Are we already looking beyond the cloud? The Internet of Things (IoT) is driving new approaches.
Own a smartphone? Odds are you do? There are an estimated 229 Million smartphone users in the U.S. in 2017. Worldwide, the number approaches 2 Billion. Pretty impressive numbers aren’t they. And, let’s face it there is a reason – these things are pretty amazing.
An old computer sales acquaintance of mine always likes to say: “When I started out, I sold mainframes – now I carry one in my pocket.” These babies have quite a computational punch and are truly more powerful that the supercomputers of the recent past.
(Image Source of IBM 1401: Wikipedia)
But there is a rub. They rely on talking to the cloud for their “smarts”. That is where vast acres of compute power and data reside that enables the magic on your phone. Your voice assistant doesn’t have all that knowledge on your phone. The same is true for when you want maps and directions, a car pick-up or language translations. The phones need to reach back to home base to actually execute these functions. Why is that a problem? Wi-Fi and cell coverage just seems to get denser and faster. Sure, there is a delay occasionally, but for most of us it is a minor nuisance.
However, now think of much bigger numbers of devices connected, like a whole order of magnitude (10X) bigger: 20 Billion, maybe 30 Billion? Whatever the number, in a short period of time there will be a lot more devices connected to the Cloud. This is the Internet of Things (IoT). The concept goes back to 1982 but really got going with a seminal paper in 1992 that called for ubiquitous computing: “The Computer of the 21st Century”.
Well, we are here. Welcome to the 21st century and the maturation and deployment of the IoT concept. All these devices of differing levels of “smarts” are talking to the Cloud. Got a new BMW 7, or a Tesla? They report measurements, check in for instructions and updates, and/or issue commands. Next time you are on a plane check out to see if GE makes the engines. (There is usually a little logo of the engine manufacturer on the nacelle.) Guess what the engines are doing? Yep, it’s the same thing.
All these coming devices mean they will consume a huge amount of bandwidth trying to get back to the vast sea of computer power in the cloud. In the U.S. especially, bandwidth is a limited commodity. We are not even in the top 20 in most measures of bandwidth. That translates into bottlenecks, slower transmission speeds, and higher network expense.
Besides, do they all need to go all the way back to the data centers in the cloud? Sure, some of the info they generate should but do we really want the IoT in an autonomous car to go back and check with “central” for every decision it could make, maybe it should just check with some of the other devices? How about all the intelligent devices we have in our home? Do they each need to download updates from their makers by themselves? Why not have it all downloaded to one device and then distributed to the others through your local Wi-Fi?
Enter the idea of Fog and Edge computing. The concept addresses the need described. Conduct some of the needed work somewhere intermediate between the device and the central data center in the cloud (“in the fog”) or make the device so smart it can handle whatever is needed by talking to other nearby devices (“on the edge”). Sometimes the terms are used interchangeably and certain vendors promote one more than another, but the idea is straightforward: avoid the backhaul if you can.
Investors are starting to bet big that the demand for IoT will drive an explosion of Edge/Fog Companies. And research today points to the possibility of putting enormous amounts of compute power into edge devices – almost “eliminating” the need for cloud.
By John Pientka