Cisco launches AI-powered network management that cuts troubleshooting to seconds

Cisco has unveiled new AI-powered networking hardware and software designed to automate network management tasks at “machine speed” through its AgenticOps framework. The announcement, made at Cisco Live 2024 in San Diego, introduces routers, switches, and management tools powered by Cisco’s proprietary Deep Network Model. The move positions the networking giant to capture enterprise demand for autonomous network operations while keeping IT teams in control.

Key hardware launches: Cisco introduced multiple product lines optimized for AI workloads across campus and industrial environments.

• New secure router models (8100, 8200, 8300, 8400, and 8500) deliver three times the throughput of previous versions to handle increased branch office network loads.
• Updated Catalyst campus LAN switches (9350 and 9610) provide up to 51.2 terabit-per-second throughput with sub-5 microsecond latency and quantum-resistant security features.
• The Meraki 9179F Wi-Fi access point and ruggedized industrial switches incorporate Ultra-Reliable Wireless Backhaul (URWB) technology as a fiber-optic alternative.

The Deep Network Model advantage: Cisco’s proprietary large language model serves as the foundation for automated network management capabilities.

• The model was trained on Cisco’s complete knowledge base, including “CCIE-level content to Cisco U. courseware,” creating domain-specific AI expertise tailored to networking challenges.
• It powers the Cisco AI Assistant, a natural language interface that can “identify network issues, diagnose root causes, and automate workflows,” potentially reducing troubleshooting time “from hours to seconds.”
• AI Canvas provides a collaborative interface for security operations, network management, and DevOps teams, with select customer testing planned for fall 2024.

AgenticOps framework: Cisco positions its approach within the emerging AgenticOps methodology, which extends beyond traditional DevOps to include autonomous AI agent management.

• The framework covers the complete lifecycle of AI applications, from initial development through continuous refinement and model retraining.
• Cisco describes it as “orchestration” of AI agents with high enterprise autonomy, similar to offerings from startups like AutoGPT, CrewAI, and LangGraph.
• The system promises to “turn real-time telemetry, automation, and deep domain expertise into intelligent, end-to-end actions — at machine speed and with IT teams still in control.”

In plain English: AgenticOps is like having AI assistants that can automatically write, test, and improve software applications while managing other AI tools in the process. Think of it as a more advanced version of DevOps (where human developers write code and deploy it quickly) but with AI agents doing much of the work autonomously while humans maintain oversight.

Unified management strategy: The new software platform consolidates control across Cisco’s hardware ecosystem, combining Catalyst, Meraki, and industrial switch management in a single interface that leverages AI automation while maintaining human oversight.