When we think of “cloud” we think of a vast collection of compute, network, and storage capabilities that resides somewhere high above us—a massive repository of functionality that can be accessed from anywhere and any device with enough bandwidth to handle the data flow.

With practically unlimited power and scalability, cloud technology has been a key enabler of the Internet. But the Internet of Things (IoT) demands something more. IoT is a broad collection of sensors, cameras, smartphones, computers, and machines—all connected to and communicating with applications, websites, social media, and other devices. To maximize value, much of the data generated by these “things” must be processed and analyzed in real time. For example, sensors and cameras in and around a large retail store may continuously collect data about customer volume and traffic flow. The store can derive some value from all this data by sending it back to the cloud to analyze long-term trends. But the value is multiplied if the system can process the data locally, in real time, and then act on it immediately by sending more cashiers to the check-out line just before a surge in customer traffic.

This sort of real-time, high-bandwidth application requires a new distributed cloud model that brings cloud networking, compute, and storage capabilities down to earth—to the very edge of the network. My friend Flavio Bonomi has worked tirelessly with both academia and other industry partners to advance the concept of fog, inspired by the way the San Francisco fog extends the cloud to the ground. Fog computing creates a platform—what we call a fog node—that provides a layer of compute, storage, and networking services between end devices “on the ground” and cloud computing data centers. Fog is not a separate architecture; it merely extends the existing cloud architecture to the edge of the network—as close to the source of the data as possible—to enable real-time data processing and analytics.

Fog Computing Extends Cloud Capabilities to the Edge of the Network. Source - Cisco, 2015
Fog Computing Extends Cloud Capabilities to the Edge of the Network. Source – Cisco, 2015

Fog computing solves some of today’s most common challenges:

  • High latency on the network
  • Challenges with end-point mobility
  • Loss of connectivity
  • High bandwidth costs
  • Unpredictable bandwidth bottlenecks
  • Broad geographic distribution of systems and clients

Fog is a key enabler of the Internet of Everything (IoE) and will drive an array of new use cases in every area of life and industry—from retail to healthcare to oil and gas production. Preventive vehicle maintenance is just one example: the sensors in new connected vehicles generate up to two petabytes of data each year. It would be impractical and prohibitively expensive to send all of this raw data back to the cloud for real-time processing, but fog is turning these vehicles into mobile data centers that can sort and index the data, then send alerts when action is required—such as checking an overheated engine, or filling an underinflated tire.

IoT is driving this architectural shift to the edge, bringing processing, analytics, and even applications close to the sources of data, and enabling real-time response to real-time information. You can find out more about Cisco’s fog computing solutions for IoT here.


Maciej Kranz

Vice President and General Manager

Corporate Strategic Innovation Group