How to Build Networking Fabrics at AI Scale – Arrcus Triple-T
AI workloads are inherently distributed, requiring unprecedented levels of network performance and reliability. This session will explore the unique characteristics of end-to-end networking fabrics designed to support distributed AI at scale. We will examine how to build high-performance, lossless, and congestion-free Ethernet networks that accelerate both training and inferencing. Emphasis will be placed on the role of open networking architectures, enhanced visibility for real-time monitoring and optimization, and intelligent automation to ensure agility and operational efficiency. The audience will gain practical insights into fabric approaches that unlock the full potential of AI by delivering scalable, resilient, and future-ready network infrastructure.
Sanjay Kumar has over 20 years of experience in the networking industry at companies like Cisco Systems, Broadcom, Aruba Networks (HPE), and IP Infusion. He has held leadership roles in Product Management, Product/Corporate Marketing, and has a proven track record of leading product development, creating and driving go-to-market strategies and developing compelling messaging for multi-billion dollar product lines across Enterprise and Service Provider segments.
Prior to joining Arrcus, he was CMO at IP Infusion, where he successfully created a new brand identity and led the overall global marketing and Strategic Alliances strategy.
