Cisco Systems (CSCO) Cisco Live! Day 2 summary

Transforming data center security for AI scale

• Enterprises face increased complexity, security demands, and AI-driven disruption, requiring a shift from traditional centralized architectures to distributed, security-infused networks.

• Cisco's approach integrates security directly into the network fabric using smart switches with both NPUs (Silicon One) and DPUs (AMD), enabling asynchronous service delivery and stateful segmentation at high throughput.

• Hypershield, the first service on these smart switches, leverages eBPF agents and a hierarchical data graph to provide end-to-end application visibility, automated policy recommendations, and distributed enforcement.

• Hybrid mesh firewall unifies policy management across Cisco Secure Firewalls, third-party firewalls, microsegmentation, and Hypershield, all managed from Security Cloud Control and integrated with Splunk for comprehensive analytics.

• Live Protect, based on Isovalent Runtime Security, enables runtime vulnerability protection on Cisco devices, reducing emergency patching needs and enhancing operational resilience.

Customer perspectives and operational impact

• Customers face challenges in simplifying security, managing complexity, and adapting to AI-driven workloads, with WWT highlighting the need for embedded security and operational efficiency.

• WWT's labs and AI proving ground help clients map out technology selection, migration steps, and security strategies, including Hypershield deployment and AI security enclaves.

• The shift from perimeter firewalls to distributed enforcement is critical for handling AI-scale data flows and agent-to-agent communication, making embedded security essential.

• AI is used both to secure infrastructure and as a target for security, with Cisco AI Defense providing guardrails for models and AI Canvas offering real-time operational insights.

• Sustainability is addressed through power-efficient designs, liquid cooling, and consolidation of security functions, freeing up resources for AI deployments.

Simplifying segmentation and future-proofing

• Hypershield's autonomous segmentation uses AI agents to learn application behaviors, recommend policies, and adapt to rapid changes, reducing manual rule creation and operational risk.

• Co-development with customers led to innovations like Tetragon network policy, drastically reducing segmentation rule counts and improving manageability.

• All new Cisco devices are smart devices, supporting future AI agents and dynamic expansion as AI use cases grow.