Radware has announced Model Context Protocol (MCP) Server Protection, a purpose-built security solution for organizations adopting MCP as a standard way for AI agents to connect to enterprise applications, data, and workflows. The company argues that MCP servers create a new attack surface that traditional web and DDoS protections were not designed to understand, leaving protocol abuse, authentication attacks, and unauthorized interactions potentially unaddressed.
The offering is part of Radware's cloud security platform, joining its Agentic AI Protection and LLM Firewall to cover agents, models, and the servers connecting them. It provides visibility into MCP infrastructure, including deployed servers, exposed tools, and agent activity, along with governance controls to enforce policies as deployments scale. The solution supports public clouds, private clouds, Kubernetes, and hardware-based data centers, and is available as a cloud service or as a cloud-augmented offering for on-premises platforms.
Radware CTO David Aviv said organizations need to know which MCP servers they have, what those servers expose, which agents are accessing them, and whether those interactions are safe. The announcement cites a Stacklok survey indicating that nearly half of software companies use MCP in production, with 19% reporting broad production use, underscoring the urgency of securing this emerging integration layer. The source is a single vendor announcement, so no independent assessment of the product's effectiveness is included.}```json lines? The user asked to return strict JSON with exactly these keys. I provided a JSON object. Need to ensure it's valid JSON. The body contains a period and quotes, but it's fine. The last sentence