Building the Agentic AI Overlay Reference Architecture – Draft Framework & Community Input
The Agentic AI Overlay Working Group has been developing a vendor-independent reference architecture that defines how large enterprises can securely and effectively deploy agentic AI systems at scale. In this session, the Working Group will present its draft reference architecture to the ONUG community for open discussion and feedback.
The objective is to ensure that the framework reflects the requirements and priorities of the user community, providing clear guidance to vendors on what is needed for interoperable, non-proprietary Agentic AI infrastructures. Attendees will have the opportunity to shape the architecture by contributing their perspectives on network, security, and data overlays, as well as cross-enterprise use cases.
This is a working session designed to align the ONUG community around a shared, ecosystem-driven approach — one that ensures agentic AI systems can be trusted, scalable, and enterprise-ready.
Nick Lippis is an authority on corporate computer networking. He has designed some for the largest computer networks in the world. He has advised many Global 2000 firms on network strategy, architecture, equipment, services and implementation including Hughes Aerospace, Barclays Bank, Kaiser Permanente, Eastman Kodak Company, Federal Deposit Insurance Corporation (FDIC), Liberty Mutual, Schering-Plough, Sprint, WorldCom, Cisco Systems, Nortel Networks and a wide range of other equipment suppliers and service providers.
Mr. Lippis is uniquely positioned to comment, analyze and observe computer networking industry trends and developments.
Parantap Lahiri was recently the VP of Network and Datacenter Engineering at eBay, based in San Jose, California. With over two decades of experience in network architecture and engineering, he has held leadership positions at major tech companies including eBay, Visa, Juniper Networks, and Microsoft.
Lahiri is recognized for his expertise in data center networks, cloud architecture, and AI infrastructure, including GPU-based systems for AI workloads. He has co-authored award-winning research, holds multiple patents in the field, and is actively involved in developing infrastructure for AI-enabled environments. Lahiri holds a Bachelor’s degree in Electronics and Telecommunication from Jadavpur University and has completed executive education at Wharton.
I focus on what I see as the next frontier of cybersecurity: securing autonomous AI agents and
systems.
I lead ONUG’s Agentic AI Working Groups, where I work alongside industry peers to define the
security standards and guardrails that autonomous AI systems will need as they take on greater
operational autonomy. This work sits at the intersection of emerging risk and practical enterprise
defense, shaping how organizations think about trust, control, accountability and safety in
agentic AI.
I was named an ONUG Fellow in recognition of my contributions to enterprise security, notably
as the creator of the Cloud Security Notification Framework (CSNF).
Prior to focusing entirely on cybersecurity research, I was a security executive at Cigna, where I
led teams responsible for cloud and AI security strategy and execution.
My work in digital forensics complements this focus, grounding the guardrails I help define in the
technical realities of how systems actually fail and how incidents actually unfold.
I’m always glad to connect with fellow researchers, practitioners, and industry groups working on
the security challenges of autonomous systems and digital forensics.
