Rodrigo Parreira looks at the model that data behemoths like Google, Facebook and Amazon adopted to cope with a requirement to process and store huge quantities of big data – a model that is on the verge of going mainstream – and asks “how do they do it?”
In the early 2000s, IT giants like Google, Facebook, Amazon and Twitter started to churn through gargantuan amounts of data; a trend that has only intensified with time. Google’s search engine alone reproduces the entire internet on its own servers, in real-time and with numerous back-up options.
Their ability to process and store vast quantities of information begs the question – how on earth do they do this? What sort of infrastructure does an organisation of their stature need to do such heavy lifting? How enormous do their data centres need to be? How much eye-watering capital expenditure are we talking about here?
The answer is surprisingly modest – thanks to a shift to services. These behemoths are developing applications internally to run on generic and low cost servers. Virtualised servers, firewalls and networked storage are basically being mimicked on White Boxes – cheap, generic and highly commoditised equipment.
For almost two decades this approach has been the preserve of these internet giants, but maybe not for much longer.
The model has been so successful it’s on the verge on going mainstream. Large corporations (banks, industries, telecom operators etc.) have begun to ask themselves, if it works for the Internet giants, why can’t it work for our big data?
Their thinking has been buoyed by the growing maturity of the open-source market, manifest in numerous initiatives over the past 10 years, such as KVM, OpenStack, Open Nebula and Open Daylight, and the thousands of globally integrated developers who have mobilised in virtual communities.
This is the story of technology today – the shift to services. Indeed, the ubiquitous move to software-as-a-service, infrastructure-as-a-service – basically any technology delivered as a service, has redefined the way we use hardware and software.
It’s easy to see why. As the lines between software and hardware have blurred, services have essentially replaced physical infrastructure with a far more flexible, on-demand virtual environment - and set businesses on the greatest transformative journey of their lifetime.
Today then, the question for some may still be “Hardware or software?” For the forward thinkers, however the answer is “Services.”