What is Edge Computing?
Soon the Internet of Things is going to give physical devices like doors, chairs, coffee makers, refrigerators, their own Internet Connection. Edge Computing will allow processing of data gathered from these devices locally and with the help of Artificial Intelligence and Machine Learning, will be able to make sense of this data. These things will become important decision makers.
So what is the Edge Device?
Edge in literal sense means the end or a boundary of a particular area. In the IoT world, Edge stands for the same thing. An Edge device stands at the boundary between a Field Device and a Cloud Storage. It is a logical controller for data to be acquired, processed and then transmitted from the field to the cloud.
What is the Difference between Edge Computing & Cloud Computing?
Cloud Computing is a Computing on demand service. When some one says their application is a cloud based one, it means that it is hosted on a distant Computer server, which is owned by IBM, Google or Amazon. Instead of having everything locally on a computer hard drive, everything is available on the Cloud. A cloud interface offers flexibility of not having to own expensive computing hardware to host an application.
Since its revenue structure is a “Pay as you use” one, running an application on a cloud storage is very profitable as it only charges you for the space and the computing power that you use. The most well known examples of Cloud usage is Google Drive and Apple ICloud.
The idea of Edge Computing is to push these cloud services closer to the network, to the devices themselves. Thus Edge Computing enables the captured data to be processed at the same location. With the help of Algorithms and AI, these machines can make decisions locally without the need of the internet or the cloud. It can be seen as a decentralized cloud.
What is the point of Edge Computing if Cloud Computing is so sought after?
Wasn’t Cloud Computing brought forward to give up hardware and only make use of the software when needed?
The answer to this is that in today’s fast moving world, avoiding latency is key. If data is processed at the time of acquisition itself, only already processed data is transferred to the cloud for a data visualization. This increases the speed of data that is transferred as processed data occupies less storage than raw acquired data.
As the Internet of Things approaches us, the size of the data that is collected is going to grow exponentially. Sending this data to the cloud for further computation is going to increase the bandwidth required and burden the network.
The benefit that Edge brings is to compute and transfer this raw data at the source itself, so as to reduce the size of the data to be transmitted and increase the speed of data transfer. Since data computation happens closer to the source of data generation, real time monitoring and analysis can take place bringing the machine closer to the operator.
Looking at the vast amounts of data that machines generate, sending all this data for computation to the cloud, is not only going to increase the processing power of the cloud, but also increase its cost of usage. With cost of usage going up and Computational speed of the Cloud decreasing, it is certain that Edge Computing only benefits the IoT to Cloud Architecture.
To give an example of Edge devices in the consumer domain, if our refrigerator were enabled with an internet connection and some data processing power, we could have our fridge know the type of food that it contains. It could also know when milk is going to run out based on consumption patterns and with its enabled internet connection and an inbuilt AI system, order a litre of milk by itself using this internet.
Speaking of Edge devices in the manufacturing and industrial sector, shop floor machines and PLC’s could be directly interfaced with edge devices. These Edge devices are critical for industries that use vast amounts of data and require immediate reaction times. The IoT to Cloud architecture is very inefficient and slow. It allows for latency in the transfer of data to the cloud due to the sheer size of data that is collected at the source.
Like Edge Computing, there is a term called “Fog Computing”.
Coined by Cisco, Fog Computing is similar to Edge Computing. Fog Computing believes in the processing, storage and networking of data between the Data Source and the Cloud. Thus Edge computing essentially is a sub division of fog computing.
For the industry, the main purpose of an Edge Gateway is to provide connectivity options to data gatherers like Sensors & Actuators. There are many industrial protocols followed like Bluetooth, Modbus, Bacnet, Profinet, OPC UA, etc.
To fulfill the need of a decision maker, the Edge gateway needs to be a programmable one, thus the term Programmable Automation Gateway Controller. In addition to running local processing and diagnostic applications, monitoring systems and custom softwares, the Edge device should store data by itself.
This allows it to be completely autonomous from the network. By running local applications and storing the data at the same time, the Edge device can operate independently, in a disconnected manner.
Thus If no internet connection is required, the processed data can be viewed locally as well. The Cloud only features as a Software to enable further analytics and poses as a customized GUI platform to view the data.
The Embedos Edge
The Flagship device of Embedos Engineering LLP is its own Edge device. It has a plethora of options and applications and is developed as a programmable automation gateway controller. The Emedos Edge device can replace the traditional PLC to perform as a Logical Controller.
The same device can be used to process the acquired data and behave like a gateway to transmit this computed data to a Cloud storage for further visualization and analytics. Applications like Protocol Converson, Programmable Controller, Data Logging, Industrial VPN are some of the options offered by the Embedos Edge programmable gateway controller.