We began this NVIDIA blog series discussing how the modern data center requires a software-defined, hardware-accelerated
network infrastructure to achieve cloud-scale efficiency. This time, we look at an essential building block within that infrastructure:
NVIDIA Spectrum switches. Learn more about the three pillars underlying Spectrum switches, which are designed for today’s data center.
Operational efficiency
NVIDIA Spectrum Ethernet switches support a wide variety of modern network operating systems, including NVIDIA Cumulus Linux, SONiC and NVIDIA Linux Switch. Cumulus Linux is an innovative, open network operating system, developed for easy implementation, management, customization and scalability. It delivers the most comprehensive end-to-end automation for a continuous integration and continuous delivery (CI/CD) pipeline. IT departments can use their existing Linux-based management tools and talent to spend more time on innovation and less on maintenance, reducing OpEx. Native integration and best-of-breed tools for monitoring and analytics enable faster IT delivery.
Cumulus Linux also powers NVIDIA Air, which makes physical deployments seamless by validating and simplifying policies and upgrades in a virtual network. The cloud-based NVIDIA Air environment is a data center digital twin platform that IT teams can use to build, simulate and experience a modern data center. This digital twin pre-production environment brings modern cloud-scale networking within easy reach. Furthermore, enterprises can reduce their CapEx due to reduced hardware needs, and lower OpEx by using the NVIDIA Air public cloud operational model.
Hardware efficiency
NVIDIA Spectrum Ethernet switch systems are powered by the purpose-built Spectrum family of ASICs which leverages more than 20 years of innovation from NVIDIA to enable accelerated Ethernet for any data center workload.
The Spectrum family of switches are built to meet the demands of modern 25/100G and 200/400G networks. Spectrum switches deliver best-in-class latency, packet rates and acceleration features such as adaptive routing. They also simplify solution design and deliver best-in-class power efficiency and a lower total cost of ownership. With a fully shared buffer architecture, the Spectrum switches enable consistent, fair and predictable performance.
NVIDIA Spectrum performance ensures business-critical applications and storage operate efficiently in enterprise cloud environments. Spectrum’s best-in-class performance accelerates NVIDIA AI platforms, including small DGX clusters, EGX architectures and Omniverse/OVX platforms—maximizing the investment in the AI infrastructure.
Visibility and telemetry
NVIDIA NetQ is a modern operation tool that provides holistic, real-time visibility, troubleshooting and continuous validation for the modern data center.
Users can test network configuration changes and software upgrades with NetQ in the NVIDIA Air environment and correct errors before deploying to their production network. In the physical production network, NetQ validations provide insight into the live state of the network, reduce troubleshooting time, and prevent network issues including MTU mismatch and VLAN misconfigurations.
NVIDIA Spectrum switches also come with What Just Happened (WJH), a hardware-accelerated telemetry feature that streams comprehensive, contextual telemetry data for analysis. WJH gives IT teams real-time visibility into network problems, such as hardware packet drops due to misconfigurations, buffer congestion and other issues.
While WJH provides always-on detection of packet drops, latency and congestion events, flow telemetry offers on-demand analysis of specific application flows. The combination of WJH with flow telemetry analysis lets network operators proactively identify root cause server and application issues and inform administrators about the possible outage or performance impact.
Data center solutions from NVIDIA and Ingram Micro
Ingram Micro supplies channel partners with NVIDIA switches, adapters and cables, either ordered standalone or as part of leading OEM servers. In addition, Ingram Micro
Integration Services can build complete hardware systems that support a broad range of applications, including hybrid cloud, hyperconverged infrastructure (HCI) and software-defined networking (SDN).
Watch for our next NVIDIA blog, which will cover NVIDIA Cumulus Linux and NVIDIA Air in more detail.
Learn more about NVIDIA Spectrum switches for the accelerated data center at
https://www.nvidia.com/en-us/networking/ethernet-switching/.
For more information, contact an NVIDIA networking expert at
https://www.nvidia.com/en-us/networking/contact/.
Contact Us