Will Edge Computing Lead to New Types of Smartphones?

It has been true for possibly a decade that smartphones have gotten so advanced it has become hard to create some big new feature to drive widespread device replacement beyond normal wear and tear cycles. Some believe foldables or 5G could prove the foundation for an upsurge of innovation and features. 

At least initially, device cost will be a limiting issue, as foldables might cost as much as $2000 each, while 5G devices could easily cost $1200 each. Over time, costs will drop, of course. 

Unknown at this point is whether edge computing, in a deployment pattern that more closely resembles Wi-Fi availability, someday will allow use of devices whose new features are supplied largely by edge computing facilities, not hosted or resident on the device itself. 

The idea is not especially new, having been dubbed by Sun Microsystems network computing back about 1984. The idea was that what we now call distributed computing or cloud computing would allow devices to be “thin” clients drawing much of their compute from a server. 

In a client-server environment that meant a lower-cost “sort of dumb” terminal instead of a personal computer., mimicing the dumb terminals of the mainframe era. It might seem crazy now, but in the 1970s there were whole word processing built on the use of mainframes and dumb terminals, built to support that single function. 

In 1983, when I was a reporter at a Gannett newspaper, I believe we used the Atex system for composing stories, featuring dumb terminals wired to a mainframe. 

source: GSMA Intelligence

Ironically, though the concept never really caught on in terms of devices--which remained smart, with local computing resources-- the idea of using centralized computing resources was common in the client-server era as well as in the cloud computing era. 

It may remain equally powerful in the coming era of edge computing, ambient computing or whatever other term eventually emerges. But the difference might be--once again--the issue of how much capability the client device must possess to supply functionality.

There is at least some hope that, with widespread edge computing with very-low latency, some functions resident on the device can be omitted, with reliance being on the edge computing fabric for those functions.

At least in principle, that would mean the ability to create functional devices available at lower cost. In the last two computing eras--whether centralized or distributed--use of remote servers has become a foundational assumption.