DriveNets to Showcase AI Ethernet at AI Infra Summit

ByRicardo
  • Learn why fabric-scheduled Ethernet delivers the very best efficiency AI back-end cloth answer
  • Explore DriveNets Network Cloud-AI – the highest-performance back-end cloth for any AI utility

 DriveNets, a pacesetter in cloud-native networking options, will take part within the AI Infra Summit, happening September 9/11, 2025 in Santa Clara, CA. DriveNets will showcase Network Cloud-AI, its Ethernet-based AI networking answer, and also will take part in an knowledgeable panel session on community expertise developments.

Attendees can cease by sales space 702 to discover DriveNets’ Network Cloud-AI – the market’s highest efficiency Ethernet-based AI networking answer. Highlights of DriveNets Network Cloud-AI embody:

  • Proven to ship InfiniBand-level community efficiency
  • Fastest time to first token – achieved with simplified end-to-end deployment and with minimal fine-tuning wanted to optimize workload efficiency
  • Delivers converged AI networking answer in a single cloth – AI frontend, backend, and storage networking
  • Supports any GPU, any optics, and any NIC

DriveNets may even be a part of a panel of specialists to talk about the subject, “Transformative Advancements in Scale-up, Scale-out, Scale-in, and Scale-across Networks.” The panel will happen on September tenth at 2:40 pm and can characteristic:

  • Dudy Cohen, VP Product Marketing, DriveNets
  • Ashwin Gumaste, Principal Architect Azure Network Hardware, Microsoft
  • David Lazovsky, Co-Founder & CEO, Celestial AI
  • Charlie Fu, Technical Director, High-Speed Optical Interconnect, Amphenol
  • Moderator – Kurtis Bowman, Director of Architecture & Strategy, AMD

In this panel, Dudy Cohen will clarify how fabric-scheduled Ethernet delivers the very best efficiency AI back-end cloth answer.

Throughout the occasion, DriveNets will discover frequent questions and optimum options for organizations who’re constructing massive AI clusters, together with:

  • What is the highest-performance back-end cloth answer?
    By combining the scalability of normal Ethernet with the high-performance of scheduled-fabric, DriveNets Network Cloud-AI delivers unmatched networking efficiency at any scale and with an optimum price construction. Featuring low-packet-loss community architectures. Built on DriveNets’ Fabric-Scheduled Ethernet (FSE) structure, the answer ensures predictable, lossless backend connectivity, delivering important enchancment in job completion instances (JCT) for AI workloads.
  • How does DriveNets ship the quickest time to first token? 
    DriveNets’ answer relies on fabric-scheduled Ethernet that’s primarily based on a predictable, lossless cloth and permits minimal deployment fine-tuning to guarantee optimized AI workload efficiency.

    In addition, DriveNets Deployment Infrastructure Services (DIS) helps AI cluster structure design and end-to-end bodily installations that speed up deployment time whereas maximizing infrastructure efficiency.

  • What is a unified AI networking answer?
    A unified networking answer, like DriveNets Network Cloud-AI, helps AI back-end cloth, front-end cloth, and storage networking, multi functional high-performance networking answer.

    Backend networking in AI clusters refers to the interconnect infrastructure that facilitates inner communication between AI GPUs inside a knowledge middle.

    Frontend networking in AI/HPC clusters refers to the community infrastructure that manages exterior information site visitors between AI workloads and customers, purposes, or different companies. It connects the AI cluster to the broader information middle, cloud companies, or enterprise methods. Frontend networking should present excessive bandwidth, low latency, and safe connectivity to guarantee seamless interplay between AI fashions and end-users or enterprise purposes.

    Storage networking in AI clusters is chargeable for dealing with the huge information switch between the AI compute nodes and exterior storage methods. Unlike typical HPC implementations, this can be a crucial infrastructure for AI workloads, as this site visitors is intense, and inadequate efficiency of the storage cloth will lead to poor general workload efficiency.

    DriveNets Network Cloud-AI gives a unified answer for each networking materials, sharing the identical Ethernet-based expertise, structure and precise implementation with the back-end cloth.

The put up DriveNets to Showcase AI Ethernet at AI Infra Summit first appeared on AI-Tech Park.

Similar Posts