Photo Credit: Reddit User, Beanful (https://www.reddit.com/user/Beanful)
The Internet of Things has opened the proverbial floodgates (of data) and is redefining the high-water mark daily. Massive quantities of information are being transferred between distributed IoT “edge” sensors and device networks with centralized computing and cloud computing resources. In this regime, low latency data processing, adequate data privacy, and robust connectivity are often deficient, inconsistent and pose significant challenges. With the advent and increasing demand for modern IoT edge applications we have found a new way to “max-out” our capacity and budgets to handle the data deluge.
In this article we’ll define exactly what Edge Computing is, and share some insights on this niche. We will also examine how it is being utilized to improve performance and reduce the costs, limitations, and dependencies associated with centralized processing and cloud computing for IoT.
In a nutshell, Edge Computing entails the shifting of data analysis and processing to the IoT edge device microprocessor itself. This is in contrast to the common practice of transferring raw, unprocessed data streams from the IoT device to a cloud or centralized processor for analysis. With edge computing, we essentially perform the dirty work of filtering, analyzing and cleaning-up data to key indicators and actionable intelligence or stimuli in situ. According to Network World Magazine, “This on-device approach helps reduce latency for critical applications, lower dependence on the cloud, and better manage the massive deluge of data being generated by the IoT.”
Furthermore, by processing and analyzing IoT edge data prior to transmission to the cloud or network, we can drastically reduce the demand for bandwidth and the problems associated with data integrity over intermittent wireless connections. This may be a non-issue for small installations of infrequently polled sensors, but for large deployments of distributed IoT or edge devices with a higher sample rate, these bandwidth and connectivity constraints can be a considerable and costly predicament. This issue becomes even more pronounced as emerging wireless services such as Narrowband IoT (NB-IoT) come online, which have greater economies of scale but with a significant trade-off in the form of reduced bandwidth and link rates when compared to Cat 1 4G/LTE and the upcoming 5G options.
Now imagine a network topology combining short-range wireless IoT edge nodes to an occasional long-range, wide-area, cellular-enabled Edge Processor capable of AI, signal processing, data storage, communications and data encryption. Such a processor can be utilized to manage the local area network of edge devices as well as to coordinate with cloud or centralized computing resources. I bet you can now really see the power and efficiency of Edge Computing! Such architectures have the potential to improve latencies, bandwidth usage, energy consumption and data security and with modest cost. The possibilities are changing the landscape drastically. Edge computing is the concept behind derivative standards such as Fog Computing, which serve as alternatives to our more centralized or cloud computing IoT infrastructures.
At NetBurner, we design our embedded Systems-on-Modules with capabilities to perform complex processing and analysis on-device. Our IoT optimized Real-time Operating System (RTOS), reliable TCP/IP stack and rugged embedded processing (ARM Cortex and NXP Coldfire options) are truly a formidable and efficient edge computing and networking solution for distributed IOT. Integrated SSL/TLS also allows for secure data transfer from the origin. As always, we are excited to enable our community to develop applications that are on the forefront of this exciting direction in networking and computing.
So how do you get involved in edge computing? Begin by assessing what raw data can be processed on site and which data or communications can be kept within the LAN. Design-in as much storage, intelligence and processing capability as feasible on the edge device. Determine what information, value or thresholds trigger a transmittal of actionable or critical data over the internet or network. Then encrypt and output the filtered or analyzed data only as necessary.
If you have questions or want to share your experience in using edge computing please comment below!
Here’s some of our suggested readings on the Edge Computing topic:
- Why edge computing is critical for the IoT – NetworkWorld Magazine – Quick reading opinion piece
- Edge computing: The state of the next IT transformation – ZDnet.com – Loaded with insight, marketing research, and helpful charts and links.
- A guide to Edge IoT analytics – IBM – Well organized and informative piece with useful diagrams