THR1027 Theater Thu 19 Nov 09:35 PST
Best practices for secure network architecture in the AI age
AKSHAY Agrawal, Mario Lopez
Summary
From the recording
Azure Networking speakers advised isolating AI workloads with no public exposure, using Private Link and inspecting egress. Private endpoint IPv6 is in public preview, with GA said to be coming soon, and over 100 services are supported. New VNets have defaulted to private subnets (no default outbound access) since March. Standard V2 NAT Gateway offers up to 100 Gbps; DDoS protection for Standard V2 public IPs is in public preview. Caveat: no GA date was given for private endpoint IPv6 or for DDoS protection on Standard V2 IPs.
Session description
Microsoft's catalogue description
AI is transforming cybersecurity on both sides of the battle. Threat actors are using AI to automate attacks, while organizations must adopt more resilient defenses. This session explores secure networking architectures for the AI era, including Zero Trust, AI workload protection, segmentation, and multi-cloud security. Learn proven patterns and customer examples for securing AI environments, reducing risk, protecting data, and accelerating AI adoption with confidence.