EDGE REGISTER

The Organizational Edge Is No Longer Only IP Space · Six Edge Types · Defend What’s Actually Reachable


Document ID CERG-GOV-EDG-001
Version 1.0
Status Approved
Classification Public
Owner Vendor Risk Analyst / Governance Pillar Leader
Parent Policy CERG-POL-001 - Cybersecurity Policy
Supporting Documents CERG-PRC-TPRM-001 · CERG-STD-AC-001 · CERG-STD-IT-001 · CERG-STD-NET-001 · CERG-STD-SDL-001 · CERG-GOV-FLOW-001
Review Cycle Annual / On material change to vendor population, SaaS portfolio, or network architecture
Frameworks NIST CSF 2.0 (IDENTIFY, GOVERN) · NIST 800-53r5 (CA, SA, SR) · NIST 800-161r1
Regulations NERC-CIP CIP-013 · CMMC L2 · SOX ITGC
Environments All organizational edges — network, identity, SaaS, vendor, software supply, data

Table of Contents

  1. The Edge — What Changed
  2. The Six Edge Types
  3. Edge Management Model
  4. Low-Control / High-Dependency Operating Model
  5. Edge Register Operations
  6. Document Control

1. The Edge — What Changed

For twenty years, “the organizational edge” meant the firewall. You knew what was inside and what was outside, and security’s job was to harden the perimeter.

That edge no longer exists in any useful form. Organizations operate across SaaS tenants, identity providers, vendor admin paths, MSP tooling, CI/CD dependencies, managed detection providers, cloud control planes, APIs, domains, certificates, subprocessors, and data egress paths. Each of these is a boundary across which trust, data, privilege, or control flows. Each of these is an edge.

The organizational edge is not a network diagram. It is the set of boundaries across which the organization’s security posture can be changed by someone outside the organization.

CERG maintains this Edge Register as the authoritative inventory of organizational edges. It is not a replacement for the TPRM procedure — it is the conceptual model that elevates third-party and supply-chain risk from “a procedure” to a first-class architectural concern. The TPRM procedure (CERG-PRC-TPRM-001) is the operational document that implements this model.

A program that truly knows its edge is less impressed by AI-powered attack branding. The edge doesn’t care whether the attacker used AI. It cares whether the path existed.


2. The Six Edge Types

2.1 Network Edge

The traditional perimeter — and still relevant. What is Internet-reachable? What VPNs, APIs, and remote-access paths exist?

Attribute Description
Examples Internet-facing hosts, VPN concentrators, public APIs, cloud load balancers, exposed management interfaces
Primary owner Network Engineering / Cloud Engineering
CERG role Risk validates external attack surface continuously; Engineering reviews architecture for new Internet-facing services
Evidence EASM export, DNS/cloud reconciliation, firewall rule review
Key question “If we didn’t know this was here, could an attacker find it?”

2.2 Identity Edge

The most consequential edge in modern environments. Who can authenticate? What can they reach after authentication? What federated trust relationships exist?

Attribute Description
Examples IdP (Entra ID, Okta), federation trusts, OAuth applications, privileged groups, service principals, API keys
Primary owner Identity Engineering
CERG role Risk monitors privileged access; Governance tracks recertification; Engineering reviews federation and OAuth grants
Evidence IdP logs, privileged group membership diff, OAuth application inventory, service principal audit
Key question “If an attacker has a valid credential, what can they reach before we detect them?”

2.3 SaaS Edge

Every SaaS tenant is an organizational edge. The organization’s data, identities, and configurations live inside a provider’s infrastructure.

Attribute Description
Examples M365 tenant, Salesforce org, Workday instance, ServiceNow, GitHub org, Slack workspace
Primary owner Business application owners + IT
CERG role Governance maintains Approved Provider Catalog; Risk runs SSPM; Engineering reviews tenant configurations
Evidence SSPM report, admin setting export, OAuth grant review, shared responsibility matrix per provider
Key question “What data lives in this tenant, who can access it, and what happens if the provider is compromised?”

2.4 Vendor Edge

Vendors and service providers with privileged access, remote connectivity, or administrative paths into organizational systems.

Attribute Description
Examples MSP tooling, remote support accounts, OT vendor maintenance access, managed detection providers, contractor admin paths
Primary owner Vendor Risk Analyst
CERG role Risk assesses and tiers vendors; Governance tracks contract clauses and evidence; Engineering validates access controls
Evidence PAM session logs, vendor access review, contract clause register, kill-switch test records
Key question “If this vendor is compromised, what can the attacker reach, and how fast can we cut them off?”

2.5 Software Supply Edge

Every dependency — open-source library, commercial SDK, CI/CD pipeline, build system — is a path through which malicious code can enter the organization.

Attribute Description
Examples npm/PyPI/Maven packages, Docker base images, CI/CD pipeline integrations, build system access, signed releases
Primary owner Engineering / Application Security
CERG role Risk runs SCA; Engineering maintains SBOM; Governance tracks integrity requirements
Evidence SBOM (NTIA minimum elements), CVE advisory subscription, signed release verification, pipeline access review
Key question “If a popular open-source package we depend on is compromised tonight, do we know within hours?”

2.6 Data Edge

Every path through which organizational data leaves the organization’s control — subprocessors, external data shares, outbound integrations, backup targets.

