ICT Redundancy, Diverse Pathways & Power Resilience

In a modern hospital the network is a life-safety-adjacent utility: nurse call, telemetry, electronic health records (EHR), barcode medication administration, real-time location, physiologic monitoring, imaging, and increasingly the building's own fire, security, and clinical-engineering systems all ride on the same converged IP fabric. This Article covers how that fabric — and the spaces and circuits that feed it — are made to survive component failures, cable cuts, room losses, and utility outages. It addresses three intertwined disciplines: logical/physical network redundancy, diverse and protected cabling pathways, and resilient power for ICT loads. The data-center spaces, network-core topology, and wireless/DAS coverage that this resilience protects are described in the sibling Article on Data Center, Network Core & Wireless/DAS Coverage; the room/riser/distribution geometry is covered under Telecom/Equipment Rooms, Risers & Distribution Topology. Here the focus is the resilience overlay on top of those.

Why ICT resilience is a clinical-continuity issue, not just an IT issue

Downtime of the network in a hospital is not a productivity inconvenience — it can interrupt care. Loss of the EHR forces clinicians into downtime procedures and paper charting; loss of nurse call or middleware breaks the communication chain between a patient in distress and the responding caregiver; loss of monitoring or telemetry gateways can mean alarms do not reach anyone. Because so many clinical workflows are now network-dependent, owners increasingly treat ICT resilience with the same seriousness historically reserved for normal/emergency electrical power and medical gas.

Several drivers push ICT redundancy into the design program:

Patient-safety exposure. A single fiber cut, switch failure, or telecom-room flood should never sever a clinical unit from the systems that keep patients safe.
Regulatory and accreditation pressure. CMS Conditions of Participation expect hospitals to maintain safe, continuous operations and to have downtime/contingency plans; The Joint Commission (TJC) and DNV survey emergency-management and information-management resilience. While these bodies rarely prescribe specific network topologies, they expect that critical clinical systems keep working — or fail over gracefully — during disruptions.
EMTALA and 24/7 operations. Emergency departments and inpatient units run continuously; there is no "after hours" window in which an outage is harmless.
FGI Guidelines technology provisions. The FGI Guidelines for Design and Construction (Hospital edition) call for information-technology and communications systems to be planned with appropriate reliability and to support the safety and operational needs of the facility, including coordination of ICT spaces with the building's environmental and electrical infrastructure.
Cyber-resilience. Ransomware and other attacks have repeatedly forced hospitals into extended EHR downtime. Physical/electrical resilience does not stop a cyber event, but segmentation, out-of-band management, and well-rehearsed failover are part of the same continuity posture.

The practical consequence: the resilience target for each ICT system should be risk-tiered to its clinical criticality, not applied uniformly. A telemetry gateway or nurse-call server deserves a higher availability target than a digital-signage controller.

Establish availability tiers and a resilience matrix

Good ICT resilience design starts with a deliberate classification of systems and spaces by criticality, agreed among the owner, clinical leadership, IT, and the design team. A common approach borrows the data-center "tier" mindset but applies it system-by-system rather than to a single building rating.

A typical (illustrative) tiering for healthcare ICT:

Tier	Description	Representative systems	Typical resilience posture
Tier 1 — Life-critical	Failure can directly threaten patient safety	Nurse call, code/alarm middleware, physiologic monitoring/telemetry gateways, clinical-alarm notification, fire alarm network links	Redundant power (UPS + generator-backed), redundant network paths, fast/automatic failover, no single point of failure
Tier 2 — Care-critical	Failure disrupts care delivery within minutes	EHR access, PACS/imaging, lab and pharmacy systems, RTLS for clinical workflow, VoIP/clinical communications	Redundant power and dual-homed network; rapid failover; downtime procedures as backstop
Tier 3 — Operationally important	Failure degrades operations but care continues	Access control, CCTV, building automation network, food service, materials management	UPS ride-through, generator backup where practical, single-path acceptable with monitoring
Tier 4 — Convenience/ancillary	Minimal clinical impact	Digital signage, guest Wi-Fi, public AV	Normal power, best-effort, no redundancy mandate

The deliverable is a resilience/redundancy matrix that, for every major ICT system and every telecom space, records the assigned tier, the required power source (normal / UPS / generator branch), the path-diversity requirement, the failover behavior (automatic vs. manual), and the recovery-time and recovery-point expectations. This matrix becomes the contract that the electrical, structured-cabling, and network designs must satisfy, and it anchors the integrated testing and commissioning effort at the end of the project.

Levels of redundancy: from component to geographic

Redundancy is layered. Each layer addresses a different failure mode, and a high-tier system generally needs several layers working together.

Component redundancy. Dual power supplies, redundant fans, hot-swappable line cards and modules within a single device. Cheapest layer; protects against the most common (single-component) failures but not against loss of the whole device or room.
Device redundancy. Two switches, two firewalls, two servers configured so that the failure of one is absorbed by the other — via stacking, clustering, virtual chassis, or active/standby pairs. Edge access switches are often deployed in redundant pairs or stacks serving each telecom room.
Path redundancy. Two physically separate cabling routes between the same endpoints so that a single cut, panel failure, or pathway fire does not isolate a space. This is the structured-cabling expression of resilience and is detailed below.