As enterprises rapidly adopt generative AI, a new class of application traffic is emerging—AI session flows. AI is not just another workload—it is a new control plane running over your network. These flows, which include interactions between users, large language models (LLMs), and autonomous agents, are fundamentally Layer 7–driven, API-centric, and highly dynamic. Securing them requires a shift in mindset: from inspecting packets to understanding and controlling prompts, responses, and agent interactions.
This session explores how AI workloads both resemble and diverge from traditional application architectures. Like web and microservices traffic, AI interactions rely on HTTP-based APIs and structured exchanges. However, they introduce new complexities: multi-hop, non-linear session paths, multiple identities participating in a single transaction (users, services, and agents), and the use of emerging protocols such as model context protocol (MCP) for tool invocation and agent coordination.
Aidan Walden has served as the Global Director of Public Cloud Architecture and Engineering at Fortinet for the past three years. Prior to Fortinet, he served in various engineering and leadership roles at a tier 1 mobility and global ISP. Throughout his 18-year career, he has navigated the evolving security landscape across mobility, IoT, and service provider infrastructure.
Register now and receive exclusive access to ONUG content and updates
Register Here