Internet in Things: Collect, Compute, Consume. Repeat

Zeina Saade - 26 January 15

As the Internet of Things (IoT) grows with the popularity of more mainstream devices, such as wearables and household smart gadgets, this is just the start of how the ways we collect and process information are changing. Enablers, such as sensor technologies, make it possible for objects to generate data anywhere in the world, and networks (both wired and wireless) exist to move data to almost anywhere in the world. My theme in this blog post is the rising importance of cloud architectures for the IoT and the data involved.

As technology improves, and device costs fall, the easier it will be to capture large amounts of information. The IoT and big data are now practical technologies for business, and we should start to ask questions such as, where are data stored, and where are they processed? Who controls the architecture and the data processing? What sorts of business opportunities emerge in this new landscape?

More and more thinking is going into ways of building IT systems around the physical world to help run systems at large physical scales. Increasingly, the term ‘big data’ will be out of fashion for the IoT. It will eventually become irrelevant to apply the term because the accumulation of terabyte after terabyte every day (or hour) will be considered the norm.

As millions of devices in the IoT churn out vast amounts of information, the next question is, what happens to it? Gartner's analysts have advised organisations building IT systems for the IoT to transfer data to multiple locations for processing, as single, centralised locations will not be technically and economically viable. With the amount of data being collected, it will be necessary for IT architects to be efficient in the way data is distributed. Matthew Finnie, Interoute’s CTO, identifies ‘machines talking to machines’ as a key driver for cloud computing in the next few years: "The Internet of Things is everything talking to central decision makers, so as the network and the computer merge, it starts to facilitate communication at the edge and the core."

I like to think of the case of a farmer working 1,000 acres of crops. The crops must be planted, fertilised, watered and picked. For maximum efficiency and to achieve a near-perfect crop, let’s imagine she divides her land into segments, and plugs sensors into the soil to monitor moisture levels, light intensity and ripeness of the produce, while actuators determine the delivery of fertiliser, water, pesticide and so on. (This scenario isn’t as far away as it might sound, if you check the numerous stories on the Internet about ‘precision farming’ and ‘wireless crop sensing’.) The sensors could gather data on virtually every single plant, even a humble head of lettuce, while the actuators control its growth. Utilising the IoT enablers, such as sensors and actuators, will help the farmer determine accurate harvest dates, and to some degree, control the time to market.

Now where would all this data go to be processed? In the current business model, an obvious location could be the centralised data centres of the supermarket companies which buy the farmer’s produce and mostly operate the supply chains between farms and consumers. However there are reasons why alternative business models might emerge. The first reason is technical: the sheer volume of data we’re able to collect with the IoT technology is only bound to increase, and at some point it will become over-costly and inefficient to move data to centralised processing locations. It will start to make sense to store and process data in regional data centres that are closer to the point of production, in order to more quickly reach the defining moment for any IoT data, where value is created out of the raw information.

In the age of public cloud computing, there are already technological alternatives to the centralised processing model. Instead of negotiating her data analysis with the supermarket, it’s conceivable the farmer could choose to pay to store her data in regional cloud data centres, and use some kind of ‘data analysis as a service’, to extract the information she needs from her data. I imagine an intriguing result of this is that the farmer could gain a technological lever, not only to produce crops more efficiently, but under a different business model: she is no longer the end point of someone else’s supply chain, but someone in control of her own data and supply chain.

From the viewpoint of cloud computing providers, and companies providing software as a service, there would be challenging questions for cloud architectures. Since public cloud computing got going with Amazon Web Services ten years ago, the tendency has been for large, centralised computer data centres, with the economies of scale which size makes possible. Today, for the data and network demands of the IoT, the most economic solutions may appear as increasingly distributed data storage and processing; that is, different kinds of economies of scale come into play. Depending on the particular business context, data might best be processed and returned locally, or cascaded through a series of processing steps, with only filtered results sent through to the next step. Multiple points of computing would reduce the size of the data which reaches the central site(s) for final-stage processing. This could be more efficient, faster and less costly than transferring raw data across large distances and under a fixed bandwidth that slows everything down. The final destination needs only to handle the summary data that’s required to make, say, strategic decisions.

While the IoT is revolutionising information for the data-driven, number-neurotic (like me), it’s as important to build IT systems which can support this growth. There’s a huge opportunity to disrupt business models and change the nature of supply chains. And some interesting questions about the directions of development for cloud computing.

 

Find out more about Interoute VDC and the Internet of Things

 

Internet of Things, Data, Devices, Cloud Computing, Network

Follow for our latest news