The cloud has many benefits, but we must be clear-eyed about the downsides as well. Here are a few things to consider.
Not so long ago I went to the opening of a gigantic new datacentre on the outskirts of London. It was a cold day and, despite a few heaters scattered around the hall, most of the heat was lost straight to the ceiling of the vast space, which would soon be filled with endless racks of servers.
One of the executives noticed me shivering. “These places aren’t really built for humans,” he said by way of apology.
But while it may make sense for individual companies to move many of their applications and data into the cloud, is there also a danger that a wholesale move cloud-wards could create new and unexpected risks?
It’s becoming very clear that the cloud will be the main way that applications and processing power will be delivered during the next era of computing.
Data used to live on the mainframe; then it moved to the PC and the corporate data centre; now it has moved on again, into the cloud.
Companies no longer need to hire expensive IT staff to manage and maintain their own infrastructure: they can buy computing as a service instead, and have their applications and data — once jealously guarded — held by others and delivered from an anonymous data centre that may be in a different country, or on a different continent.
For their part, the giant cloud providers can generate significant economies of scale by delivering services from their own mega data centres, in which every factor has been calculated and refined to make the installation as efficient as possible.
Few businesses can spare the effort to optimise their systems with such dedication, and so increasing numbers are deciding that it’s better to trust their data and apps to the cloud giants. From a cost point of view, it seems that the cloud makes sense financially for applications with a variable or ‘bursty’ need for computing power (those used just once a month, for example). For applications with a predictable workload, however, the cloud may cost more than more traditional models.