A partitioning strategy divides a system into independent subsystems so one subsystem can fail without causing systemic or unacceptable damage. Each partition must remain manageable while its failure is tolerated for the time and cost required to restore it. section 9.2.2
Apply the strategy by:
- Assigning each subsystem a bounded area of responsibility and allowing it to operate autonomously. section 9.2.2
- Detecting or reporting localized failures to an operating center, then triggering remediation under applicable service-level agreements. section 9.2.2
- Adding failover, replication, or rapid repair when the required coverage or recovery time demands it. section 9.2.2
Mobile phone networks
Each network cell serves subscribers in its allocated area. A failed subscriber database or antenna-steering strategy creates a localized outage that can be detected centrally and remediated without disabling the entire network. Life-support systems may also partition functions, but require a qualified person to operate failed elements. section 9.2.2
