Fog Computing

05/05/2021 184

Since 2006 with the creation of Cloud, by Amazon, we have been able to access resources of all kinds, have scalable architectures at the click of a button and be able to use them from anywhere. In 2008 Cisco announced IoT, Internet of Things, although its beginnings date back to 1999, is one of the technologies that will benefit from the Cloud. IoT will allow me to store sensor data and act on them, automate tasks and react in real time to situations that previously required manual intervention.

Since its emergence we have heard how IoT has spread to professional and personal environments, and how as a result of these new needs the world of sensorization and communication protocols have emerged or evolved. This, together with the combination of all that sensor information and its use through Artificial Intelligence has allowed the emergence of new paradigms. Terminologies such as Smarthome have begun to be used, allowing us to easily and automatically regulate energy consumption, the lowering of blinds or the connection of the air conditioning, among others.

On a larger scale, in cities the concept of SmartCity  arises, and in the industrial world, the terminology Smart factories is coined. What they all have in common is automation, data and its use, as well as automatic decision making.

More and more devices are connected to the Internet and we all benefit from it. In fact it is estimated that by 2025 we will have 75 billion connected devices according to Statista. In the early days of technologies such as IoT, integration with Cloud, seemed sufficient, until the need for exponential growth of data, faster, better communication and real-time actions, requires new standards in the market.

The concept of fog computing emerged in 2012 with the same philosophy as Cloud but covering the needs of the evolution of IoT, among others, which requires lower latency and data processing at the source. Usually the concepts of fog computing and edge computing are confused or interchanged when they are not exactly the same, but we will focus only on the concept of fog computing to understand the differences with Cloud as well as the need for it.

By 2025 we will have 75 billion connected devices according to Statista.

Fog computing, which refers to computing in the fog, is a new architecture in which data, computation and storage are located somewhere between the data source and the cloud. The fog layers consist of compute networks and storage devices such as switches, routers and servers. These devices are called fog nodes. Fog nodes closer to IoT devices avoid cross network traffic and help improve the latency problem that is always a disadvantage in traditional cloud systems. The fog layer preprocesses the data and then transports it to the cloud if there is a greater need for processing.

Examples include changing the configuration of a computer, applying the brakes of a train, sending alerts for a preventive repair or braking an autonomous car at a traffic light. By the nature of the examples, it seems clear that the decision making and execution of the actions cannot be performed in the Cloud, but on devices closer to these sensors that capture the information.         

The use of fog computing has a number of advantages over the Cloud, which technologies such as IoT take advantage of. Firstly, faster and real-time processing, since it is not necessary to resort to the Cloud, but processing is done locally. Secondly, less network traffic, which benefits communication, considering the high number of IoT devices today and in the future, and thirdly, greater efficiency when developing applications and being able to run them wherever they are needed.

The emergence of fog computing led to the emergence of new standards, driven by Cisco, which facilitated communication between devices and the Cloud, in terms of better latency, as well as simpler calculations at the source, i.e. in the sensors themselves or at a closer point.

It was in 2012 when Openfog was designed. Its main objective is the definition of a reference architecture for the deployment of Fog Computing networks, which will serve as an intermediary between sensor networks and cloud applications. Openfog merged in 2019 with Industrial Internet Consortium (IIC) and counts on the collaboration of more than 300 participants, among which are companies such as, Cisco, IBM, Huawei, AT&T and Intel. 

As a recent Cbinsights report shows, there are many industries that can benefit from these new architectures. First of all, medicine, any of the devices measuring sugar, blood pressure or other metrics that have to be sent and processed continuously. Decisions and actions are made faster and locally thanks to fog. Another great application area is the consumer market and logistics, where thanks to this technology and the rapid evolution of sensors the tracking of distribution and stock can be done almost in real time. And another great example is smart cities.

In short, the convergence of different technologies will provide opportunities for new business models, as well as covering business needs that are changing faster and faster. Steve Jobs said:

Stay close to your customers, so close that you are the one who tells them what they need long before they realize it.