Skip to content

Control 2.8: Encryption (Data in Transit and at Rest) — Verification & Testing

Control 2.8 requires a manual evidence pack. The tests below deliberately separate evidence that proves an individual negotiated connection from evidence that proves a configured connector, Customer Key deployment, or Microsoft-managed service boundary.

Test cases

Test 1: Microsoft 365 service-encryption review

  • Objective: Record the Microsoft-managed service encryption boundary without inventing a tenant-wide portal setting.
  • Steps:
  • Review the current Microsoft 365 encryption documentation.
  • Review the current technical encryption reference.
  • Obtain the applicable Service Trust Portal report through the organization’s approved process.
  • Record source version/date, workload scope, reviewer, and limitation.
  • Expected result: The evidence pack identifies Microsoft-managed service assurances and clearly distinguishes them from tenant-configurable evidence.
  • Evidence: Documentation review record and approved Service Trust Portal artifact.

Test 2: Actual TLS handshake

  • Objective: Capture the protocol and cipher suite actually negotiated with representative approved Microsoft 365 endpoints.
  • Steps:
  • Run Script 1 from a representative managed client/network location.
  • Retain the raw JSON output, endpoint inventory reference, timestamp, protocol, cipher suite, and certificate data.
  • Verify that each recorded negotiation is TLS 1.2 or TLS 1.3.
  • If TLS 1.3 is not negotiated, record that it is service- and client-dependent rather than declaring it unavailable across Microsoft 365.
  • Expected result: Each sampled endpoint has timestamped TLS 1.2+ negotiated-handshake evidence.
  • Evidence: Raw handshake output and approved endpoint inventory reference.

What this test does not prove

A successful sample does not prove every Microsoft 365 endpoint, client, proxy path, or service-to-service flow. It is a scoped observation that must be paired with Microsoft service documentation and the organization’s endpoint inventory.

Test 3: Exchange connector and legacy SMTP AUTH exception

  • Objective: Review forced-TLS or mutual-TLS settings only where the tenant uses those Exchange mail-flow paths.
  • Steps:
  • Run Script 2.
  • For each relevant connector, review TLS requirement, certificate identity, partner domain, and smart-host configuration.
  • Review AllowLegacyTLSClients only for SMTP AUTH use.
  • If the legacy SMTP AUTH endpoint is enabled, document the exception owner, affected devices, compensating controls, and retirement plan.
  • Expected result: Connector evidence matches approved connector design, and any SMTP AUTH legacy TLS exception is explicitly governed.
  • Evidence: Connector export, approval/change record, and legacy-exception record where applicable.

Test 4: Multi-workload Customer Key DEP, configuration, and onboarding validation

  • Objective: Verify the tenant-level multi-workload Customer Key DEP (MDEP) and its assignment without accepting Exchange-mailbox evidence or a single-subscription design as Copilot coverage.
  • Steps:
  • Run Get-M365DataAtRestEncryptionPolicy and Get-M365DataAtRestEncryptionPolicyAssignment after connecting to Exchange Online. Preserve every returned property and the property names, using Script 3.
  • Confirm both a tenant-level multi-workload DEP policy and its tenant assignment are present and match the approved Copilot scenario. Get-DataEncryptionPolicy is Exchange-mailbox DEP evidence only and cannot satisfy this test.
  • Record two different paid Azure subscription IDs in the Customer Key scenario evidence.
  • Verify one Azure Key Vault (Standard or Premium) vault or Managed HSM instance in each subscription for the scenario; if Azure Key Vault is used for production data, confirm the Premium SKU and HSM-protected keys.
  • For Azure Key Vault, verify 90-day soft-delete retention and purge protection; for Managed HSM, verify purge protection and the configured recovery or soft-delete retention settings.
  • Verify production keys are HSM-protected where that design is required.
  • Run the Customer Key Onboarding Service in Validate mode and retain the property-preserving onboarding request output.
  • Before enabling, confirm ValidationResult is successful. Retain enablement evidence only after approved enablement completes.
  • Expected result: The multi-workload DEP policy and tenant assignment are both present; the onboarding record verifies two unique subscriptions and the correct recovery/key configuration; it reports no unresolved validation failure.
  • Evidence: Multi-workload DEP policy/assignment export, subscription/vault/HSM export, onboarding request output, and change approval. Exchange-mailbox DEP output cannot replace the multi-workload evidence.

Test 5: Sensitivity-label encryption and Copilot behavior

  • Objective: Verify documented EXTRACT behavior and exceptions rather than asserting that encrypted content is categorically inaccessible.
  • Steps:
  • Test a requesting user with VIEW but no EXTRACT. Confirm Copilot does not summarize the encrypted item and can reference it with a link.
  • Test an OWNER user. Confirm that OWNER includes EXTRACT and record the expected outcome.
  • Test an unopened SharePoint/OneDrive item encrypted with user-defined permissions, a direct / reference where supported, and the same item open in an Office app.
  • If Edge DLP is not deployed, test the active encrypted browser-tab exception and record the result.
  • Test each external plugin or Graph connector source separately; do not assume sensitivity labels/encryption applied to external data are recognized by Microsoft Copilot Chat.
  • Test a DKE-protected item separately. Confirm that it is not returned by Copilot/agents and that Copilot cannot be used in the app while the DKE item is open.
  • Expected result: Outcomes match the documented user, source, and surface boundary; deviations have an owner and escalation path.
  • Evidence: Test matrix with user role/rights, source, surface, expected result, actual result, date, and tester.

Test 6: Current HSM validation evidence

  • Objective: Confirm the organization has current evidence for the HSM assurance level it requires.
  • Steps:
  • Record the actual key type and service configuration.
  • For Azure Key Vault Premium or Managed HSM, review the current firmware validation notice.
  • Record the current validation reference and review date; recheck after firmware or certification changes.
  • Expected result: The evidence names the actual HSM-backed key configuration and current validation source, not only a vault SKU.
  • Evidence: Key configuration export and validation-reference review record.

Evidence collection

Evidence item Format Scope / limitation Storage location
Microsoft service-encryption review Review record + approved service-assurance artifact Microsoft-managed service boundary; not a tenant setting Compliance evidence repository
Negotiated TLS handshake JSON/text Representative endpoint/client/network observation Compliance evidence repository
Exchange connector export JSON/text Only configured partner/hybrid/forced-TLS or mutual-TLS mail paths Compliance evidence repository
SMTP AUTH legacy exception record Change/exception record Only smtp-legacy use controlled by AllowLegacyTLSClients Compliance evidence repository
Multi-workload Customer Key DEP/assignment and onboarding state JSON/text Tenant-level Copilot MDEP policy and assignment; Exchange-mailbox DEP output is a separate scope Compliance evidence repository
Azure Key Vault/Managed HSM configuration JSON/text + portal export Subscription, recovery, key type, and role configuration Compliance evidence repository
Sensitivity-label/DKE test matrix Test record User/source/surface-specific behavior Compliance evidence repository

Compliance mapping

Regulation Requirement How this control supports it
GLBA §501(b) Safeguards for customer information Documents encryption evidence and key-management boundaries for Copilot data flows.
NYDFS Part 500 §500.15 Encryption of NPI in transit and at rest based on risk assessment Supports a risk-based review of transport, at-rest, and key-management evidence.
FFIEC IT Examination Handbook Cryptographic controls Supports examination-ready evidence of actual handshakes, key configuration, and exception decisions.