Attribute Description
Examples External data shares (SharePoint, Box), subprocessor data flows, outbound API integrations, backup replication targets, analytics pipelines
Primary owner Data Governance / Business application owners
CERG role Governance tracks subprocessor inventory; Risk assesses data egress paths; Engineering reviews integration architecture
Evidence Subprocessor register, data flow diagrams, DLP policy coverage, external share audit
Key question “Where does our data live outside our control, and who can see it?”

3. Edge Management Model

Every edge is managed through a common lifecycle:

Phase Activity Owner
Discover Identify and inventory the edge. What exists? Risk / Engineering
Assess Evaluate the risk. What’s the blast radius if compromised? Risk
Control Implement controls. What reduces the exposure to acceptable levels? Engineering
Monitor Continuously validate. Has anything changed? Risk
Respond React to compromise. How fast can we cut access? Engineering / Risk / SCCT

Each edge type has its own discovery cadence, assessment criteria, control catalog, and monitoring signals. The table below summarizes the primary tooling and evidence for each type:

Edge Type Discovery Assessment Control Monitoring
Network EASM, external scan, cloud inventory CVSS + exposure + asset criticality Firewall, WAF, ACL, segmentation Continuous external scan, drift detection
Identity IdP export, privileged group review Access review, privilege analysis PAM, JIT, phishing-resistant MFA IdP logs, anomalous access detection
SaaS SSPM, tenant inventory, OAuth review Configuration assessment, data classification Tenant configuration baseline, OAuth app approval SSPM continuous monitoring, drift alerting
Vendor Vendor register, contract review Tier-based assessment, evidence collection PAM, session recording, kill-switch Access review, SOC 2 / attestation monitoring
Software supply SBOM, pipeline inventory, dependency scan CVE matching, exploitability analysis Signed releases, pipeline access control, dependency pinning SCA continuous monitoring, advisory feeds
Data DLP discovery, share audit, integration map Data classification, egress path analysis DLP policies, encryption, access controls DLP alerts, external share review, subprocessor change alerts

4. Low-Control / High-Dependency Operating Model

Not all edges can be controlled. Some can only be influenced, contracted, or monitored. CERG recognizes four control levels:

Control Level What the Org Can Do Examples
Direct Control Configure, monitor, patch, test Owned servers, self-managed cloud workloads, on-premise network
Shared Control Configure tenant, require evidence, validate logs SaaS tenants (M365, Salesforce), IaaS/PaaS (AWS, Azure)
Contractual Control Negotiate clauses, notification, audit rights MSPs, managed security providers, subprocessors
Opaque Dependency Track residual risk, exit plan, kill switch, compensating monitoring SaaS vendor’s underlying infrastructure, open-source library maintainers, upstream cloud provider outages

Operating at each level:

  • Direct and Shared control environments follow standard CERG controls — baselines, scanning, change management, evidence collection.
  • Contractual control environments require: security clauses in contracts, annual evidence collection (SOC 2, ISO 27001, or equivalent), incident notification commitments, and a tested kill-switch procedure.
  • Opaque dependency environments require: documented residual risk acceptance, an exit/migration plan, compensating monitoring (e.g., CSPM/SSPM for what IS visible), and a documented assumption of what the dependency provides. CERG does not pretend to control what it cannot control.

5. Edge Register Operations

5.1 Register Maintenance

The Edge Register is maintained as a living document. The Vendor Risk Analyst (or delegate) reviews each edge type quarterly:

  • New edges added within 30 days of discovery
  • Edges retired when the relationship, service, or path no longer exists
  • Control level re-assessed annually

5.2 Integration with CERG Procedures

Procedure Edge Type Integration
TPRM-001 Vendor, SaaS, Software supply Primary operational procedure for third-party edges
VM-001 Network, Software supply Exposure management feeds edge assessment
AC-002 Identity, Vendor Access reviews and privileged access management
AR-001 All New services trigger edge registration during architecture review
IR Plan All SCCT activation when a vendor edge is compromised

5.3 The Principle

A program that knows its edge is less impressed by attack branding.

AI-powered attacks, supply-chain campaigns, and nation-state actors all exploit the same thing: an edge the organization didn’t know it had, or didn’t manage. The edge doesn’t care about the attacker’s tooling. It cares whether the path existed.


6. Document Control

Revision History

Version Date Author Change Summary
1.0 2026-06 CERG Governance Initial release. Six edge types, edge management model, low-control operating model.

Review Triggers

Annual review or upon: material change to vendor population, SaaS portfolio, network architecture, or regulatory scope.

Document ID Relationship
TPRM Procedure CERG-PRC-TPRM-001 Operational implementation of vendor, SaaS, and software supply edges
Access Management Standard CERG-STD-AC-001 Identity edge controls
IT/Cloud/SaaS Security Standard CERG-STD-IT-001 SaaS and cloud edge controls
Network Security Standard CERG-STD-NET-001 Network edge controls
Secure Development Standard CERG-STD-SDL-001 Software supply edge controls
Cross-Pillar Operational Flows CERG-GOV-FLOW-001 SCCT activation and edge compromise response
Exposure Management Procedure CERG-PRC-VM-001 Exposure discovery feeds network and software supply edge assessment

Know your edge. Manage it deliberately. Never be surprised.


CERG-GOV-EDG-001 · Version 1.0 · PUBLIC


Source: governance/CERG-GOV-EDG-001_Edge_Register.md · Download .md · View on GitHub