When responsible for the collection, storage and retrieval of digitised information, whether for one’s own purposes or for those of one or more clients, the ability to ensure that the servers involved remain operational at all times is a mission-critical requirement. Without some means to maintain power, should there be a mains outage, new information can be lost and that which has already been stored can become irreversibly corrupted. To protect against such risks, installing an uninterruptable power supply (UPS) is a vital precaution for any data centre regardless of whether it is running just a couple of networked PCs or an extensive array of servers.
There are a variety of available options and it is important to select one that is best suited to the needs of a given operation. While we have all had concerns regarding the continuity of the mains supply, given the pressure on Eskom to maintain its services despite aging infrastructure and in the face of increasing demand. However, even had there been no need for load shedding and programmed outages, it would not have been possible for the national service provider to guarantee a supply free from occasional power sags and surges. More than a mere inconvenience, power spikes can cause actual damage to IT equipment and so a UPS system will always be important to a data centre, even if the national supplier was able to guarantee continuous delivery.
When selecting a system then, it will be important to choose one that will not only provide the necessary backup power during outages, but that can also protect the IT equipment from the potentially damaging consequences of transient irregularities in the supply, like power sagging and spiking, by smoothing their effects to ensure a clean supply. These are the fundamental requirements of any UPS that is suitable for use in a data centre.
UPSs differ in the manner in which they store energy. This, for instance, may take the form of chemical energy, as is the case with a battery-based system and, to agree parallels that employed in laptop PCs where the battery acts as a backup supply when the unit is disconnect ted from the mains. While, in this case, switching between power sources is a manual action, the type of UPS employed in a data centre operates as a separate unit that can switch sources automatically when necessary.
An alternative stores its energy in the form of angular momentum captured by a flywheel. In the event of an outage, the kinetic energy of the flywheel is used to generate electricity, thus maintaining the supply to essential equipment such as computers. Although a flywheel is generally able to store less energy than a battery, nevertheless, this type of UPS is adequate to protect data centre equipment long enough for it to be safely switched off.
For a battery-based system, two additional components are necessary. A charger/rectifier is required to convert the alternating current from the mains source into direct current in order to charge the battery. Conversely, an inverter will then be needed to reconvert the DC output from the battery to AC whenever the backup power source may be called for.
These systems fall into three categories, the cheapest of which is the standby offline variant. Other options are the line-interactive models which, unlike the entry-level system, is also designed to provide protection from transient sags and spikes in the mains supply, and the double conversion models. The latter type provides the highest degree of protection and is the ideal choice of UPS for use in a dedicated data centre handling sensitive information for multiple clients.
If the integrity of digitised information is your concern, consult our specialists at the PacB Group about a cost-effective protection system to meet your company’s needs.
Our qualified technicians offer support and advice in the selection of the right power solution for your needs by calculating your power requirements.