![Story image](https://itbrief.com.au/uploads/story/2024/06/04/img-RGINTAHiw7Tu3bofB6Iv3zjX.webp)
Qumulo joins Ultra Ethernet Consortium, partners with Intel, Arista
Qumulo has made history as the first networked storage supplier to join the Ultra Ethernet Consortium, signalling a significant advancement in the realm of IT infrastructure.
In conjunction with this membership, Qumulo has also announced a collaborative effort with Intel Corporation and Arista Networks to push forward the state-of-the-art at the intersection of networking, storage, and data management. This collaboration is set to enhance the performance and operations of Qumulo's Scale Anywhere Data Management platform, spanning from edge computing to data centres and public clouds.
The Ultra Ethernet Consortium is poised to influence the way data moves across networks. Kiran Bhageshpur, Chief Technology Officer at Qumulo, stated, "The Ultra Ethernet Consortium’s work will shape the way data flows across the network, bringing systems, storage, and networks closer together while simplifying architectures with massive performance and reliability improvements." Bhageshpur highlighted the importance of coupling the consortium's forthcoming enhancements with Qumulo's ongoing development efforts with Arista Networks, which operate in numerous real-world production environments.
By leveraging Arista Networks' EOS-based switching and routing systems, Qumulo has successfully deployed over an exabyte of storage across various customers. The robust architecture of Arista’s 7280/7800 series and the proven leaf/spine setup has enabled Qumulo’s customers to deliver primary enterprise storage on a converged network, thereby reducing the costs associated with legacy storage networks.
Brian Balderston, Director of Infrastructure at the San Diego Supercomputer Center, described the impact of this solution at the University of California San Diego. "Our UC San Diego customers require dedicated networks for storage interconnect, based on a standard ethernet infrastructure at up to 200 Gbps of provisioned performance. The Data Science Machine Learning Platform (DSMLP) runs on Qumulo, where thousands of students execute performance-sensitive AI workloads concurrently. These students require optimal configurations to ensure efficiency when running GPUs over thousands of NFS connections," he explained. Balderston noted that the deep buffer leaf-and-spine architecture provided by Arista combined with Qumulo’s Scale Anywhere data management platform allowed them to lower costs and deliver an efficient, unified filesystem for various student organisations.
The collaboration aims to integrate modernised architectures with material product integration to deliver operational value, simplify configuration and troubleshooting, and leverage the strengths of each IT technology effectively. This unified approach aims to yield a consistent primary storage system, applicable across the enterprise for critical business systems, security information and event management (SIEM) systems, backups, AI workloads, and large-scale distributed systems and applications.
Ed Chapman, vice president of business development and strategic alliances at Arista Networks, noted the significance of Qumulo's move to join the Ultra Ethernet Consortium. "As we develop high performance and scale networks for the largest AI pods on open and interoperable IP and Ethernet protocols, the evolution to Ultra Ethernet is to simplify the network and bring compute, AI processing, and storage together," said Chapman. He highlighted that Qumulo's participation in the consortium validates Ethernet and IP as the foundational technologies for the next generation of general-purpose, cloud, and AI computing and storage.
This initiative underscores the companies' ambitions to innovate and reshape the IT infrastructure landscape, making technologies more efficient and accessible for enterprises aiming to handle vast amounts of data effectively. The collaboration between Qumulo, Intel, and Arista Networks is anticipated to drive substantial improvements in the performance and reliability of data management systems, offering robust solutions for contemporary and future storage and networking needs in both public and private sectors.