Fog computing is an essential development in the cloud and IoT era. Combining it with AI/ML, you have the Intelligent Edge. Like many enterprises, you have made some investments in cloud computing systems, as a way to control costs for IT endeavors as well as to make it more convenient and useful for your users to access and update information for the company.

This may also involve data that you push out to vendors, customers, or other partners. But now, as more and more Internet of Things or IoT devices are getting connected to the global network, it’s difficult to justify a cloud-only setup since there are issues with moving around so many automated data.

As a forward-thinking CIO, you will naturally have a stake in advances beyond ordinary cloud computing systems. That’s why you will want to pay more attention to fog computing.

Characterizing Fog Computing

Fog computing is an activity that happens at the “Edge” where data is gathered or created. IT professionals know that it differs from traditional, earlier approaches to data, where you would collect data from various sources, then route it to a central server for processing and analysis before using it locally.

What’s key here is that fog computing is ideal for users in the enterprise to analyze, process, and use data more efficiently, at the point where it’s created. In many cases, the new information you’re dealing with comes increasingly from iOT devices.

TechBeacon refers to fog computing as a metaphor for a ground-level cloud. “Fog computing moves the cloud closer to the devices collecting the data. This is done to address a big problem: lots of data.” One example is a scenario where a smart grid network assesses, manages, and reports on a set of data. The information holds value for homeowners on the local level, as well as higher up for the manager of the city block it’s on, and then onto the town’s mayor, followed by leaders of the state and national government.

Those who need the information the most are often going to be those who are closer to it on the ground. So a homeowner would appreciate real-time data via fog computing nodes on demand, while a local politician may benefit from daily or weekly reviews of such information (such as crime statistics).

Higher up, there is not much of a priority for someone like the governor of a state or a member of the president’s cabinet to get immediate data, hence by the time fog data has served the most pressing concerns or questions at the local level, it’s appropriate that a governor’s office would access the same information after it had been synced to a cloud server.

Similar hierarchies apply to data for in corporate settings, such as iOT sensors on the ground providing data for local processing in a factory, while daily reports on its activities go to a manager and detailed analysis of data for the entire month goes to the president of a division that uses the information less frequently.

What’s the difference between fog computing and cloud computing? ITProPortal states that “the use of the term ‘fog’ is meant to convey the idea of cloud computing closer to the ground. You can think of it as a layer sitting between devices and a cloud or conventional data center.”

For context, fog computing began in earnest in 2015, when technologists from Microsoft, Intel, Dell, Cisco, ARM Holdings, and Princeton University met to create the OpenFog Consortium, which develops and promotes standards and methods for using fog systems.

“Managing the data generated by the Internet of Things (IoT) sensors and actuators is one of the biggest challenges faced when deploying an IoT system,” according to the National Institute of Standards and Technology. “Traditional cloud-based IoT systems are challenged by the large scale, heterogeneity, and high latency witnessed in some cloud ecosystems.”

The NIST explains that one potential solution is for an enterprise to decentralize applications as well as data management and any data analytics, in a distributed model under fog computing. That’s becoming increasingly necessary, as the NIST projected that the number of iOT devices for 2020 50 billion, with smart sensors, mobile devices, networks of autonomous vehicles, and a whole host of industrial controls all contributing to the rise in deployment.

The main characteristics of fog computing setups, according to the NIST, include:

* Contextual location awareness and low latency: fog nodes are aware of where they are logically in the system, and shorter distances allow for better latency levels.

* Geographical distribution: an example is serving streaming media to vehicles while they move, facilitated by a network of access points placed on highways, roads, and streets.

* Heterogeneity: Data that comes in different formats because it originates from different systems can be processed at the node level, much faster than sending it to a central location and then delivering the results.

* Interoperability and federation: As data moves from one system to another in real-time, the fog computing system will interoperate with various information providers.

* Real-time interactions: Fog computing applications work in real-time, instead of delaying interaction to perform it all at once in a batch.

Fog Computing in Operation

It’s useful to examine a few cases of fog computing to see how an enterprise might use it.

Fingent suggests the idea of a user with a hand-held device who needs to look at the most current video from an iOT security camera. In a traditional cloud system, the technician would need to wait until the video footage was synced in the cloud since the cameras do not have their own storage capacity. Using a fog setup, the video would be immediately accessible to the technician connecting at a local node.

Enterprises that include vehicle fleets or are otherwise involved in tracking the movement of vehicles (such as for verifying delivery) can readily profit from fog computing. TechBeacon offers a scenario where four different parties have an interest in data being gathered from a truck set up with iOT devices. There is the driver, the customer awaiting delivery, the logistics company, and federal or state authorities that are charged with monitoring vehicle safety.

“There are three key places the data can reside: 1. locally on the vehicle; 2. regionally for the logistics manager; and 3. in a central cloud for regulatory reporting,” noted TechBeacon. “The goal is to move the data to the cloud, but the benefit of keeping data locally or close to specific customers provides the opportunity of real-time data.”

Fog computing in this case gives the driver real-time information about when it’s time to take a mandated rest (for safety), while managers use the incoming data to make changes in the delivery schedule. There is also an obvious benefit to the customer, who relies on the delivery date estimation, getting updates as the package gets closer. And there are obvious benefits to authorities investigating accidents if they can access data about the delivery vehicle before the crash.

Fog Computing Benefits and Drawbacks in the Enterprise

It’s wise to assume there will be increased deployment of Internet of Things devices in industrial processes, going forward. An enterprise can use fog computing applications to keep track of data coming from a series of sensors, and when there is a state change, the signal will prompt an action such as opening or shutting a valve or pushing an updated report to another system.

With machine learning or artificial intelligence providing input, ITProPortal explains that your enterprise could use a fog computing node to predict equipment that’s about to fail, so you can automatically schedule and dispatch one of your technicians to arrive on the scene and repair or replace it before there is a catastrophic shutdown.

You’re getting much timelier access to information right at the location it’s needed, without the added pressure of first putting it in your cloud environment. It’s safe to say that no one in your organization will require immediate access to all information being generated and stored. So, a fog solution allows better allocation of urgently needed data versus slower access for less important information that is typically not needed at the local level.

Be aware of the potential drawbacks of deploying fog computing for your organization. Chief among them will have to do with security. It’s much easier to keep a collection of data that’s stored in one place, as in the cloud on a distant server. But when you involve fog computing, the data is kept locally, distributed on more devices. That means there are more opportunities for criminals to break in and grab sensitive information.

In addition to the additional complexity that comes with the fog computing method, there are security issues you’ll need to address, shoring up fog nodes so you can rely on the data being protected just as rigorously as you’d expect from a cloud services provider.

Will Your Enterprise Add Fog Computing to the Mix?

If your organization has been involved in cloud computing for some time already now, chances are you are getting ready to add fog computing to your resources too. With benefits such as enabling faster access to information right where it’s needed and the ability to handle the multitude of iOT devices that continue to be added to networks, a fog system can help you extract more value out of the flood of data.

Instead of trying to shovel all data directly to your cloud solution first, you can designate the real-time data that should move out more efficiently via cloud nodes. Later, you can sync newer fog data to the cloud, at a more leisurely pace. That’s an ideal approach for managing bandwidth and computing assets, especially as you grow in scale and see a corresponding rise in your user base.