Distributed Redundancy: Maximum Availability at Minimal Cost

Data center power planning is not only about having enough UPS capacity to support the connected load. Facilities also need to consider what happens when equipment is being maintained, capacity is being expanded or part of the critical-power path is temporarily unavailable.

Distributed redundancy is one approach to designing critical-power infrastructure so protected loads can be supported through multiple UPS systems and power paths rather than relying on a single centralized power source.

When properly engineered for the application, a distributed redundant architecture can provide additional flexibility for maintenance, equipment replacement and future expansion while making more effective use of available UPS capacity.

CI Group helps organizations source UPS systems, automatic transfer switches, backup generators and other critical-power equipment for data centers and mission-critical facilities.

What Is Distributed Redundancy?

A distributed redundant power system uses multiple UPS systems and distribution paths to provide the required critical load capacity while maintaining additional capacity that can support the load if equipment or a power path becomes unavailable.

Instead of placing all reserve capacity within one centralized UPS system, redundancy can be distributed across multiple UPS modules or systems serving different portions of the facility.

The exact design varies by data center, but a distributed redundant architecture may include:

• Multiple UPS systems
• Independent power distribution paths
• Shared reserve UPS capacity
• Dual-corded IT equipment
• Static transfer switches for appropriate single-corded loads
• Automatic transfer switches
• Redundant battery configurations
• Multiple generators or generator capacity designed around redundancy requirements

The goal is to create an architecture in which the required load remains supported while allowing appropriate capacity for equipment service, planned maintenance and other operating requirements.

Capacity and Redundancy Are Different

A UPS system can have enough capacity to support the entire connected load and still provide limited redundancy.

For example, if a facility requires 800 kW of protected power and installs a single UPS system capable of supporting that load, the facility has sufficient capacity. The redundancy question is different: what capacity remains available if that UPS needs to be taken out of service?

Redundancy planning evaluates what remains available when a component or power path is unavailable.

That distinction is especially important for facilities where maintenance needs to occur while critical equipment remains operational.

How Distributed Redundancy Can Work

Consider a data center with several UPS systems supporting different sections of the critical load.

Rather than providing a completely duplicated UPS system for every load, the facility may design the architecture so reserve capacity is shared across multiple systems.

If one UPS system or path is unavailable, the remaining systems may be configured to support the affected load within the capacity and switching strategy established by the facility design.

This requires careful coordination between:

• UPS capacity
• Load distribution
• Electrical distribution paths
• Static transfer equipment
• Bypass arrangements
• Battery systems
• Generators
• Automatic transfer switches
• Monitoring and controls

The specific architecture should be engineered around the facility’s load, equipment configuration and availability requirements.

Distributed Redundancy vs. N+1 and 2N

Distributed redundancy is one of several approaches that can be used when designing critical-power infrastructure.

Architecture	General Approach
N	Provides the capacity required to support the intended load.
N+1	Provides the required capacity plus an additional capacity component.
2N	Provides two complete capacity paths capable of supporting the required load.
Distributed Redundancy	Distributes reserve capacity and power paths across multiple systems so redundancy can be shared within the overall architecture.

No single architecture is automatically appropriate for every facility. The design should be matched to the required load, maintenance objectives, physical infrastructure, growth plans and desired level of redundancy.

Maintenance Is an Important Part of Redundancy Planning

One of the practical reasons to build redundancy into a critical-power system is to provide more flexibility when equipment requires planned maintenance.

UPS systems, batteries, capacitors, fans, transfer switches and generators all require inspection and maintenance during their operating life.

A well-planned redundant architecture can make it easier to perform service while maintaining appropriate support for the critical load.

CI Group provides UPS preventative maintenance, battery replacement and other critical-power services that can be incorporated into the facility’s maintenance strategy.

Redundancy Should Extend Beyond the UPS

UPS redundancy is only one part of the critical-power path.

The complete infrastructure may also need to account for:

• Battery systems
• Electrical distribution
• Static transfer switches
• Automatic transfer switches
• Switchgear
• Backup generators
• Fuel systems
• Controls and monitoring

A facility with redundant UPS capacity should also understand what happens if another component in the power path is unavailable.

