Control Plane vs Data Plane

Overview

Peridot uses a dual-plane architecture that separates governance from execution.

The control plane manages policies, routing, and auditability, while the data plane handles AI requests and data processing within your environment.

Why This Architecture Exists

Enterprise AI systems require:

  • Centralized governance

  • Strict data isolation

  • Consistent policy enforcement

  • Auditability across all interactions

A single-plane system cannot meet these requirements without exposing risk.

Control Plane Responsibilities

The control plane (Peridot Cloud) handles:

  • Policy definition and evaluation

  • Model routing decisions

  • Governance configuration

  • Metadata and audit logs

No customer data is processed or stored here.

Data Plane Responsibilities

The data plane (customer environment) handles:

  • AI request execution

  • Model interaction

  • Data processing

  • Integration with internal systems

All sensitive data remains within this layer.

How Requests Flow

  1. Request originates in customer system

  2. Control plane evaluates policies

  3. Routing and enforcement decisions are applied

  4. Request executes in data plane

  5. Logs are recorded

In Production

  • Governance is centralized

  • Data never leaves your environment

  • Systems scale independently

  • All interactions are auditable

Security Implications

  • Eliminates unnecessary data exposure

  • Ensures consistent enforcement

  • Enables compliance with strict data policies

Next Steps


Was this article helpful?