What the fog? How fog computing and IoT mix

Edge computing

With the advent of IoT, we see more and more data being generated by sensors, somewhere in the field (often quite literally). In some use cases, we want to do a lot of local processing. Take, for instance, sensors on a modern-day car. Processing of the data needs to be done locally; and with the increasing amounts of data being sent over it, it would be impractical to send all of that to centralized location. But mention the fact that action that needs to be taken is local, and an example would be a distance sensor on the front of the car that start registering the diminishing distance between sensor and the car in front. The action, of course, is to brake.

Fog computing

Fog computing, fog networking or fogging is an architecture that uses edge devices to carry out a substantial amount of computation. The National Institute of Standards and Technology (NIST) defines fog computing as follows:

Fog computing is a layered model for enabling ubiquitous access to a shared continuum of scalable computing resources. The model facilitates the deployment of distributed, latency-aware applications and services, and consists of fog nodes (physical or virtual), residing between smart end devices and centralized (cloud) services.

The document from NIST is something that is nice to read and will give you more insight into the concepts of fog computing. For me, it opened up a whole new set of concepts including mist computing (I am not kidding). But what would I’d like to do rather than going into all the technical details is to discuss a use case of these new paradigms and what they mean for you!

Everyone can be a provider

You can see it as an alternative to cloud providers like Amazon, Microsoft or Google. Rather than a centralized and massive cloud computing platform, you are now creating computing resources at what they call the “edge” of the network. You can add your computing resources to a network and start making some money with some spare computing resources that you might have. Together, these devices at the edge of the network make up a massive computing platform, very much like the grid computing efforts in the beginning of the 21^st century. One such company that offers such a service is SONM.

SONM offers general-purpose, cloud-like computing services (IaaS, PaaS) based on fog computing as a back end. Computing power suppliers (hosts) all over the world can contribute their computing power to SONM marketplace. Users will, according to SONM, get cheaper computing power compared to cloud providers. The system SONM is offering is Linux- and Docker-based, with payment based on Ethereum smart contracts.

Use case

Although SONM says it is for general-purpose computing, some of the examples that it gives are for a very specific use case — for instance, rendering training models in the case of machine learning and video rendering. There are also examples that are more toward the nature of fog computing (the edge) — for instance, video distribution and content distribution networks.

Would you use it?

The question is, of course, would you use it? Or even would I use it? At this moment, it’s hard to say — we get an idea about what fog computing is, but the devil is always in the details. As a company, we use cloud computing for the services we offer to clients, and we pay for what we use. We hardly have any spare computing resources that we can add to such a market — and I’m not sure if we would want to if we did.

The world is changing, and we see more and more IoT devices arising, along with the demand for edge processing. As a part of a digital transformation strategy, fog computing, be it the form that SONM is offering or not, could very well be part of the roadmap. It’s a question of additional research from our side that is necessary to see if this will be beneficial to use as an alternative to cloud computing, or even to identify if there would be any cases for our clients use for computing. But this doesn’t mean that fog computing is useless; it just means that we haven’t fully identified the use cases.

In case you do have some spare computing power that you might want to recover some of the cost associated with, look into SONM or a similar offering; if you are on a shoestring budget, cheaper computing resources might be beneficial. So, it all depends on your use case.

What do you think? Is fog computing a hype? A trend? A fad? Let me know in the comments!

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.