For that reason, redundancy planning is most useful when the entire electrical path is evaluated as one coordinated system.

Distributed Redundancy and Data Center Expansion

Redundancy planning becomes particularly important when an existing data center is adding new critical loads.

Higher-density computing, AI infrastructure and GPU environments can increase the load placed on UPS systems, generators and electrical distribution equipment.

Before adding capacity, operators should determine:

• How much UPS capacity is currently available?
• How much reserve capacity remains?
• What happens if one UPS system is unavailable?
• Can existing distribution paths support the planned load?
• Does battery runtime remain appropriate at the higher load?
• Can the generator plant support the expanded facility load?
• Do transfer switches and switchgear have sufficient capacity?
• Is additional UPS equipment required?

For additional planning considerations, see Power Infrastructure Requirements for AI Data Centers.

Redundancy for GPU and AI Data Centers

High-density GPU environments can make power-path planning especially important because a large amount of compute capacity may be concentrated within a relatively small physical footprint.

Operators should evaluate whether the supporting power architecture provides the capacity and redundancy required for the planned equipment.

This includes UPS systems as well as backup generation, transfer equipment, batteries and distribution paths.

CI Group discusses these considerations further in Why Redundant Backup Power Matters for NVIDIA GPU Data Centers.

Temporary UPS Equipment Can Support Infrastructure Changes

Redundancy requirements can also change temporarily during equipment replacement or facility upgrades.

For example, a UPS replacement project may remove part of the normal system capacity while the work is being completed.

Temporary UPS equipment can provide another option for maintaining required critical-power capacity during:

• UPS replacement
• Planned maintenance
• Capacity upgrades
• Facility expansion
• Infrastructure transitions

CI Group offers UPS rental systems that can be evaluated as part of a planned equipment transition or temporary critical-power requirement.

Evaluate Existing Equipment Before Expanding the System

An existing redundant design should be reviewed whenever the critical load changes significantly.

Equipment installed for the original facility may still be operating properly but may no longer provide the same reserve capacity after substantial load growth.

A review should consider:

• Current UPS loading
• Available UPS capacity
• Battery condition and runtime
• UPS age and serviceability
• Generator capacity
• Transfer-switch capacity
• Electrical distribution capacity
• Planned future loads

The goal is to understand whether the existing architecture can support both the required operating load and the intended redundancy strategy.

Balancing Redundancy, Capacity and Equipment Investment

Critical-power infrastructure should provide the level of redundancy required by the application without treating maximum duplication as the only possible approach.

Distributed redundancy can offer another way to allocate UPS capacity across a facility while still supporting maintenance and availability objectives.

The appropriate architecture depends on the specific facility, and the design should be evaluated by the engineering team responsible for the electrical system.

For equipment planning, CI Group can help organizations identify available UPS systems, generators and transfer equipment that fit the capacity and electrical requirements established by that design.

Distributed Redundancy Planning Checklist

When evaluating a distributed redundant power architecture, consider:

• What is the current critical load?
• What is the planned future load?
• How much UPS capacity is installed?
• How much reserve capacity is available?
• What happens if one UPS system is unavailable?
• Are there multiple power paths to critical equipment?
• Are single-corded loads appropriately supported?
• Does the battery configuration support the redundancy strategy?
• Is the generator plant appropriately sized?
• Are transfer switches and switchgear coordinated with the design?
• Can equipment be maintained without unnecessarily reducing critical-load support?
• How will future expansion affect reserve capacity?
• Is temporary UPS capacity needed during planned upgrades?

Plan Redundancy Around the Complete Critical-Power System

Distributed redundancy can provide a flexible approach to data center power architecture by combining multiple UPS systems, power paths and shared reserve capacity.

The most effective design is one that provides the required capacity today while also accounting for maintenance, component availability and planned growth.

CI Group supports data centers and other critical facilities with UPS systems, generators, automatic transfer switches, rental equipment and critical-power services across a wide range of capacities and manufacturers.