EDGE COMPUTING
Industry 4.0 is the fourth industrial revolution which brings a lot of opportunities to industries and to the world. With the ever increasing outreach of this revolution we come across new technology every now and then. Here, one such technology - Edge computing - is discussed.
DEFINITION: Edge computing is a distributed information technology architecture where data collection points at the periphery of the network are provided with storage and computation for processing data near its generation point.
Let us break down the definition. As the name goes, the edge devices (i.e., data collection points or sensors) which are located at the periphery of a network, are provided with storage to collect data and computation power to process data. The main data center is not bombarded with all the data collected from sensors. Only vital data required for analysis and key decision making are sent to the main data center.
WHY EDGE COMPUTING: It is rightly said that data is the new oil. In modern times, the data generated with IIoT devices is humongous. The task of sending these myriad data to main data center raises the following issues:
Bandwidth limitation
Latency issues
Network unpredictability
In order to address these issues, data processing is done near the source of data generation. For this storage & compute is required near data or data collection point. After processing the data, critical information (not all the data collected) is sent to the main data center for analysis decision making. Gartner predicts that by 2025 around 75% of the enterprise data will be created and processed outside a traditional centralized data center or cloud.
EXAMPLE: Consider a situation where pressure generated in a certain device is a parameter to be monitored with the help of a pressure sensor. So, an IIoT based storage and computation device is provided for this pressure sensor. The pressure is recorded every 0.5 seconds. However, all the values of pressure recorded are not sent to the main data center. Only those values are sent to the main data center which are beyond a certain range. This is done with the help of IIoT based storage and compute provided for the sensor.
USE CASES:
Manufacturing - Sensors are used on machines in an assembly line. These sensors (or group of sensors) are connected to dedicated IIoT devices with storage and compute. After processing this bulk data locally, only critical data e.g., failure prediction data, can be sent to the main data center. This data is used for analysis and decision making regarding predictive maintenance, RUL (remaining useful life) assessment.
Traffic control - Data collected from video analysis of traffic (in smart cities) at a traffic signal is sent to the main data center. This data is used for traffic decongestion by varying the signal length, infrastructure planning etc.
Retail - Real-time data is collected at retail stores. One such data is the number of customers of a particular product. This data is analyzed at the edge to group the number of customers by age or gender once in a month or season or year. This data is sent to a main data center for analysis and business decision making e.g., inventory management.
CHALLENGES: Like any other technology, edge computing comes along with its own set of challenges:
Security - IIoT devices put alongwith the edge devices pose its own security threat. This security threat is not just for data but also for physical devices.
Connectivity - Although edge computing alleviates issues arising from network vagaries by reducing bulk data transmission, the connectivity issue is not completely eliminated. IIoT devices should be capable of handling erratic and poor internet connectivity.