Investment in innovation has enabled industries to achieve the highest levels of performance and productivity. Business digital transformation has been driven primarily by Cloud computing and the Industrial Internet of Things, which enable optimized monitoring and decision making – see more in IIoT. In this context, the concept of Edge Computing has become a great ally to accelerate the growth of the company’s digital transformation. In the Internet of Things world, data analysis has become the main mechanism for a complete understanding of industrial processes. Much is said about Machine Learning and Artificial Intelligence techniques to extract information from this wave of sensor data. But where does the information from these sensors go?
The most widely deployed architecture in the industry today is sensors that only collect data and send it to local servers or straight to the cloud. Data traffic has grown a lot and tend to increase. Gartner estimates that there are 14 billion connected devices in use around the world by 2019, a figure that will double by the end of 2021, and could reach a whopping 25 billion IoT devices. Will we be able to keep up with this growth to accommodate so much information? Will we be guaranteed low network latency due to data congestion? There are many questions that need to be considered in the short term.
Edge computing is a concept that refers to computing infrastructure existing near or near the data source. Instead of sending data from untreated machines and sensors to the cloud, computationally powered devices can do an intermediate treatment and send only consolidated information to the servers. The idea of bringing intelligence close to the data source will be indispensable for industries and IT infrastructure.
Over the past few years, organizations have incorporated cloud solutions to handle big data for the purpose of gaining insight and becoming competitive. This has allowed for reduced costs, increased productivity etc. Cloud is essential to enable this to happen, and Edge Computing only ensures better management of data from the shop floor.
The standard communication and information distribution architecture on the shop floor of recent years makes it difficult to insert new technologies. In the current model, there is a chain of information that goes through the process in a chained way, from the machine / PLC to the SCADA system, then to the MES, until arriving at ERP and CRM. Through the use of IIoT and edge computing, industries benefit by triggering the process, making architecture decentralized. This operating model enables better supply chain and industrial process management. The next article will be entirely focused on this subject, where more details will be analyzed.
Smart cars can be considered as the next revolution of technology. In fact, they are already reality and are present in some countries. There is a lot of technology embedded in vehicles that incorporate dozens of cameras, radars, GPS, sonar and numerous other sensors that allow you to travel without human interference. An autonomous vehicle is estimated to deliver up to 30 terabytes in one day. The concept of Edge Computing is crucial for managing this data, as well as allowing M2M communication between sensors and actuators to occur in real time to, for example, prevent accidents. The infrastructure of cities should allow the use of numerous gateways operating on the concept of Edge Computing, enabling the rapid exchange of information with autonomous vehicles.
The growth of IoT devices coupled with streaming services like Netflix has made cellular networks reach very high levels of data consumption. In addition, many smartphone applications leave the computing part in the cloud, which demands more data traffic. In the short term, problems of high latency and low network availability inevitably exist. When the 4G network was created, between 2006 and 2010, our daily life did not demand as much bandwidth demand as today. The telecom industry’s answer to such problems is investment in the 5G network, which will allow greater network availability and speed. To maximize network efficiency, decentralized architecture with Edge computing will be crucial to meeting the prediction of tens of billions of connected devices.
Key Advantages of Edge Computing in the Industry
- Low latency
Due to the physical proximity between an intelligent sensor and a machine, reaction time decreases when analysis is done locally rather than on a remote server. The latency between sending data to remote servers to action from it can in some scenarios be considered high and even unpredictable. Edge Computing enables predictability and low latency, making it ideal for critical, real-time situations.
- Cost reduction
Entering intelligence into manufacturing means collecting sensor data, sending it to a database, analyzing and making decisions. If we are interested in applying this concept across the board, the amount of existing data would mean high bandwidth, storage, computational and analytical costs. If instead we put computational power and storage into intelligent sensors on the periphery of machines, it will be possible to apply filters and remove noise from the generated data, drastically reducing the costs mentioned.
- Shopfloor ERP Integration
Just as Edge Computing connects devices and processes without having to send data to the cloud, it enables the connection between the company’s ERP and shopfloor. This framework enables the IT architecture to be more responsive and present data in real time. The ERP system will have more accurate manufacturing process numbers, making planning and supply chain more efficient.
Challenges for Edge Computing Deployment
One of the key challenges for deploying smart sensors is related to safety. From the moment we insert devices into the network periphery, the exposure to attacks also grows. Decentralization of the architecture makes the network more vulnerable and scale as more sensors are added. However, there is another point of view regarding network security: Given the fact that less data circulates between sensors and the cloud, the intensive use of Edge Computing allows information to be stored locally, reducing overall risk.
Another challenge is related to the deployment and management of edge computers. As devices gain computing power, managing their configurations and servicing increases the operational cost. It is up to the company to check case by case whether the investment justifies the use of sensors in the network periphery.