ONUG Triple-T session outlines evaluating AI for enterprise networks

The ONUG Triple-T session brief outlines a framework for adopting AI in enterprise networks using a vendor-agnostic architecture, starting with small pilots, and scaling automation and observability under data governance.

Research Overview

The session is framed around two tracks: networks built for AI and AI used for networking capabilities.

It describes an approach to assess where AI fits in network environments while emphasizing control over data and LLM training.

Key Findings

The brief says the method begins with selecting a foundation that covers GPU, switch, and DPU components as part of a vendor-agnostic ecosystem.

It also states that teams should start small, demonstrate ROI, and then expand automation and observability while maintaining strong data governance.

Technical Breakdown

The networks-for-AI portion identifies a “front-end infrastructure” ecosystem spanning GPUs, switches, and DPUs.

The AI-for-networks portion emphasizes “data governance” as a constraint for how automation and observability scale.

Operational Impact

The brief describes a pilot-first pattern: begin with initial deployments to prove ROI before scaling.

It also positions observability as part of the scaling process under governance controls tied to data and LLM training.

Leadership Perspective

The session is led by Vishal Shukla, CEO of Aviz Networks, whose background includes 20+ years in IT networking at organizations including NVIDIA/Mellanox, IBM, Cisco, and Nortel.

The brief lists experience and output spanning SDN, IoT, cloud orchestration and automation, and performance monitoring, along with authorship on SDN, OpenFlow, and OpenStack.

The ONUG Triple-T session summary focuses on evaluating AI for enterprise networks through a vendor-agnostic foundation, an initial ROI-driven pilot phase, and scaling automation and observability within data governance controls, with a live demo and session guidance for decision-makers; this “Blog Signals brief” is a fact-based summary of the vendor blog.