Traditional datacentres ill-equipped to cope with IoT demands, warn technology experts

Enterprises need to look beyond the datacentre when it comes to selecting the best place to store, collect and analyse the data generated by internet of things (IoT) devices.

Speaking at the DatacenterDynamics Converged conference in London on 18 November 2015, Brandon Butterworth, chief scientist at the BBC, said the datacentre may be ill-equipped to cope with the sheer volume and velocity of data IoT devices generate.

“The problem is this just doesn’t scale. If every product is going to make data, and require a datacentre to process it, the products are going to get really expensive – and what are people actually going to use that data for?” he said.

Therefore, it might be more efficient for enterprises to store the data their IoT devices collect more locally and process it on the fly.

“If a datacentre requires information to be distributed and redundant, why not leave it in local places? Why pay to store it yourself? Why not leave it on devices and pick up bits and pieces of it when you need?” he added.

“The problem with big data is that people want to collect it because you might find something useful to do with it later, but you might find it is all noise.”

Speaking at the same session, Jane Gilmour, international chief technology officer (CTO) of Coca Cola, said the drinks manufacturer is on a mission to move 80% of its computing environments to the cloud in the next three-to-five years.

“We have moved away from traditional datacentres in part because of the latency. We have milliseconds to make a decision before the consumer walks away, and you can’t do that from a traditional datacentre,” she said.

Coca Cola has a number of IoT initiatives on the go, particularly on the supply chain side of its business. A third of its vending machines are internet-enabled, so the drinks manufacturer can keep close tabs on customer buying habits and product inventories.

It also collects the data generated by the selections customers make when using its “freestyle machines”, which allow users to choose from 100 or so pre-selected flavours to create the drink of their choice 

By tracking the flavour combinations customers opt for, the company can use this information to put a product into production that users could buy off the shelf at a later date, Gilmour explained. 

Coca Cola has embarked on these projects to try and answer some clearly defined business questions around how the consumer enjoyed the drink they have purchased, she said.

The company answers a lot of these using traditional questionnaires, but the IoT could allow it to embed sensors in drinks fountains and at other points of sale later down the line.

The company’s questioning approach is one Gilmour said other firms should consider taking, as there is a danger they may end up collecting IoT data for the sake of it.

“When you get down to business drivers and why you’re doing it in a company, it pays to have very specific business questions you’re answering,” she said.