Storage considerations for containerised data centres

Containerised data centres are useful disaster recovery infrastructure. But if you use a lot of storage, don't expect off-the-shelf units to do the job.

If a containerised data centre appeals as a disaster recovery option, prepare to speak to a specialist provider to ensure your storage gear survives the trip. That’s the advice from Mario Baecker, a Solutions Architects at containerised data centre vendor Rittal, which recently installed one of its products at Orion, the provider of electricity networks in earthquake-hit Christchurch.

Containerised data centres are often used in situations like Christchurch. Some organisations also maintain a containerised data centre, complete with a collection of equipment mirroring their permanent facilities, as a disaster recovery tactic. Others deploy the rigs because they have either run out of interior space or know an office move is imminent.

Whatever the application, Baecker says storage needs some special consideration in the design phase. Rittal’s containerised data centres are, he said, purpose built as shipping containers cannot be sealed against dust and are not sufficiently strong to house densely-packed racks of IT equipment. Extra-strong raised floors are standard, as are fixtures that allow racks to be fixed to floor and ceiling to ensure greater strength and lower the likelihood equipment will move in transit.

Shock absorbing equipment is another option to protect equipment, and Baecker said this is sometimes requested by clients who use containerised data centres in locations where vibrations can be expected (Disk performance can be degraded by vibration, as this video famously demonstrates).

Baecker told SearchStorage ANZ that Rittal assumes a reasonable mix of equipment will be housed in its containers, and that a customer that wanted to deploy a dense (and therefore hot and heavy) storage array would need to consider a custom-designed interior.

“Before we build the container we would speak about what racking systems are required,” he said. “To be perfectly honest we need to look in detail because each application is different.”

That custom design could include extra mounting points to keep arrays stable. Baecker added that the company would also consider the route used to transport a special container, to further reduce the chance of damage.

However you chose to design a container, Baecker added that it makes sense to keep it running as the environmental management systems used to monitor the devices rely on SNMP data output to preserve optimal environments. While running a disaster recovery data centre will cost you a little money, being able to monitor it to ensure it is ready to get to work in the event of the disaster is a sensible investment.

Read more on Data protection, backup and archiving