PingOne Privilege

Key concepts

Learn about the foundational concepts and terms for PingOne Privilege.

Core concepts

Privileged access management (PAM)

PAM is a set of principles and practices to control, monitor, and secure access to critical resources by both human and machine identities. PAM ensures that users have the right-sized permissions to access critical resources, such as infrastructures, applications, or data essential to an organization’s operations, and that could cause significant harm if compromised or unavailable.

Just-in-time access

PingOne Privilege replaces long-lived, standing privileges with just-in-time (JIT) access: temporary permissions that are granted when needed, expire automatically, and are revoked when they expire. Users request access through the self-service portal, or policies grant it, and every grant is time-bound and auditable.

Trusted Platform Module (TPM) and Secure Enclave

Modern Windows and Mac laptops include hardware-based security features like the TPM chip or Apple’s Secure Enclave. These provide a secure, tamper-resistant location for storing private keys, offering stronger security than file-based credentials because the keys can’t be directly read or extracted.

Deployment models

PingOne Privilege offers Endpoint Controller-based and No Endpoint Controller-based deployment models. In the Endpoint Controller-based model, users install the Endpoint Controller on their device for hardware-backed identity assurance. In the No Endpoint Controller-based model, users authenticate through PingOne SSO without installing the Endpoint Controller.

Endpoint Controller

The Endpoint Controller, installed during setup, creates, and stores a non-exportable private key in the device’s TPM or Secure Enclave. The Endpoint Controller uses this key to establish a secure mutual TLS (mTLS) connection with PingOne Privilege, which verifies the identity of both the user and the device.

The Endpoint Controller also serves Agent Privilege by hosting the sensors that report personal agent discovery findings from managed devices.

No Endpoint Controller

For situations where the Endpoint Controller can’t be installed, the No Endpoint Controller provides secure access to resources like SSH, cloud CLIs, and Kubernetes. When a user authenticates through the No Endpoint Controller, they’re redirected to PingOne SSO for authentication. After successful authentication, the user can access resources based on their permissions.

PingOne Privilege products

PingOne Privilege includes two complementary products with their own unique concepts.

Human Privilege concepts

Gateways

Gateways are entry points for accessing resources on a private network. The Endpoint Controller connects to a private proxy to establish a secure session, and the Endpoint Controller determines the optimal gateway path between the user and the resource. Relays extend gateways to networks that don’t allow inbound connections.

Learn more in Configuring gateways.

Access protocols

Human Privilege secures access to resources through industry-standard protocols: certificate-based SSH for Linux servers, with short-lived certificates issued per session, and RDP for Windows servers, with sessions managed and audited through PingOne Privilege.

Learn more in Access protocols.

Agent Privilege concepts

AI gateway

The AI gateway is the enforcement point for agent traffic. Agent credentials and tokens are validated and terminate at the gateway and are never forwarded to backend services. At each request, the gateway injects the credential needed for the resource being accessed. The gateway hosts MCP Gateway, LLM Gateway and AI Guard functionality. Learn more in AI gateways.

Agentic Apps

Agentic Apps are the applications agents reach through PingOne Privilege: MCP Server, MCP Aggregate, OpenAPI MCP, and Remote Agent applications, integrations from the MCP Catalog, and Git Servers. Each application is associated with an AI gateway that enforces policy on the agent’s traffic.

Learn more in Agentic apps.

AI Guard

AI Guard inspects agent traffic across MCP, A2A, and LLM connections for threats such as prompt injection, data exfiltration, and tool abuse, and enforces allow, alert, sanitize, or block responses for each detection.

Learn more in Agent registry

AI discovery

AI discovery builds an inventory of the AI in your organization. Personal agent discovery observes the agents users run on their devices and browsers, while workload agent discovery inventories the agents in your cloud platforms.

Learn more in Workload agents and Personal agents