DCN December 2016 - Page 22

modular systems BALANCING ACT Wendy Torell of Schneider Electric discusses how N+1 UPS designs can help find a balance between redundancy, cost and speed of deployment. T he primary purpose of any uninterruptible power supply (UPS) is to ensure back up power to the equipment in a data centre in the event that the electrical or mains supply becomes compromised. But how do you ensure the UPS itself doesn’t lead to downtime? The answer lies in the utilisation of redundant systems. For highly critical loads, the deployment of a dual-path architecture, in which two separate power paths, each supported by a separate UPS, is commonplace. While this offers the best protection in that there is a fully redundant UPS to take over should the other one fail or require maintenance, such a solution also represents the most expensive option. There are more cost effective redundancy options in which an acceptable level of resilience can be provided using N+1 systems. A number of N+1 configurations are possible, with a highly modular option offering a good trade-off between system availability and cost. N+1 options Virtualisation is now the norm in many data centres, in particular those 22 providing colocation services. In these scenarios, applications are not hosted on dedicated individual servers with their own integral storage arrays; rather they are allocated virtualised resources according to need from racks of servers and storage which host multiple applications and services, frequently on behalf of a number of clients. In such scenarios, fault tolerance is provided by software that automatically reassigns an application to a properly functioning server resource in the event that a malfunction occurs in the IT equipment. In such situations, instead of having a 2N arrangement - or full redundancy for each piece of critical infrastructure such as a UPS - an N+1 arrangement can be more cost effective and still provide adequate back up. In an N+1 layout, one extra unit can provide resilience to ‘N’ required units. There are several different configurations through which N+1 redundancy can be achieved. In an Isolated Redundant configuration there is a main or ‘primary’ UPS module that normally feeds the load, but this primary module has a separate input for the static bypass circuit. The secondary UPS module feeds the static bypass of the main UPS module or modules and is completely unloaded in normal operating mode. A Parallel Redundant configuration consists of several UPS modules of the same size and capacity connected in parallel on to a common output bus. This approach is more common than Isolated Redundant. In both of the above N+1 configurations, they are comprised of separate UPS units, one of which provides redundancy for all the others. An alternative approach is the Internally Modular Redundant configuration in which the ‘+1’ redundant UPS is built inside the frame of the system, thereby sharing backplane, control system and battery plant within the UPS. Compared with th H\[[Y[[ۙY\][ۋB[\[H[[\ۙY\][ۈٙ\œYHY[Y\[\\و\][[YXY\[[Y\˂ۜY\[HYK[ٙ]Y[X\[\XX]Y] K