Cisco pushes security into control plane

Cisco used its Cisco Live conference in Las Vegas on June 2 to introduce Cloud Control, a platform that moves policy, identity and automation closer to the network’s operating core. The company said the system is designed for humans and AI agents to manage, monitor and defend infrastructure together, rather than treating security as a separate dashboard layered on top. Cisco described the product as the foundation of its “AgenticOps” model, which combines networking, security, compute, observability and collaboration in one environment. Network World reported that the change is notable because policy and identity are built into the control path itself, allowing real-time enforcement rather than post-hoc visibility. ### Why is Cisco talking about the control plane instead of another dashboard? Cisco said Cloud Control is a “unified platform” with one login, a single inventory and real-time topology across its product estate. In Cisco’s own description, the platform converges identity, assets, topology, authorization, telemetry and workflow so that humans, applications and agents can reason over the same operating data. (newsroom.cisco.com) Network World wrote that Cloud Control is not a passive management console because policy and identity sit directly in the control path. That means the system is meant to decide and enforce in-line, not merely show operators what happened after the fact. ### What does Cisco mean by humans and AI agents running infrastructure together? Cisco said Cloud Control is built for “humans and AI agents to manage, monitor and defend critical IT infrastructure,” and tied the launch to its broader AgenticOps push. (newsroom.cisco.com) Computer Weekly reported that Cisco is positioning the platform so customers can build applications and agents in natural language, extending the operating model beyond manual administration. (networkworld.com) Cisco’s AI Canvas, which sits inside Cloud Control, is described by the company as a workspace where Cisco data, third-party tools and customer-built agents can take action together. Cisco said Cloud Control Studio also lets customers connect outside tools and custom agents, broadening the system beyond Cisco-only environments. ### How is security being folded into this system? (newsroom.cisco.com) Cisco paired the launch with security products aimed at AI-era threats. The company said the window between vulnerability and exploit has compressed from weeks to minutes, and announced additions including expanded Live Protect coverage and policy controls for agentic environments. (cisco.com) SC Media reported that Cisco’s Cloud Control AI defense suite is intended to let customers create AI agents that monitor systems and block exploitation attempts. That report, citing Cisco executive DJ Sampath, said the company is arguing that human-scale defenses are no longer enough against machine-speed attacks. ### Why would this matter in a warehouse or other physical operation? (newsroom.cisco.com) Cisco’s product pages say Cloud Control spans networking, security, compute, observability and collaboration, and Cisco’s security control layer says customers can “define once” and “enforce anywhere” across Cisco and third-party enforcement points. In a connected warehouse, that architecture would sit near the systems that tie together cameras, badge access, wireless networks, handheld devices and industrial sensors. (scworld.com) That is an inference from Cisco’s stated cross-domain scope and enforcement model, not a separate Cisco case study. Computer Weekly said Cisco is framing the platform around critical infrastructure, where humans and agents operate side by side. In settings where uptime depends on connected operational technology, the practical question becomes how much authority an automated policy engine gets before a human intervenes. ### Where does governance get harder? Cisco said Cloud Control is a secure environment for humans and AI agents, but the same design raises questions about who sets policy, who approves changes and how automated actions are audited. (cisco.com) Cisco’s March security announcement on the “agentic workforce” said Duo, Identity Intelligence and Secure Access add discovery, policy enforcement and intent-aware monitoring for AI agents, indicating the company is already building governance layers around non-human actors. (computerweekly.com) Cisco Live materials said the company is moving agentic AI from concept to production and building “the trust to act at machine speed.” The next test will be customer deployment: Cisco said Cloud Control extends to third-party tools, and Cisco Live sessions and product pages are now the main public record for how those controls will be implemented. (ciscolive.com) (newsroom.cisco.com)