Edge implies closer to the devices and an edge agent is an intelligence that sits on a connected device, to collect, process, and analyze data locally without pushing the entire volume of data to the server sites. It also helps achieve faster analysis, and quicker detection of anomalies. When a connection fails between edge and the cloud server, edge simply stays cool, and makes all the decisions - even calling the shots for critical actions, that helps the overall system or service function smooth. Putting in simple terms, edge computing is nothing but the ability to collect and process data at the device level, without having to wait for device-cloud server integration.
Today most enterprises are brimming with data from multiple assets being monitored over the internet. Devices are getting smarter and becoming their own decision makers. Internet of Things is making it possible for businesses to use the data to cut unwanted expenses, improve efficiencies and create strategic business plans to sustain competition and create new revenue streams.
Why edge computing?
In a hyper connected environment, connectivity issues, legacy systems from brownfield deployments, server downtimes, lack of visibility, and general latency are all the pain points edge computing irons out. Sometimes the devices might not be able to handle the data load and seamless communication between device and server might pose a challenge. Streaming the data overload from device to IoT server can become slow and inefficient. Deploying a technology that seamlessly handles the growing data deluge is imperative for any enterprise IoT application. Edge takes care of the data at device side, so that all the data need not be transferred to the IoT server for processing. This saves bandwidth, avoids latency and improves pro-activeness of the device.
Today, devices connected to the IoT network are diverse. They could be traditional ones that cannot establish communication with IoT server or smart devices and sensors that communicate directly. Volumes of raw data that are processed into actionable insights needs to pass on from end nodes to the server without any interruption. Besides, there should be seamless two-way communication and control between IoT server and the sensors.
To begin with, there are certain challenges that need to be addressed in the existing enterprise scenario.
1. Most organizations have already invested in capital intensive assets that aren't IoT ready. It is not economically viable to scrap down the entire framework and start from the scratch. For instance, sensors in high CAPEX Diesel generators across an enterprise might not be smart enough to collect data and talk to the central IoT Server. In this case it is necessary to facilitate the communication through a light weight application agent that helps in fetching the sensor data seamlessly and push it to the IoT server without any delay. The light - weight application is the 'edge intelligence' integrated with the IoT platform on the device side.
2. Secondly, not all devices are compatible with modern IoT-related protocols. It is crucial that sensors devices on board start communicating using protocols such as MQTT, HTTP, AMQO, Co-AP, LORA, etc for uninterrupted data acquisition and processing. However, some of the traditional devices aren't capable of doing so without the help of an agent application.
3. At times of intermittent connectivity or poor network coverage, device - server communication gets stalled. The downtime isn't acceptable for critical operations. When integrated with edge computing capability, it is possible to achieve localized data processing & analytics that initiate actions based on the insights got from the edge.
Benefits of edge computing are huge, and are solid proof of what edge-empowered devices could achieve. Edge helps in achieving faster responses, reduces latency due to a great extent, makes legacy and modern devices interoperable, and empowers local decision making that results in huge cost savings. Edge intelligence is not restricted to just industrial IoT, and cuts across multiple verticals. Edge helps achieve high fleet efficiency and evolved logistics, helps in better energy management, better run facilities in a sustainable manner, enable remote management of assets and building spaces.
Thus it is promising to note, the move towards strengthening edge computing capabilities is gathering steam. While at present, only 10 percent of all enterprise-data is created and processed outside of traditional data centers and cloud servers, Gartner predicts this figure to rise to about 75 percent by 2022.
Karen Ravindranath is the director – WebNMS IoT (Zoho’s IoT division).
Disclaimer: This article is published as part of the IDG Contributor Network. The views expressed in this article are solely those of the contributing authors and not of IDG Media and its editor(s).