Thousands or even millions of devices directly to the cloud is often not feasible due to cost, privacy, and network issues. Edge computing facilitates data processing at the edge of a network, close to the sensors, thus helping avoid the issues of sending all data directly to the cloud.
Edge computing addresses many of the technical challenges experienced by large-scale IoT applications with the following advantages:
Low network latency – Network latency is a significant issue for safety critical systems, like a self-driving car or a factory control system. The difference between a response time of 100ms and 1ms can be life threatening. Edge computing allows for reduced latency due to local processing of key compute decisions.
Data privacy and security – Reducing the attack vector of an IoT application is a way to create a more secure system. Edge computing reduces the number of devices connected to the internet and local data filtering reduces the amount sensitive data being transmitted.
Reduced network load – By processing data locally, edge computing can significantly reduce network bandwidth requirements. This is particularly important in cases where bottlenecks might occur due to unreliable and constrained network connectivity.
Computational efficiency – It is often more efficient to perform data analytics and data processing on smaller data sets and with longer inter-arrival time. For example, more efficient analysis can be performed on a smaller data set (such as smaller geo-specific areas) and data that is arriving with a longer time interval (in the order of seconds instead of milliseconds).
Reduced cloud costs – Local storage and processing not only reduces network load, it also avoids sending unnecessary data to the cloud, consequently decreasing the costs associated with cloud storage and processing.
Autonomy – For certain use cases, it is critical that the local application can continue to run in disconnected mode. Edge computing allows for local execution so a system can continue to run autonomously even if there is no network connectivity.
Using an edge solution such as an IoT gateway is applicable to a wide range of industries and use cases. The following use cases provide additional insight into how edge computing can solve some of the challenges of IoT deployments.
Industry
Typical factory floor IoT applications require quick local response times enabled by edge computing. Examples include:
Safety-critical applications - protect humans and machines from physical injury or damage with fast, sub-second response to sensor input.
Predictive maintenance and condition monitoring - perform data processing locally on an edge device and reduce the amount of sensitive data sent outside a factory.
Industrial automation - operate a programmable logic controller (PLC) that runs directly on a high performance gateway to process sensor information and control actuators such as robots.
Retail
Supermarkets and retail stores can benefit from IoT applications to improve the efficiency of their buildings and customer experience. Some examples how edge computing helps retailers:
Inventory management and operational efficiency - send employees push notifications for re-stocking vending machines to avoid customer dissatisfaction and ensure availability.
Food safety and regulatory compliance - monitor temperature, humidity, and air pressure locally, without the need to involve the cloud. Reduce cost and response time with locally triggered notifications and timely corrective action.
Mobility
Mobility applications such as autonomous driving, and remote monitoring and diagnostics of automobiles are well suited to edge computing. In particular, an edge gateway can run on a vehicle's telemetry control unit to enable the following functionalities:
Communication: Edge software can act as the communication client to enable the exchange of control information for device management and telemetry data.
Applications: Edge software can run user applications on the automobile, such as emergency call service (eCall) that provides precise location in case of accidents. Another example is a diagnostics application that collects logging information coming from different components within the vehicle to identify problems.
Updates: Edge software can be used to coordinate software updates of the vehicle's electronic control units, head units, or navigation map material.
