Organizations traditionally deployed large, centralized power systems to provide battery backup to their infrastructure, but this approach was often inefficient if admins sized the backups to support future capacity. The emergence of modular uninterruptible power supply setups makes it easier for admins to size backup power infrastructure, increase overall load efficiency and extend hardware lifespan.
The main advantages of a modular UPS system are its ability to grow capacity as needed and reduced maintenance cost. The modules are hot-swappable, and admins can exchange or repair hardware through a vendor. Modular system designs can accept more modules than the required rated capacity, which makes them inherently redundant at much lower cost than a large backup power system.
For example, a three-node N+1 UPS runs each module below 33% load power, and 2N architectures run each UPS at half load. In either case, if one module fails, the remaining modules assume the load.
Modular UPS systems also bring power infrastructure efficiency. A UPS system runs at highest efficiency when near maximum rated capacity; efficiency drops when the load level decreases. These efficiency losses may not seem great on the surface, but they add up over time.
Because these systems run close to capacity, it is easy for admins to configure and reconfigure setups. Organizations that purchase large, traditional UPS systems with future capacity in mind usually end up with systems that run well below capacity. Redundancy always means running below capacity, which also means reduced efficiency. But admins can get higher efficiency levels with an N+1 modular system and careful power management.
Hyperscale data centers and colocation providers use modular UPS technology with multiple architectures to minimize operating costs and keep redundant capacity available.
Known as distributed reserve and block redundant UPSes, these systems can mix different sized modules with unequal loading to provide both required capacity growth and redundancy at minimal capital cost.
They also let admins borrow additional capacity from an idle system until they install new modules, often from warehoused units. This setup means redundancy is only slightly compromised.
As of 2020, vendors offer UPS options that include the following:
More organizations use smaller modules that support 10 kilovolt amperes to 50 kVA. N+1 designs offer lower cost for redundancy and higher loading and efficiency in each module than multiple parallel UPSes, while 2N designs provide accurate load matching. Certain UPS designs enable admins to add modules of different capacities and nonuniformly allocate loads.
But smaller is a relative term. For most organizations, modules are 25, 50 or 100 kilowatts, but mega data centers may use 250 kW or 500 kW units for multi-megawatt systems.
Admins must manage any 2N redundant configuration so that no system load is beyond 50% of its capacity. If capacity is over 50%, the UPS overloads, and the duplicate, load-sharing system fails. As a result, every UPS that runs in 2N mode must operate at less than maximum efficiency.
With careful management, admins can configure a modular UPS more closely to the optimal power load than a larger, fixed capacity system; this may result in some long-term power savings. If the UPS frame is not fully populated, admins must control power so there is always at least one module's worth of unused capacity, or there can be redundancy loss.
Modular UPS system disadvantages depend on how an organization sets up the hardware. Most organizations install the smaller modular systems as additional in-row cabinets. This means added space and weight in the machine room.
Depending on how many equipped cabinet rows and wired distribution circuits admins install, there may be reduced economy of scale because extra capacity in one UPS may not immediately reach another UPS that requires more capacity.
Admins can offset any potential performance issues and move the UPS modules -- if the frame size is adequate -- to reduce power latency.
It's common for most data center batteries to be valve-regulated lead acid (VRLA). This type of battery is used in the majority of UPSes, but it carries certain failure risks and lifetime limitations that can add up in replacement costs. If organizations run a large data center with VRLA batteries, then it is not practical to place smaller UPSes among cabinet rows.
Newer Li-ion batteries offer several advantages over flooded wet cells and VRLA batteries. They are smaller, lighter and can withstand an unlimited number of discharges; this makes them ideal for peak shaving, which can help reduce power costs. With a Li-ion battery, admins should consider centralized UPS offerings.
Li-ion chemistry and packaging are much different for UPSes than consumer-grade electronics, but they can still pose hazards. Fire protection requirements around Li-ion products are much more stringent, which may require organizations to upgrade alarm and suppression systems.
25 Jan 2021