The Stack Archive Article

Containerised data centres

Thu 28 Jul 2016

We’re all aware of our dependence on technology, as exemplified by smartphones, automated booking systems, satnav and much else besides. However the depth of this dependence, and the scale on which we use technology, may not be quite so apparent.

Consider a music festival, for example. An event of any size calls for highly competent management to ensure an enjoyable and safe experience for all, and you’d expect the management team to have a level of IT resource – maybe a few laptops – at their disposal to facilitate their situation awareness, communications and control. And you’d be right, except that the IT resource may well extend to a complete data centre rather than just a few items of personal kit.

‘Pop Up’ data resources like this are implemented using containerised data centres, providing IT in a box on any scale from a single cabin to a large multi-module solution. With wide use in the military, they can also be deployed wherever IT and communications resources are needed temporarily, for example in offshore and land drilling oil and gas projects, sporting as well as music events, office relocations and major construction sites.

As their name suggests, containerised data centres are typically housed within shipping containers so that they can survive transport by planes, cargo ships or road, or even slung under helicopters, for delivery at short notice to any required location. As complete, self-contained units they can be made operational within a day of arrival. They can withstand extreme temperatures and prevent ingress of water, dust and sand.

To be self-contained, the data centres include all necessary components; compute, storage and networking hardware but also cooling, power protection and conditioning, and possibly a generator for full-time or back-up operation.

The power protection and conditioning is handled by an uninterruptible power supply (UPS), which has to share some of its host data centre environment’s properties. It must be easily scalable, since containers may be populated with one or several equipment racks. Remote control and monitoring must be supported, as the data centre may be deployed in a remote and unmanned location. These circumstances also require the UPS to have good resilience to internal faults.

UPSs with a modular topology are well-equipped for these applications. They can be easily and quickly scaled by populating racks with plug-in modules, and then adding more racks if necessary. Costs, and above all demands on very precious floorspace, are kept to a minimum. Scalability also provides resilience, because adding an extra module provides incremental capacity; N+n redundancy can be efficiently set up, rendering the UPS resilient to the failure of one or more modules.

Remote monitoring and control can also be facilitated by providing the UPS with simple network management protocol (SNMP) capability if it isn’t already built in. This vendor- and platform-independent protocol will allow communication with other manufacturers’ networks locally or over the Internet, so that the UPS can be managed from a remote location. Central operators can log events, monitor power supply quality, review battery conditions and initiate UPS self-diagnostics. SNMP also facilitates remote control and rebooting of UPS-protected devices, graceful load shutdown of the critical load during extended power failures, and other management functions.

Containerised data centres allow an effective response to today’s dynamic, high volume data processing and communications demands. They can be built cost-effectively, deployed quickly, and operate reliably in remote locations. Choosing the right UPS contributes significantly to these centres’ ability to realise their full potential.

Send us a correction Send us a news tip