Cisco and NVIDIA Partner to Create a Cross-Portfolio Unified Architecture

0

Cisco is expanding its partnership with NVIDIA to accelerate AI adoption in the enterprise by bringing together two networking portfolios, Cisco Silicon One and NVIDIA Spectrum-X.
NVIDIA will enable Cisco Silicon One, coupled with NVIDIA SuperNICs to become part of the NVIDIA Spectrum-X Ethernet networking platform.

Cisco will build systems combining NVIDIA Spectrum silicon with Cisco operating system software, allowing customers to simultaneously standardise Cisco networking and NVIDIA technology in the data centre.

The proposed collaboration will open new market opportunities for Cisco by unifying the architectural model between front-end and back-end networks, making it easier to manage various enterprise and cloud provider networks.

The expanded partnership aims to give organisations flexibility and choice as they look to meet the demand of AI workloads for high-performance, low-latency, highly power-efficient connectivity within, and between, data centre, clouds, and users.

By enabling interoperability between both companies’ networking architectures, the two companies will prioritise customers’ needs for simplified, full-stack solutions. The goal of the expanded partnership is to allow customers to optimise their AI infrastructure investments with a common architecture, using their existing management tools and processes, spanning front and back-end networks.

“Enterprises are under immense pressure to deploy AI quickly and effectively, and many leaders struggle to justify the investment while balancing the risks,” said Cisco CEO Chuck Robbins. “Together, Cisco and NVIDIA are partnering to remove barriers for customers and ensure they can optimise their infrastructure investments to unlock the power of AI.”

“Advancing at lightspeed, AI will revolutionise every industry,” said NVIDIA CEO Jensen Huang. “NVIDIA Spectrum-X is ethernet enhanced and supercharged for AI. Together with Cisco’s enterprise platforms and global reach, we can help companies worldwide build state-of-the-art NVIDIA infrastructure as they race to transform with AI.”

By making it easier to integrate and standardise on both Cisco and NVIDIA technology, customers will benefit from current and future technology advancements in the NVIDIA Spectrum-X platform, such as adaptive routing, telemetry, congestion control and low latency, as well as Cisco’s broader networking, security, and digital resilience portfolio, including the Splunk data platform. As businesses develop AI capabilities and train models on their own data, having a comprehensive security strategy becomes crucial.

Cisco and NVIDIA are planning to develop joint solutions for customers. Cisco will develop data centre switches with the NVIDIA Spectrum ethernet platform. The open ecosystem approach will provide customers with more choices and flexibility. Organisations can standardise on the NVIDIA Spectrum-X networking platform with both Cisco and NVIDIA switch silicon-based architectures, bringing the industry-leading technologies from both companies under a single management fabric.

Cisco will collaborate with NVIDIA to create and validate NVIDIA cloud partner and enterprise reference architectures based on NVIDIA Spectrum-X with Cisco Silicon One, Hyperfabric, Nexus, UCS Compute, Optics, and other Cisco technologies.

Cisco and NVIDIA are committed to collaboration and joint development, aiming to deliver the highest-performing ethernet solutions for customers. The companies plan for continued collaboration to address significant challenges like congestion management and load balancing so that customers can scale and secure AI deployments.

Cisco plans to upgrade its Silicon switches to be compatible with Spectrum-X and NVIDIA’s reference architectures. These updates will be available in the middle of 2025, including existing and upcoming Cisco Nexus, Cisco Nexus Hyperfabric, and Cisco UCS products.

The timeline for new Cisco Spectrum switches will be announced at a later date.

Share.

Comments are closed.