Fog Computing vs Cloud Computing – What’s The Difference?

Difference between fog and cloud computing: a glowing blue cloud symbol inside a futuristic data center with transparent digital overlays and numeric data visualizations.

You may have heard the terms “cloud computing” and “fog computing.” And you may be a bit “foggy” on what exactly is the difference between the two. What exactly is the difference between fog computing vs cloud computing, if any? And if so, just how different are they?

The main difference – at least as it is being defined these days – comes from the fact that the cloud exists via a centralized system. All things connect and report to, the cloud. Whereas in a fog computing environment, everything is decentralized, and everything connects and reports via a distributed infrastructure model.

Why Do We Call Them Fog Computing and Cloud Computing Anyways?

Despite its seemingly ubiquitous nature, The Cloud has its shortcomings. Although it’s a powerful platform for many applications, most notably the diverse array of X as a Service (XaaS) offerings, latency issues make it a less than perfect vehicle for supporting applications that require virtually instantaneous information processing. Moreover, that list of applications is growing day by day as the Internet of Things (IoT) continues to expand and connect things we never thought were connectable,  let alone worthy of a connection. Conservative estimates put the number of connected IoT devices at 55 billion by the year 2025.

The idea of the Fog as it applies to computing was coined by Cisco back in 2014, who registered the name “Cisco Fog Computing.” The soon-to-follow associated term “fog computing” then attempted to define a suite of new product capabilities developed to push computational and storage functionality even closer in proximity to end users – closer than the cloud, at least. Since this application is an extension of the cloud, the fog metaphor pretty well reflects the relationship between the two (The cloud residing in the distant sky, and the fog manifesting itself close to the ground) via their meteorological linkage. Think of the fog as cloud you can touch, if you will. A cloud that is even closer to the inevitable influx of data that the IoT is ushering in (remember that figure – 55 billion devices).

How Fog Computing Works

Edge and fog computing are often used interchangeably as both move computational processes and data storage closer to end users. It’s the subtle distinction in where the intelligence and computing power reside which serves as the primary difference between the two. In the fog architecture, the LAN is the home for these functions. Connected devices transmit data to a gateway where it is forwarded for processing and return transmission. In the cloud model, these computational and storage resources are anywhere but local. The distinct advantage here is that in the fog model, your latency is significantly reduced, as are transmission and backhaul costs associated with longer transmission times and distances. In short, if you need instant, you need to get out of the cloud.

The Fog and the Edge: Six of One, Half Dozen of Another?

If you think that the fog and edge are terms of distinction without a difference, you’d be mostly correct – which also means you’d be partially wrong. In advocating one technology over the other, supporters point to a narrow set of variations. That “narrow set of variations” is still enough, however, to warrant a distinction.

Edge supporters see a structure that has fewer potential points of failure since every device operates autonomously to determine which data is processed and stored locally or forwarded to the cloud for more in-depth analysis. Fog enthusiasts (Foggers? Fogheads?) believe that the architecture is more scalable and provides a more comprehensive view of the network and all of its data collection points.

In many respects, fog and edge computing are, in fact, complimentary. The coming of 5G wireless paves the way for the placement of sensors and antennas in any number of unique locations ranging from park benches, bus stop enclosures to streetlamps, thereby requiring an additional remote computing layer to cloud/edge networks. In these instances, fog structures will merely act as extensions of strategically located edge data centers.

Too Clever by Half

What is “too clever by half”? For the uninitiated out there, Merriam-Webster defines the phrase as “clever in a way that is annoying or causes problems.” So, the idea is being clever in a sense that is worse than being clever merely for the sake of being clever, but a degree of “clever” that actually winds up being detrimental.

Here’s an example: in the movie The Princess Bride, Vizzini thinks he’s killed Westley by tricking him into drinking poison, “You fell victim to one of the classic blunders! The most famous is “Never get involved in a land war in Asia,” but only slightly less well known is this: “Never go in against a Sicilian when death is on the line!” Of course, as he makes this boast, he falls over dead.

Naming conventions for technology sometimes are the results of being “Overly clever,” and while their initial intention might have been pure, they often wind up confusing the issue rather than illuminating it. “Fog Computing,” like its namesake is murky, obscure, even mysterious, and in the context of edge computing – not very clearly understood.

Fog and Edge Data Centers

If we think of edge data centers as performing the same essential functions as traditional data centers, and they do, and define fog operations as running applications from network devices, routers, and switches for example, then in many instances edge facilities will serve as a significant point of convergence from both a hardware and software perspective. In this application, edge data centers, like their larger cousins, will provide the underlying platform to agnostically support fog network operations be they from Cisco, EMC, VMware or Intel.


What constitutes “the edge” is still a topic of vigorous debate, but when we consider that in a few short years getting as close to the end user as possible will likely mean that everything from someone’s refrigerator to a stop sign – any boundary – will be artificial, dynamic, and as technology develops: temporary. The inexorable extension of networks will result in networks consisting of the cloud, the edge, the fog and whatever comes after that (The Mist? The Humidity?) How we refer to these elements is irrelevant to anyone who’s not an industry analyst or marketer. New terminology will continue to arise in an attempt to put a name to a new mode of operation or application. However, if we are destined to remain “too clever by half” then it’s better that we apply said cleverness towards the development of the actual supportive architecture rather than what we choose to call it.