The Internet of Things (IoT) promises businesses of all shapes and sizes better visibility of their estate, whether the owner’s infrastructure is comprised of orchards or oil fields. Indeed, analyst firm Statista stated that 75 billion IoT devices will be installed globally by 2025. The resulting surge in raw data reported back by all these devices, such as thermometers, pressure monitors and security cameras, is set to create a bandwidth and data bottleneck. Unless these predicaments are solved, the value of the intelligence sent back will be neutralised.
In order to unlock the potential value that IoT-enabled intelligence promises, organisations must invest wisely in edge computing devices. These data processors will prove invaluable in managing data, undertaking timely analyses and giving machine learning processes a head start. Underpinning this need is the eye-popping stat that we will collectively invest $6.7 billion on them by 2022, according to analyst firm CB Insights.
IoT undoubtedly puts every company at the scene of every single transaction that takes place in its value chain. However, the true value is only unlocked if IoT has the presence of mind to act – the power is diminished if the data or associated processing is delayed. This is indeed a developing crisis when you consider that, in 2019, we had already reached the stage where 507 zettabytes of data had been created by edge devices.
Before we rush to make judgements, all those figures need context. History teaches us that our perception of any new technology can be obscured by over-enthusiastic manufacturers jumping into the market with inappropriate offerings. In the clamour to compete for any lucrative new opportunity, many are tempted to hastily re-package their existing products in a bid to meet the market’s new requirements.
Let us start then with the definition of ‘the edge’, which can in fact vary substantially. It can refer to the capabilities built into an embedded device, a device with more computing power, or just a gateway to which multiple devices connect. However, one universal belief is that the edge of IoT is smart – or should be. Processing workloads can be offloaded to it in the interests of expediency.
The petrochemical industry illustrates this point. Oil rigs, for example, have IoT networks that create masses of intelligence on, say, the underwater landscape, the crude oil pockets below the rock and the capacity of the drilling equipment to retrieve it. Sending all these terabytes of data from the rig to the mainland takes a long time. In the oil industry, every minute is worth millions, given the opportunity cost of lost drilling time or failed equipment. And, as anyone in the industry will confirm, the goal for the IT department is to constantly cut the mean time to oil extraction.
If that information is collated, calculated and decided upon locally, the drilling operation saves millions. Or indeed, can be empowered to make millions. A powerfully intelligent local presence, at the edge, enables this. It doesn’t have to be a giant data centre, just something to take the ‘edge’ off the workload. In effect, a mini data centre, but in reality, a collection of equipment that provides the memory, storage and computing power to make local processing work.
Decisions are made in a fraction of the time, partly because the length of the data journey is decimated – but mostly because those terabytes can be processed locally. Hence, the mean time to drilling drops. This is IoT that thinks global, but acts local!
The same economies can be applied to any vertical market, from manufacturing and motorsport, to hospitality and horse racing.
As with drilling for oil, the IT landscape must be carefully considered before the data miners can begin – if there are masses of data to be processed, it helps if the machinery is in close proximity. Beyond the hardware architecture itself, there is a layer of technical detail, such as the raw capacity of the units, central processing (CPU), memory, data storage and battery life.
The associated software architecture is an even more critical configuration to consider, since performance advantages can be obtained by adjusting parameters such as the operating system, the network, various protocols and syntaxes and supporting software runtime capabilities.
In edge computing, the ‘edginess’ – or the capacity to push boundaries – is a product of a number of factors, including the protocol, network stability, and the need for the device to operate when not connected. The intelligent edge computer is connected, but independently minded. It should be placed in the right conditions, with appropriate networking in both directions, be impenetrable and close with regards to proximity distance.
That’s not all. It may seem like there are hundreds of tiny configurations to get right. Edge scale is about detail, but once nailed, there are no limits to the possibilities.
The next test is something that we may call DeviceOps, where one installs software on or close to the device, constantly monitors it and plans the upgrades to both software and firmware. The Device Life Cycle, the rate at which the hardware gadgets must be replaced, is an additional consideration, along with the device’s function and its effect on data. Are the devices just collecting and forwarding data, or do they have some transforming effect on it? Are they analysing, streaming or imposing any kind of rules on the data?
Such questions are important because edge devices are becoming increasingly functional and getting smarter by the day. Today, they’re mostly used for collecting and forwarding data, but in the future, they will exert increasing amounts of influence on the information they receive, and on the equipment to which they are associated. Their employer (the enterprise) will demand that they collect the information and make key decisions. These might be based on simple rules, or the devices may have some transformative role, such as stream preparation, algorithmic analysis or streaming logic aggregations. The important thing is that by doing this locally – at ‘the edge’ – the actions that follow these edge decisions will happen a lot quicker.
Thanks to the IoT, all enterprises – regardless of their sector – can look forward to a time of vastly increased efficiency, driven by opportunities being seized and developing problems being nipped in the bud.
Interested in hearing industry leaders discuss subjects like this? Attend the co-located 5G Expo, IoT Tech Expo, Blockchain Expo, AI & Big Data Expo, and Cyber Security & Cloud Expo World Series with upcoming events in Silicon Valley, London, and Amsterdam.
