Choosing a UPS: Balancing availability, reliability and modularity.

Choosing a UPS: Balancing availability, reliability and modularity.

By Jack Ward, MD of Powermode

Thanks to the vagaries of the Eskom power supply, availability concerns are very familiar to South African businesses, particularly those with mission-critical installations, such as a data centre.

Given the high cost of downtime (as much as R 40,000 to R 60,000 per minute), data centre managers are well aware that availability has become the most important metric on which they and their operations are evaluated. This positions an alternative power supply source at the very top of their list of the most essential devices in the data centre.

In this light, the selection of an uninterruptible power supply (UPS) system is driven by the need to totally eliminate any possibility of a power outage and consequent downtime.

When (static) UPS systems first made their appearance on the market years ago they comprised a rectifier, battery and inverter. The reliability of the device depended predominantly on the reliability of the inverter. An inverter failure meant an immediate load crash.

As the development of UPSs ramped up, the static bypass switch was introduced to enable an interruption-free load transfer to the standby power source (batteries or generator) in the event of an inverter failure or overload.

While this (then) new technology substantially improved the overall reliability of the backup power architecture, it paved the way for the introduction of computer-controlled real-time information systems which are today a requirement of the highest reliability (99.99% uptime) UPS configurations.

Of necessity, these systems can no longer rely on legacy single UPS/static bypass systems. So, while the importance of continuous power availability has in no way diminished, it has been joined by another, most important consideration in standby power design reliability: redundancy.

Redundancy comes in a number of forms. The parallel (n+1) redundant system improves availability and reliability and simplifies maintenance of individual UPS modules. (n+1 stands for the number of UPS modules that are required to handle an adequate supply of power for essential connected systems, plus one more.)

Taking the idea a step further, two conjoined units (1+1 redundancy) offers the advantage of additional failover transparency in the event of component failure. This level of resilience is referred to as active/active or ‘hot’ as backup components actively participate with the system during normal operation and are ‘hot swappable’ in the event of failure.

As modularity in terms of UPS design gains traction, so 2+1, 3+1 and 4+1 redundant configurations with modular hot-swappable UPS components are appearing. Driven by the need to minimise MTBF (mean time before failure) values, these parallel redundant chains offer redundancy at module and system level, thus minimising or even negating the impact of any single element failure on the overall reliability of the installation.

As UPS system designs advance – centred on the principals of ‘flexibility’ and ‘scalability’ - so the concept of ‘intelligent paralleling’ is gaining attention. This improves the efficiency of redundant UPS systems by deactivating UPS modules that are not required to support the load, thus taking advantage of the inherent efficiency improvements available in UPS systems under higher (more optimal) loads.

This feature is particularly useful for data centres looking to offset the spiralling costs of electricity by optimising periods of low demand, such as when operating at low capacity on weekends and holidays.

Find us:

  

www.powermode.co.za