Edge Computing, are we going back to physical DPCs or is it the trendy buzzword?
In a future filled with data from thousands of millions of devices connected to the internet, faster and more reliable data processing will be crucial for industries development. In recent years, we have witnessed the Cloud computing consolidation thanks to its centralized and scalable nature. However, the expanding IoT and mobiles demand have put on trial the network width band apart from increasing Cloud server’s data storing costs.
On the other hand, we should be aware that not all mobile devices need the cloud to operate and here is where Edge computing is ready for action. But, let´s start by the beginning, Edge computing is “the optimization of computing systems in the cloud as they allow to simplify the flow of local traffic towards such cloud. In other words, physical servers are installed near the data source for analyzing the (raw) data obtained in real time and then send only to the cloud the processed data worth keeping, and once it is there advanced analytics techniques such as Machine Learning o Deep Learning can be applied.
For example, a muffin factory that has installed sensors to IoT. If a data processing center is installed next to the factory, many servers together received the data, the center will analyze the data in real time to detect any anomaly in the muffin production. Early detection will allow to act immediately and fix the oven temperature if it is too high. Once this data is processed, the rest of the collected data is sent to the cloud for other departments of the muffin factory processing such as sale forecasts, price fixing, and consumers service satisfaction. Therefore, the data sent to the cloud will help to carry out other data analytics (predictive and prescriptive) that do not depend on real-time but rather upon agreed timeframe (response times set).
Nevertheless, what is the advantage of using Edge Computing? Here are a few examples:
Power Network Analysis and monitoring
With Smart Grids, besides the myriad of data power companies analyze and process, they must add data of power distribution, transmission, and generation. The shift from fossil fuels to renewable sources of energy involves adapting current electricity power network to other electricity generating resources and contribute to the distribution in the existing network. Smart servers implementation, Grid Edge, allow the industry to supervise and analyze the additional resources of generation and transmission of energy and integrate them into the network in real time. In short, Smart Grids generate a myriad of data that help electric power companies to see their availability and required energy, leading to more efficient demand avoiding peak demand and reducing costs.
Remote Monitoring for oil and gas
Oil and gas infrastructure is critical for any country therefore monitoring and control become of utmost importance. This is the reason these two industries are implementing IoT devices for their processes monitoring and avoid possible disasters.
The use of IoT devices for temperature, humidity, pressure, control (among others), generates a lot of data to provide key information of the systems state and short-term forecast of critical operational aspects such as demand (demand forecasting) and failure forecast (predictive maintenance, condition-based monitoring).
With Edge Computing, data is analyzed, processed and delivered to final users in real time, allowing to have access to data while it is created preventing failures and accidents before they occur.
Traffic Management Systems
Smart Transport systems, particularly for traffic management, are examples of how this type of solution is implemented today.
Due to the increasing use of IoT devices and sensors for smart transport and traffic management systems, a great deal of data is collected that requires data processing. However, only a selection of this data is useful for future traffic improvement.
At this point, we should ask ourselves if Edge computing, for the cases before mentioned, is disruptive technology or an attempt to go back to the local “iron” (physical servers inside companies´ offices or warehouses) at the expense of current Cloud Computing. As a matter of fact, suppliers of critical infrastructures would rather choose Edge Computing as for type of applications they cannot rely on the Cloud relatively low reliability and would prefer deployment of redundant and their own IT and communications infrastructures. Thus, data processing is faster and not dependent on communication failures.
On the other hand, most industries applying Big Data analytics use a physical hybrid DPC and cloud system, which brings us back to raise the same question, are we going back to physical DPCs or is it the trendy buzzword?