Information Systems Operations and Business Resilience
What this domain tests
Domain 4 covers day-to-day delivery of IT services and the organisation's ability to keep them running through disruption: service management, change and configuration control, scheduling, backup, capacity, and disaster recovery/business continuity planning.
Service and change management
Incident management restores service quickly; problem management finds and removes the root cause. Change management requires a request for change, risk and impact assessment, approval (CAB), a backout plan, and post-implementation review - and emergency changes still need retrospective approval and documentation. Configuration management, supported by a CMDB, is what makes impact analysis possible. Developers should not hold production access, and production job schedules should be controlled by operations rather than the application owner.
Backup
Full plus incremental: the incremental resets the archive bit, so restoration needs the last full plus every subsequent incremental. Full plus differential: restoration needs the last full plus the latest differential. Backups must be encrypted, stored offsite or immutably, and proven by test restoration. RAID protects availability against disk failure only - never against deletion, corruption, ransomware or site loss, so it never replaces backup.
Resilience
A business impact analysis establishes the maximum tolerable downtime, from which the RTO derives; the RPO sets the maximum tolerable data loss and therefore backup or replication frequency. Alternate sites run from hot (near-immediate, fully equipped) through warm (partial, hours to days) and cold (shell, weeks), plus mirrored and mobile options. Test rigour increases from checklist review to tabletop walkthrough, simulation, parallel test and full interruption.
Worked example
A trading platform has an RPO of five minutes and an RTO of two hours. Nightly tape backup misses the RPO by up to 24 hours of trades. Fitting controls are continuous disk mirroring or journaling to a remote site, plus a hot site with pre-staged configuration and replication to meet two hours. A cold site is acceptable only if the BIA-supported MTD is measured in weeks.
Traps
Accepting RAID as a backup substitute; approving a DR plan that has never been tested; treating 'the plan exists' as evidence of recoverability; and forgetting that DR restores IT, while BCP sustains the whole business.