Architecture & deployment

One platform. Three places to run it.

SaaS runs today. Your own cloud account for residency, or on-premises for maximum control, are available by arrangement and scoped per customer. In every model a connector cluster inside your network handles collection with read-only credentials, brokered through CyberArk (opens in a new tab) where available.

Deployment models

Where REAP runs

SaaS · REAP Cloud

  • Available today; quick start and time to value
  • Fully managed, always current
  • AWS (opens in a new tab), U.S. regions; tenant isolation enforced at the application layer
  • Connector cluster (agent + proxy) inside your network
Diagram of the three REAP deployment models: SaaS in REAP Cloud, Customer Cloud, and Customer On-Prem
Deployment models (REAP technical appendix, September 2026) · Open full size

The connector cluster

How data reaches REAP

A connector cluster runs inside your network and is the only thing that touches your devices. It collects configurations, state and telemetry with read-only credentials, brokered through your privileged-access system where available, and opens an outbound TLS connection to REAP, so no inbound firewall rule is needed. You can revoke its access at any time. Ports, encryption and the data classes collected are documented in the security packet, under NDA.

Multi-tenant SaaS

Isolation by design

Authentication and role-based access, multi-tenancy, alerting, scheduling and audit logging are platform services shared by every deployment model. In REAP Cloud, tenants are isolated with access enforced at the application layer; SSO (SAML/OIDC) with your identity provider is supported.

Regulated environment? Ask about deployment in your own account.

Most pilots run in REAP Cloud with read-only access. Where policy requires your own AWS (opens in a new tab), Azure (opens in a new tab) or GCP (opens in a new tab) account, the deployment is scoped with your cloud and identity teams and its timeline agreed before week 0.