Big firms and start-ups are vying to stake their claims as computing moves to the “edge”, a new frontier of decentralised data crunching.
Over the decades computing has alternated between being centralised, as in the cloud, and being distributed. With cloud computing dominant, the tide is turning yet again in the technology sector. Driven in no small part by the Internet of Things (IoT), more and more data is being processed not in large, centralised data centres, but on the “edge” of local networks and smart devices. A major technological shift is underway, and the race is on as start-ups and big incumbents strive to colonise the edge and control the IoT.
Demand for computing at the edge is growing, as companies discover that keeping it local can be more practical, particularly when there is a need to process data quickly. Self-driving cars offer a prime example, as they generate up to 25 gigabytes per hour, by some estimates. As The Economist illustrates, “Before so many data are uploaded, and driving instructions sent back, the vehicle may well already have hit that pedestrian suddenly crossing the street.” In many industrial settings, the ability to analyse data as they are captured translates directly to higher revenues.
All of this does not spell the end of cloud computing; in some cases, the two may continue to work in tandem, particularly with devices that use artificial intelligence. For example, the algorithms of autonomous cars are initially “trained” in the cloud using recorded driving data. They are deployed only after this process is completed, then begin to drive by interpreting live data themselves. Still, in the case of some devices that depend on live data, training algorithms locally makes better sense.
Peter Levine of Andreessen Horowitz, a Silicon Valley venture-capital firm, predicts that centralised clouds like those of Amazon, Google and Microsoft will continue to grow. Concurrently, start-ups with smaller, more local data centres are proliferating. Representing yet another iteration, telecoms equipment makers such as Ericsson and Nokia, as well as network operators, are getting into “mobile edge computing”.
For computer makers like Dell EMC and HP, the shift to the edge presents an opportunity to regain lost ground by providing gear for firms that want to crunch data locally. The edge is also being populated by big cloud computing providers. Microsoft changed its slogan from “mobile first, cloud first” to “intelligent cloud and intelligent edge” last year, and sells services that send AI algorithms to devices. AWS has a service called Greengrass, which handles data for both local and cloud communications. IBM’s motivation to buy the Weather Company in 2015 was to acquire not only weather data, but numerous “points of presence” for edge computing.
What all of the different players, entrenched and new, seem to be moving towards is a world in which data collection and analysis takes place exactly when and where it is needed.