SIAM practices and processes across the service lifecycle
Domain 4 tests the SIAM process model: which practices the integrator owns, which the providers own, and how they interface across the lifecycle.
Grouping the practices
The BoK groups practices across the lifecycle and adds specifically SIAM practices:
- Strategy and governance: service strategy, demand, portfolio, service governance, financial management and service costing.
- Design and transition: service design, change, release and deployment, knowledge management, service asset and configuration management - an integrated CMDB is the backbone of multi-provider SIAM.
- Operation: service desk as single point of contact, incident and major incident, problem, service request, event, access.
- SIAM integration practices: service level management (end-to-end SLAs), service performance management and reporting, supplier and contract management, service improvement.
The design rules
- One process, one owner, one instance, one tool wherever practical; duplicated provider processes destroy end-to-end control.
- The integrator typically owns the cross-provider process instance and end-to-end measurement; providers own execution inside their tower and their own component-level records.
- Interfaces are documented - inputs, outputs, escalation, timescales, data - so every hand-off is a defined transaction.
- Agreement hierarchy: business-facing SLA for the service, OLAs between the integrator and internal providers or teams, underpinning contracts with external suppliers. Only the integrator can assemble these into one end-to-end commitment.
Worked example
A card payment fails; provider A runs the application, provider B the network. The integrator's service desk logs one ticket in the shared tool, declares major incident, runs the bridge, tasks A and B, owns business communication and keeps the end-to-end SLA clock. A finds the root cause in its component; afterwards the integrator opens one cross-provider problem record with joint root cause analysis and adds the fix to the improvement backlog. Reasoning: each provider sees only its own segment, so integration - not escalation - restores service and prevents recurrence.
Traps that cost marks
- Assuming the integrator owns every process; providers still own internal execution.
- Letting each provider run its own incident process and tool.
- Confusing SLA, OLA and underpinning contract.
- Omitting cross-provider problem management and knowledge management, which prevent repeat failures.
- Treating the CMDB as a provider deliverable rather than a jointly maintained integrated asset.