Start with risk and service scope
The team is expanding a fictional funds service into a colocation room. The design shows two racks, two network contracts and two cooling units. Before accepting redundancy, relate each component to the service that must continue and the failure it must tolerate. Request evidence of location, paths and shared dependencies, including team access during an emergency. Distance between rooms does not establish isolation from flooding, construction or a common operational error. Assessment should involve facilities, networking, security and business owners. NIST references help formulate requirements; they are neither a facility certificate nor evidence of compliance with a particular organization’s internal rules.
Calculate capacity remaining during maintenance
In the fictional worksheet, each power path supports 80 kW of usable capacity and the stated current load is 110 kW. With one path out of service, you cannot add both capacities. Shedding 25 kW leaves 85 kW, still above the available 80; shedding 35 leaves 75. The second calculation passes only the stated nominal limit. It measures neither transients, margin, runtime, transfer, cooling nor actual capacity confirmed by qualified staff. A UPS may support transition or controlled shutdown; a generator’s presence does not establish that it starts, receives fuel and powers every required dependency. The PM should request criteria and results of an authorized exercise without performing electrical interventions outside their role.
Find common environmental and network dependencies
Units cooler-A and cooler-B serve different rooms, but the worksheet shows a shared pump P and control panel C. Failure of P can affect both rooms even when both units are available. Two telecommunications carriers can also use the same duct and building entry. During review, request physical paths and aggregation points as well as commercial names. Sensitive details need not appear in the general report; retain evidence within the authorized scope. For environmental alarms, define equipment-appropriate limits, representative sensors, recipients and response. A dashboard showing normal temperature does not establish that a notification reaches the on-call team at night.
Integrate personnel safety, access and equipment controls
A contractor arrives at the weekend to replace equipment. Permission to enter the building does not establish authority to open any rack or remove media. The procedure should identify the work, accountable owner, escort when required and records of components entering and leaving. For fire or water leakage, the plan assigns physical actions to qualified personnel and site procedures; an availability objective does not authorize obstructing a life-safety response. Detection, suppression, valves and emergency shutoff require suitable inspection and maintenance. This lesson’s exercise is a document review: no data-center inspection or protective-system test was performed.
Turn the review into a project deliverable
Complete the worksheet below with facilities, networking and APS: service, failure, remaining capacity, dependencies, evidence, owner and acceptance condition. If nominal load fits but transfer has never been exercised, record that distinction rather than declaring proven resilience. If both circuits share the building entry, assign the gap and compare an alternative route with the impact of accepting the risk. Plan out-of-hours maintenance with an interruption decision owner and a post-change service validator. Summary: component count, capacity during failure and independence are different questions. Connect this lesson with service classification, continuity, change management and monitoring.
FICTIONAL WORKSHEET: maintenance and dependencies
Load: 110 kW. One available path: 80 kW.
Option A: shed 25 kW. Calculate remaining load and compare with 80.
Option B: shed 35 kW. Distinguish nominal capacity from physical acceptance.
Rooms A/B: separate units; shared pump P and panel C.
Network: two contracts; shared duct D and entry E.
Deliverable: analyzed failure | affected service | missing evidence | owner | acceptance condition.
Do not perform physical interventions from this worksheet.With a 110 kW load and only 80 available, shedding 25 leaves 85; the proposal still fails the nominal limit.
Common pitfalls
Counting components without paths; adding unavailable capacity; treating tabletop as a technical trial; confusing restoration with service acceptance.
Related topics: Continuity and impact analysis · Change, suppliers and operations
Assess remaining capacity, common paths and physical-protection evidence.
Reference: Security and Privacy Controls for Information Systems and Organizations · CCSP examination outline effective 2026-08-01; January2026 V2 PDF