IT Operations, Service Management and Business Continuity
Domain 5 covers the operational backbone: service management practices, batch and backup operations, and the continuity arrangements that keep commitments to the business when something fails.
Service management essentials
- Incident management restores service; problem management removes root cause; change management controls modification. Major incidents need a review with actions, not just a closed ticket.
- SLAs bind the provider to the customer; OLAs are internal; underpinning contracts bind suppliers. Test that SLAs are measured from the same data the business sees and that credits or remedies actually apply.
- Batch operations need scheduling, dependency checks, failure alerting, restart and checkpoint logic, and reconciliation of output to source.
Backup and recovery
- Full, incremental and differential backups differ in restore chain and window. The 3-2-1 principle (three copies, two media types, one offsite) plus an immutable or offline copy is the ransomware-era baseline.
- Backup job success is not evidence. Restore testing, with measured time and integrity checks, is evidence.
- Backup is not disaster recovery, and neither is business continuity.
Continuity
- The business impact analysis sets RTO (acceptable downtime) and RPO (acceptable data loss) per process; these cascade into system recovery tiers.
- Recovery options include hot, warm and cold sites, reciprocal arrangements and multi-region cloud.
- Test types escalate in rigour: checklist or tabletop, simulation, parallel test, full failover. Every test needs success criteria, measured actuals against RTO/RPO, and logged issues with owners.
Worked example
The BIA gives order processing an RTO of four hours and an RPO of 15 minutes. The annual failover test restored service in nine hours with two hours of lost transactions. The gap is five hours against RTO and one hour forty-five minutes against RPO, so the plan does not meet the business requirement. The finding names the tested actuals, not the plan's stated targets, and requires either remediation or a re-approved BIA that formally accepts the longer outage.
Traps
- Using RTO and RPO interchangeably.
- Accepting backup logs or a supplier's DR certificate as test evidence.
- Reviewing the plan without checking it was updated after the last architecture change.
- Running tests with no success criteria, so nobody can say whether they passed.