(Source: Wright Studio/Shutterstock.com)
Unlike enterprise IT, industrial applications operate in “‘real-world’” environments and extreme outdoor conditions. Consider the example of a military drone deployed in a rugged, remote combat terrain to gather mission-critical surveillance data. The weather conditions are extreme, and internet connectivity is intermittent. In spite of these circumstances, the drone has to collect data, and then process and communicate the data in real-time. In addition, the drone has to make intelligent decisions based on data analytics. Reliance on cloud computing alone would be impractical in such scenarios. For time-sensitive, real-time actions, much of the processing, storage, and analytics must happen locally in the drone itself. That’s where edge computing comes in.
Cloud computing has been a central component in Internet of Things (IoT) architectures to process and analyze sensor data and to generate actionable insights. Processing data in centralized datacenters or third-party cloud facilities require additional transit to and from the cloud for every decision. The associated delay is unacceptable for time-sensitive industrial applications such as military drones, energy power plants, connected fleets, etc.
Each bit of data sent to the cloud costs money and precious bandwidth. Reliable internet connectivity is a major requirement to transport data. Besides, due to a single point of failure, a security breach in the cloud could adversely affect mission-critical industrial applications.
Edge computing complements cloud computing. In edge (or fog) computing, much of the processing gets done on-premise, near the source of data. In many cases, this could mean that compute, storage, and analytics happens in edge gateways collocated with the IoT devices. An autonomous vehicle, for example, can make immediate maneuvers and control decisions while driving on the road by processing time-series data in the vehicle itself.
Edge computing devices usually are lightweight and supported on a wide variety of form factors. Sometimes edge computing can also refer to a hardware-agnostic software component that can run on on-premise hardware or in virtualized environments.
Edge computing is a major breakthrough for Industrial Internet of Things (IIoT) applications as it meets some of their fundamental requirements.
Real-time decisions: By processing data near its source, latency could reduce from minutes to milliseconds. Smart grid distribution networks, for example, have highly time-sensitive control and protection loops. When an anomaly is detected in the grid network, edge computing corrective control commands can be immediately sent to actuators in real time. Localized computing doesn’t rely on network connectivity. So reliable performance is guaranteed even where connectivity is intermittent.
Overcome big-data challenges:Industrial applications involve a high volume of time-series data. Not all of this data requires the massive computing power of data centers. Local processing significantly reduces the burden of uploading data to the cloud. Moreover, machine-learning applications can run more efficiently to detect anomalies, and to predict failures.
Expedite IoT Integration with legacy platforms: Edge gateways have protocol translation capabilities that support data processing for legacy equipment running proprietary technologies, which cannot otherwise integrate with the open networks.
Improve data security: Edge computing allows the management of a large number of assets within firewalls and reduces the attack surface by limiting round-trip data transfer to the cloud. Edge computing, however, is not a replacement of cloud computing. Offloading less resource-intensive tasks to the edge enables smarter industrial applications. Advanced analytics, which are less time-sensitive, still rely on powerful cloud-based services. The edge and the cloud both maintain device state information (digital replica) synchronized with the physical device in quasi-real-time, as illustrated in Figure 1.
Figure 1: Edge and Cloud computing are state-synchronized. (Source: Practical Industrial Internet of Things Security, Packt Publishers)
Let’s revisit the example of a military drone: While the drone is airborne, edge-computing enables it to communicate with ground troops and the command center in real-time without having to rely on cloud connectivity and the associated latency. After returning to the base station, the drone can upload data to the cloud for advanced analytics and insights. Thus for IIoT apps, edge and cloud computing work together.
Because of exposure to extreme outdoor conditions and physical tampering, it is important to design edge devices with adequate security controls such as secure device operating system, tamper-resistant hardware, and hardware-based root-of-trust, and key storage. The operating system needs to support secure boot and updates, encrypted tunnels for data transport, policy-based whitelisting, etc.
Cloud computing technologies are mostly sector-agnostic. Both banking and healthcare applications, for example, can utilize the same cloud platform. Edge computing, however, involves use-case specific requirements for processing, storage, latency, etc. The latency requirements in a deep-water drilling application are quite different from that in a smart power grid. This opens wide opportunities across industries for edge computing hardware.
A market study by McKinsey and company covered 11 industry sectors including more than 100 use cases where edge computing hardware represents a potential value of $200 billion by 2025. The hardware value includes opportunity across the various components of the full stack namely, sensor, processor, on-device firmware, and storage.
The edge software is usually based on Docker containers and agnostic of both the underlying hardware and the operating system.
Edge computing is a major breakthrough for IIoT as it enables mission-critical IIoT applications to operate in real-time. The technology improves bandwidth utilization, saves money by reducing opex, and allows smooth operations even where internet connectivity is intermittent. Also, edge computing does not replace but augments cloud computing, opening new market opportunities for hardware vendors.
Sravani Bhattacharjee has been a Data Communications technologist for over 20 years. She is the author of “Practical Industrial IoT Security,” the first released book on Industrial IoT security. As a technology leader at Cisco till 2014, Sravani led the architectural planning and product roadmap of several Enterprise Cloud/Datacenter solutions. As the principal of Irecamedia.com, Sravani currently collaborates with Industrial IoT innovators to drive awareness and business decisions by producing a variety of editorial and technical marketing content. Sravani has a Master's degree in Electronics Engineering. She is a member of the IEEE IoT Chapter, a writer, and a speaker.
Visualizar em dispositivo móvel
Centro de privacidade |
Termos e condições
Mapa do site
Direitos autorais © 2021 Mouser Electronics, Inc.
A Mouser® e a Mouser Electronics® são marcas comerciais da Mouser Electronics, Inc. nos EUA e/ou em outros países.
Todas as demais marcas comerciais são de propriedade de seus respectivos proprietários.
Matriz e centro de logística em Mansfield, Texas, EUA.