ISO 27001 controls

Control 5.23 treats a cloud service as something you enter, run and eventually leave.

The 2022 edition added a control just for cloud services. It asks you to set out how you pick a provider, how responsibility is split, how you configure and watch the service, and how you would get out. Most companies already do part of this. The gap is usually the written split of duties and a credible exit.

Back to ISO 27001 services

01 · What it asks

A control for the way most companies now run IT

New in 2022, and close to the supplier controls.

  1. 5.23
    Information security for use of cloud servicesSet the rules for acquiring, using, managing and exiting cloud services, matched to your security needs.

A cloud provider is a supplier, so the supplier controls (5.19 to 5.22) still apply. Control 5.23 adds what is particular to cloud: the provider runs part of the stack, the customer runs the rest, and the line between them changes with every service model. The control wants that line written down and the customer's half actually done.

It covers every cloud service in scope, not only the main hosting platform. A single SaaS tool that holds customer records is in scope too.

02 · Lifecycle

Five phases from choosing to leaving

Each phase leaves a record an auditor can read.

  1. Select

    Decide what the service will hold and check the provider can protect it.

    What happens
    • Classify the data the service will hold
    • Review the provider's assurance and data locations
    • Compare against your cloud policy requirements
    You get
    • A short selection record with the decision and its owner
    Good practice
    • Assurance scope checked against the exact service
    Watch for
    • Buying on a card before anyone asks what data goes in
  2. Agree

    Fix the terms and the split of duties.

    What happens
    • Read the provider's shared-responsibility statement
    • Map each duty to provider, you or both
    • Note gaps in standard terms
    You get
    • A responsibility map
    • Signed terms or a recorded risk decision
    Good practice
    • Every customer duty has a named owner
    Watch for
    • Assuming the provider handles backups or logging by default
  3. Onboard

    Set up the service securely before real data arrives.

    What happens
    • Apply single sign-on and multi-factor sign-in
    • Apply the configuration baseline
    • Turn on logging and alerts
    You get
    • A configured tenant that matches the baseline
    Good practice
    • Admin accounts kept separate and few
    Watch for
    • Production data loaded during a trial on default settings
  4. Operate

    Keep the configuration where you set it.

    What happens
    • Check for drift from the baseline
    • Review admin access
    • Track provider notices and changes
    You get
    • Drift reports, access reviews, change notices handled
    Good practice
    • Automated checks where the platform offers them
    Watch for
    • One-off hardening never checked again
  5. Exit

    Be able to leave without losing data or leaving copies behind.

    What happens
    • Know the export format and method
    • Plan the move to an alternative
    • Get deletion confirmed
    You get
    • An exit plan, and on exit a deletion record
    Good practice
    • Export tested once, not just read about
    Watch for
    • Discovering the export limits during a dispute

03 · Shared responsibility

Who does what, by service model

The line moves; your data and access never leave your side.

LayerInfrastructure as a servicePlatform as a serviceSoftware as a service
Data and its classificationYouYouYou
User accounts and access rightsYouYouYou
Application configurationYouYouShared: provider builds, you configure
Application codeYouYouProvider
Runtime, middleware, operating systemYouProviderProvider
Network controls inside the tenantYouSharedProvider
Physical hosts, data centres, hardwareProviderProviderProvider

This is a general pattern. Each provider publishes its own split, and the details vary by service. Use theirs as the source and keep your map beside it.

04 · Baselines

Configuration baselines and watching for drift

Write the settings down once; check them often.

A configuration baseline is the set of settings a service must keep: who can be an administrator, how sign-in works, whether storage is private by default, which logs are on and where they go, how long they are kept. Public benchmarks for the major platforms are a sound starting point. Adjust them to your risk and record why.

  • Identity. Single sign-on, multi-factor sign-in for every human, a small set of break-glass accounts kept apart.
  • Storage. Private by default, encryption on, public sharing an exception with an approval.
  • Logging. Administrative activity logged to a place the service's own admins cannot erase.
  • Network. Management interfaces not open to the internet without a reason that is recorded.
  • Keys and secrets. Who controls encryption keys, and where secrets are stored.

05 · Small vs large

What proportionate looks like

Scale the method, not the intent.

Small company

One platform, a handful of SaaS tools

  • A one-page cloud policy that names approved services and who may buy new ones
  • A responsibility map per main platform, kept beside the provider's own statement
  • The provider's built-in posture checks used as the drift report
  • An exit note per critical service: export method, format, alternative

Large company

Several platforms and many teams

  • Cloud services approved through architecture review before procurement
  • Baselines written as code and enforced at deployment
  • A posture management tool reporting drift to service owners
  • Exit plans for critical services rehearsed, with data portability checked

06 · Evidence

Evidence and the usual gaps

What a reviewer asks to see for 5.23.

  • The cloud policy or the section of the supplier policy that covers cloud
  • A list of cloud services in scope, each with an owner and the data it holds
  • The responsibility map for a sampled service
  • The baseline and a recent drift or posture report, with fixes traced
  • The exit plan for a critical service
  • Only the main platform listed. SaaS tools holding customer data sit inside 5.23 too, and auditors find them in the expense records.
  • Provider report treated as cover. The provider's assurance covers its half. Your half needs its own evidence.
  • Exit plan as a sentence. "We would migrate" is not a plan. Name the format, method and destination.
  • Baseline with no check. A written standard that no report ever compares against is an intention, not a control.

Sources: ISO/IEC 27001:2022 Annex A control name 5.23 and its supplier neighbours 5.19 to 5.22; ISO/IEC 27002:2022 as background (not quoted); ISO/IEC 27017 on information security controls for cloud services; the major cloud providers' published shared-responsibility statements for the general pattern. The table and phases are our own illustration.

Cloud control questions

Does 5.23 apply if we only use SaaS?

Yes. Every cloud service in scope counts.

The control is about using cloud services of any kind. A SaaS-only company still chooses providers, configures tenants, manages access and needs a way out. The provider simply runs more of the stack.

Do we need ISO/IEC 27017 to meet 5.23?

No. It is guidance you may use.

ISO/IEC 27017 is a code of practice for cloud security. It can help you design controls, and some providers are assessed against it. It is not required for an ISO 27001 certificate.

How detailed must the exit plan be?

Enough that someone else could follow it.

For each critical service, record how data is exported, in what format, where it would go, and how you would confirm deletion. Testing an export once makes the plan far more credible.

Is the provider's SOC 2 or ISO certificate our evidence?

For their half only.

It shows the provider runs its controls. The auditor will still look at your configuration, access and logging, because those sit on your side of the line.

Draw the line between you and your cloud provider before an audit does.

An ISO 27001 gap analysis checks your cloud services against 5.23: the responsibility map, baselines and exit plans.