Cisco partners with VAST Data for simplified AI infrastructure

Cisco has selected VAST Data as the first data platform for its newly unveiled Cisco Nexus HyperFabric, a Generative AI infrastructure system that has been developed in collaboration with NVIDIA.

The partnership aims to provide enterprises with a simplified, high-performance AI infrastructure combined with an Ethernet-based AI fabric that can manage enormous amounts of data at an exabyte scale.

In an announcement in Sydney, VAST Data revealed the collaboration, which is intended to benefit companies building AI workloads.

The AI infrastructure will facilitate high-speed access to massive datasets, a critical requirement for AI applications. According to Cisco Executive Vice President and General Manager of Cisco Networking, Jonathan Davidson, this ecosystem approach will assist enterprise customers in managing the complexities of networking, data, and security challenges when constructing AI infrastructures at any scale.

The VAST Data Platform, which offers unified storage, database, and data-driven function engine services built specifically for AI, becomes the first data platform optimised and supported with Cisco Nexus HyperFabric.This enables Cisco’s cloud-management solution for data centre networking to orchestrate, deploy, and manage the stack, integrating VAST APIs and incorporating network and storage telemetry.

As a result, Cisco customers will gain the ability to monitor and correlate performance and latency to optimise complex applications.

The Cisco Nexus HyperFabric AI Cluster solution aims to help enterprises build AI data centres using NVIDIA’s accelerated computing resources, including NVIDIA Tensor Core GPUs, NVIDIA BlueField-3 DPUs and SuperNICs, and NVIDIA AI Enterprise software, combined with Cisco networking and the VAST Data Platform. This end-to-end solution provides visibility across compute, networking, storage, and data management, enabling organisations to construct and scale their AI operations seamlessly.

The collaborative integration also extends to VAST’s DASE (Disaggregated Shared Everything) architecture, powered by NVIDIA BlueField DPUs, which further enhances the performance and scalability of AI applications.

Generative AI demands unprecedented scale and performance to ensure rapid access to data, driving a significant investment in Ethernet networking technologies. According to IDC, the Ethernet switching market grew by 20.1 per cent in 2023 to USD $44.2 billion, largely due to the increasing number of enterprise AI deployments.

Renen Hallak, CEO and co-founder of VAST Data, emphasised the importance of this collaboration for enterprise adoption of AI technologies. He noted that enterprises have traditionally been reluctant to adopt new systems due to integration complexities. However, this partnership with Cisco and NVIDIA simplifies the process, making it more accessible for enterprises moving from proof of concept to production stages.

Jonathan Davidson reiterated the accelerating business innovation spurred by AI applications and the necessity for efficient, high-speed data access.

By including VAST Data and NVIDIA in its ecosystem, Cisco aims to support its enterprise customers in overcoming their most challenging networking, data, and security issues as they build scalable AI infrastructures.

The VAST Data Platform has also been certified with Cisco Nexus Ethernet-based switches, delivering validated designs that ensure seamless deployment and the highest levels of support for joint customers.