Mellanox 01LT122 InfiniBand SX6720 36x QSFP+ 56Gb Infiniband Switch

The SX67X0 systems provide the highest performing fabric solution in a 1U form factor by delivering up to 4Tb/s of non-blocking bandwidth with 200ns port-to-port latency. These systems are the industry's most cost-effective building blocks for embedded systems and storage with a need for low port density systems. Whether looking at price-to-performance or energy-to-performance, these systems offer superior performance, power and space, reducing capital and operating expenses and providing the best return-on-investment. The systems are an ideal choice for smaller departmental or back-end clustering uses with high performance needs, such as storage, data base and GPGPU clusters. Powerful servers combined with high-performance storage and applications that use increasingly complex computations are causing data bandwidth requirements to spiral upward. As servers are deployed with next generation processors, High-Performance Computing (HPC) environments and Enterprise Data Centers (EDC) need every last bit of bandwidth delivered with Mellanox’s FDR InfiniBand systems. Built with Mellanox’s sixth generation SwitchX®-2 InfiniBand FDR 56Gb/s system device, these standalone systems are an ideal choice for top-of-rack leaf connectivity or for buildingsmall to extremely large sized clusters. These systems enable efficient computing with features such as static routing, adaptive routing, and advanced congestion management. These features ensure the maximum effective fabric bandwidth by eliminating congestion. The managed systems comes with an onboard subnet manager, enabling simple, out-of-the-box fabric bring-up for up to 648 nodes. MLNX-OS® software delivers complete chassis management of firmware, power supplies, fans, ports and other interfaces. Mellanox’s edge systems can also be coupled with Mellanox’s Unified Fabric Manager (UFM®) software for managing scale-out InfiniBand computing environments. UFM enables data center operators to efficiently provision, monitor and operate the modern data center fabric. UFM boosts application performance and ensures that the fabric is up and running at all times. InfiniBand systems come as internally or externally managed. Internally managed systems come with a CPU that runs the management software (MLNX-OS®) and management ports which are used to transfer management traffic into the system. Externally managed systems come without the CPU and management ports and are managed using firmware tools. Mellanox's InfiniBand to Ethernet gateway, built with Mellanox's SwitchX®-2 based systems, provides the most cost-effective, high-performance solution for data center unified connectivity solutions. Mellanox's gateways enable data centers to operate at up to 56Gb/s network speeds while seamlessly connecting to 1, 10 and 40GbE networks with low latency (400ns). Existing LAN infrastructures and management practices can be preserved, easing deployment and providing significant return-on-investment.
This device is perfect for:

• High-Performance Computing (HPC): Infiniband is the standard for connecting nodes in HPC clusters, where minimal latency and maximum bandwidth are critical for parallel processing and data-intensive tasks.
• Artificial Intelligence (AI) and Machine Learning (ML) Clusters: Training large AI models and running machine learning workloads requires massive data transfer between GPUs and CPUs. An Infiniband switch like the SX6720 can provide the necessary high-speed fabric.
• Data Center Interconnect (DCI): It can be used to create a very fast interconnect between data center racks or clusters, especially for applications like IP storage where a lossless, low-latency network is essential.
• Financial Services: In high-frequency trading and other latency-sensitive financial applications, every microsecond counts. Infiniband provides a significant advantage over traditional Ethernet.
• Cloud and Virtualized Environments: While Ethernet is more common, Infiniband can be used in cloud environments for specific workloads that benefit from its superior performance, such as storage virtualization or high-speed inter-VM communication.