Living on the edge in stormy weather

As South Australia’s recent storms (which knocked out the state’s entire power supply) showed, natural disasters are unpredictable and powerful. What is predictable, however, is that IT clients and consumers now demand uninterrupted access to their data and applications regardless of the weather. For this reason, avenues to futureproof data infrastructure against downtime should always be top of mind. Combining cloud with an edge computing strategy is one way to ensure this availability.

In the case of natural disasters, organisations with a centralised data centre architecture can be affected — regardless of whether the centralisation is cloud, on premise or colocation. With this considered, organisations should be looking at a re-architecture with a hybrid approach to leverage a mix of edge technology and geographic dispersion.

Edge places computing power, control, storage and applications closer to the end users who are using them. Cloud computing is already quite reliable, but combining it with an edge computing strategy results in faster response times and a considerable increase in dependability. Edge will transform a cloud computing implementation from a wholly centralised architecture to a distributed environment. Disruptions would be limited to just the point in the network where it occurred, instead of the entire cloud implementation.

There are three key options for employing edge computing in order to increase resiliency for a cloud implementation. The first is a hybrid strategy using simple devices to support specific, defined applications. An example is a cloud storage gateway, which is a local device that acts as a network appliance or server that translates cloud storage APIs such as SOAP or REST.

The second approach is more complex and involves employing small, micro data centres of 10 racks or fewer. Such data centres are often available as pre-engineered, configure-to-order systems that can be assembled on-site. Or, they may be fully assembled in a factory in a single enclosure and simply dropped on-site. These single-enclosure systems come in durable versions that can withstand harsh environmental conditions.

A third option is to create geographic dispersion by employing a series of regional data centres. These are larger facilities, with 10 racks or more, that have more processing power and storage capacity than a micro data centre. In essence though the idea is the same, but this option enables cloud applications and storage to live closer to the users who are employing them, improving response time while also adding another layer of reliability and resilience.

Natural disasters are certainly unpredictable, but data infrastructures cannot afford to be. Combining cloud computing with a robust edge computing strategy is one of the key ways organisations can ensure they are running effectively 24/7, whatever the weather.

Adam Wilkinson drives Schneider Electric’s on-premise data centre, network and edge computing business. He has more than 20 years’ experience in IT, 15 of them with Schneider Electric/APC and Dimension Data.