling, header rewrite, and hierarchical QoS.

• AI training and inference clusters requiring 400Gb/s InfiniBand or Ethernet connectivity
• Scientific computing and HPC clusters with SHARP and collective operations acceleration
• Hyperscale cloud data centers modernizing network infrastructure
• NVMe-oF and disaggregated storage architectures with GPUDirect Storage
• Virtualized and containerized environments (VMware ESXi, Kubernetes)
• Timing-sensitive infrastructures requiring IEEE 1588v2 PTP and SyncE
• Security-sensitive edge-to-core deployments with inline IPsec/TLS/MACsec

• InfiniBand: IBTA Spec 1.5 compliant; NDR 400Gb/s, HDR 200Gb/s, EDR 100Gb/s
• Ethernet: 400GbE, 200GbE, 100GbE, 50GbE, 25GbE, 10GbE; RoCE supported
• Interface technologies: NRZ (10G, 25G), PAM4 (50G, 100G)
• PCIe: Gen 5.0 compatible, 32 lanes, bifurcation support, NVIDIA Multi-Host (up to 4 hosts)
• Operating systems: in-box drivers for RHEL, Ubuntu, Windows; virtualization and containers: VMware ESXi (SR-IOV), Kubernetes
• Management: NC-SI, MCTP over SMBus/PCIe, PLDM DSP0248/DSP0267/DSP0218/DSP0257, SPDM DSP0274, SPI, JTAG

The MCX75310AAS-HEAT is a PCIe HHHL single-port adapter. For OCP 3.0 form factors or multi-port configurations, refer to other ConnectX-7 ordering part numbers.

Specifications are based on NVIDIA product documentation. Exact power consumption, physical dimensions, and port configuration for this specific OPN should be confirmed with the manufacturer or your sales representative before ordering.

The MCX75310AAS-HEAT is a single-port PCIe HHHL adapter. For dual-port, quad-port, or OCP 3.0 variants, refer to the ConnectX-7 user manuals or contact us for the correct ordering part number.

• 400Gb/s performance: Ultra-low latency and extreme throughput for AI, HPC, and cloud workloads.
• In-Network Computing: SHARP and collective operations offloads free CPU cycles and boost application efficiency.
• Security built in: Inline hardware IPsec, TLS, and MACsec encryption protect data in transit without performance penalties.
• Storage acceleration: GPUDirect Storage and NVMe-oF offloads improve storage performance and CPU utilization.
• Precision timing: IEEE 1588v2 PTP with 12ns accuracy and SyncE support for timing-sensitive applications.
• Flexible deployment: Supports InfiniBand and Ethernet, PCIe Gen 5.0, multi-host, and a broad range of operating systems.

Hong Kong Starsurge Group Co., Limited provides global delivery, multilingual support, and tailored network infrastructure solutions. Our experienced sales and technical team supports industries including government, healthcare, manufacturing, education, finance, and enterprise. We offer:

• Pre-sales technical consultation and compatibility verification for ConnectX-7 deployments
• Custom software development and network management systems
• IoT solutions and system integration
• Responsive after-sales support and RMA assistance

Contact us for volume pricing, lead times, and certification documents for this product.

• ESD handling: Observe standard ESD precautions during installation and removal. Use grounded wrist straps and ESD-safe work surfaces.
• Thermal: Ensure adequate airflow in the server chassis. High-speed 400Gb/s adapters require proper cooling to maintain performance and reliability.
• Power: Verify that the server provides sufficient power and PCIe Gen 5.0 slot bandwidth for the adapter's x16 host interface.
• Firmware and drivers: Install the latest NVIDIA OFED or in-box drivers and firmware for optimal performance and compatibility.
• Compatibility: Confirm that the server's PCIe slot, bracket type, and cooling configuration match the adapter variant before installation.

• Confirm required port count and speed (single-port 400Gb/s for MCX75310AAS-HEAT).
• Verify PCIe Gen 5.0 slot availability and x16 bandwidth.
• Check bracket type (HHHL) and server chassis compatibility.
• Confirm cable/transceiver type (QSFP112) and reach.
• Ensure adequate cooling and airflow for a 400Gb/s adapter.
• Request firmware/driver package and certification documents if required.
• Confirm quantity, lead time, and warranty terms with supplier.

• ModelMCX75310AAS-HEAT
• BrandNVIDIA
• Ports1x QSFP112
• Max Speed400Gb/s
• Host InterfacePCIe Gen 5.0 x16
• Form FactorPCIe HHHL
• ProtocolsInfiniBand & Ethernet
• SecurityIPsec, TLS, MACsec

• NVIDIA ConnectX-7 dual-port 400Gb/s PCIe adapter (refer to ConnectX-7 PCIe adapters manual)
• NVIDIA ConnectX-7 OCP 3.0 400Gb/s adapter (refer to ConnectX-7 OCP 3.0 adapters manual)
• NVIDIA ConnectX-6 MCX653106A-HDAT — dual-port 200Gb/s
• NVIDIA ConnectX-6 MCX653105A-HDAT — single-port 200Gb/s
• NVIDIA Quantum-2 InfiniBand switches for NDR 400Gb/s fabrics

• NVIDIA ConnectX-7 PCIe Adapters User Manual
• NVIDIA ConnectX-7 OCP 3.0 Adapters User Manual
• NVIDIA In-Network Computing and SHARP Technology Overview
• InfiniBand Trade Association (IBTA) Specification 1.5
• Data Center ESD Handling Best Practices

Hong Kong Starsurge Group Co., Limited is a technology-driven provider of network hardware, IT services, and system integration solutions. Founded in 2008, the company serves customers worldwide with products including network switches, NICs, wireless access points, controllers, cables, and related networking equipment. Backed by an experienced sales and technical team, Starsurge supports industries such as government, healthcare, manufacturing, education, finance, and enterprise.

The company also offers IoT solutions, network management systems, custom software development, multilingual support, and global delivery. With a customer-first approach, Starsurge focuses on reliable quality, responsive service, and tailored solutions that help clients build efficient, scalable, and dependable network infrastructure.