As the Internet of Things (IoT) continues to grow, so does the amount of data that needs to be stored, processed, analysed, and transmitted.
In its infancy, edge computing sought to lessen the burden on network bandwidth by storing data closer to its point of origin rather than in a central data warehouse or in the cloud. Recent developments in low-latency applications, such self-driving cars and video analytics using several cameras, are propelling the idea further.
Because 5G allows for speedier processing of these cutting-edge, low-latency use cases and applications, its continued global rollout has a direct bearing on edge computing.
What is edge computing?
Edge computing, according to Gartner, is “a component of a distributed computing topology in which information processing is positioned close to the edge—where things and people produce or consume that information.”
Edge computing, in its simplest form, is the practise of bringing computer and data storage closer to the devices that are gathering it, as opposed to relying on a central site that may be thousands of miles away. This is done to ensure that latency issues do not impact the performance of an application and prevent data, especially real-time data, from becoming corrupted. Furthermore, businesses can cut costs by processing data locally rather than sending it to a centralised or cloud location.
Consider factory floor monitoring equipment or a video camera streaming live from a faraway office via the internet. Transmission of data from a single data-generating device is straightforward across a network; but, as the number of data-generating devices increases, so do the potential for collisions and delays. Think about hundreds, maybe thousands, of cameras sending live footage in place of just one. The latency will lower the quality, and the bandwidth prices can be sky high.
To address this issue, many of these systems are beginning to rely on edge-computing hardware and services as their primary means of processing and storing. For instance, an edge gateway can filter out irrelevant information from data collected by an edge device before sending it back to the cloud. For real-time applications, it can also relay data back to the edge device. (Also see: Edge gateways are a robust and adaptable IoT enabler)
What is the relationship between 5G and edge computing?
While previous networks, such as 4G LTE, may be suitable for deploying edge computing, this may not be the case with 5G. That is to say, businesses will not get the full benefits of 5G until they establish an edge computing infrastructure.
According to Dave McCarthy, IDC’s research director for edge solutions, “by itself, 5G reduces the network latency between the endpoint and the mobile tower, but it does not address the distance to a data centre,” which might be troublesome for latency-sensitive applications.
Prof. Mahadev Satyanarayanan of the Computer Science Department at Carnegie Mellon University, who co-authored the article that paved the way for edge computing back in 2009, concurs. “What difference does it make, even if zero milliseconds on the last hop, if you have to go all the way back to a data centre across the country or other end of the world?”
Although there will be a growing dependence on 5G wireless as more and more 5G networks are rolled out, wired and even Wi-Fi networks can still be used to build edge computing infrastructure if necessary. However, 5G’s superior speeds make it more likely that edge infrastructure will employ this technology, especially in unserved rural areas.
How does edge computing work?
Although the edge’s physical architecture can be complex, the concept is simple: for faster and more reliable processing and operations, client devices hook up with an edge module that’s physically close by. IoT sensors, employee laptops, smartphones, surveillance cameras, and even the break room microwave oven can all be considered edge devices.
The edge device may be a self-driving mobile robot or a robot arm in a manufacturing scenario. It can also refer to a high-tech surgical apparatus that allows surgeons to operate from afar. An edge gateway is a device at the very periphery of a distributed computing network. The components may also be referred to as edge servers or edge gateways, depending on the context.
Businesses that want to use a private edge network will need to think about edge gateways or servers, even if many of these devices will be deployed by service providers to support edge networks (Verizon, for example, for its 5G network).
How to buy and deploy edge computing systems
The acquisition and installation of an edge system might take several forms. On the one hand, a company may prefer to take on most of the work themselves. Choosing an appropriate network architecture for the use case, purchasing management and analysis tools, and deciding on edge devices (likely from hardware vendors like Dell, HPE, or IBM) are all necessary steps.
While this is a lot of work and would require a lot of in-house IT experience, it may be a viable solution for a large firm that wants a fully tailored edge deployment.
On the opposite end, suppliers in specific industries are heavily promoting edge services that they will administer on your behalf. If a company decides to take this path, all it has to do is hire a vendor to set up the necessary equipment, software, and networks, and then pay a monthly or annual price for access and upkeep. Products and services in this category include those provided by GE and Siemens, two industry leaders in the IIoT space.
While this method has the benefit of being simple and trouble-free to deploy, it is possible that highly managed services are not always readily available.
What are some examples of edge computing?
In the same way that the number of internet-connected devices keeps growing, the variety of scenarios where an organisation might benefit from using edge computing to take advantage of its low latency or cut costs is also growing.
Edge-enabled sensors to provide detailed imaging of crowds in public spaces to improve health and safety; automated manufacturing safety, which leverages near real-time monitoring to send alerts about charac… Verizon Business, for example, describes several edge scenarios, such as end-of-life quality control processes for manufacturing equipment; using 5G edge networks to create popup network ecosystems that change how live content is streamed with sub-second latency;
There will be a sizable gap between the hardware needs of the various deployment models. For instance, industrial customers would prioritise dependability and low-latency, necessitating ruggedized edge nodes that can function in the harsh environment of a factory floor and dedicated communication lines (private 5G, dedicated Wi-Fi networks, or even cable connections).
Users in the connected agriculture sector, on the other hand, will still need a rugged edge device to handle outdoor deployment, but the connectivity piece may look very different; for example, low-latency may still be necessary for coordinating the movement of heavy equipment, but environmental sensors are likely to have both higher range and lower data requirements. In such a case, a low-power wide-area network (LP-WAN) link, such as Sigfox or something similar, may prove to be the most suitable option.
Other applications have their own unique difficulties. Edge nodes can serve as a central hub for a variety of in-store functions for retailers, including the consolidation of POS data with targeted promotions, foot traffic tracking, and more.
It’s possible to keep things straightforward by providing Wi-Fi to all devices within the building, or it could get more complicated by using Bluetooth or another form of low-power connectivity for things like traffic monitoring and advertising, with Wi-Fi saved for things like the register and automated checkout.
What are the benefits of edge computing?
Expense reduction is a primary motivation for many businesses to use edge-computing. Organizations who relied heavily on the cloud may have experienced unexpectedly high bandwidth bills and are now looking for a cheaper alternative. You could find a use for edge computing.
However, the capacity to process and store data more quickly is becoming increasingly the main benefit of edge computing, since it enables more effective real-time applications that are crucial to businesses. Edge computing has reduced the processing time required to run a facial recognition algorithm from a smartphone scanner all the way to the cloud. The algorithm could be executed locally on the edge server or gateway, or even on the smartphone, thanks to the edge computing architecture.
Fast processing and response are essential for many emerging technologies, including virtual and augmented reality, autonomous vehicles, smart cities, and even building-automation systems.
Edge computing and AI
In response to the growing need for edge processing, hardware manufacturers like Nvidia are hard at work creating solutions, such as modules with integrated AI functionality. Developers of robots, autonomous devices, and next-generation embedded and edge computing systems will benefit from the company’s newest product in this field, the Jetson AGX Orin developer kit, a tiny and energy-efficient AI supercomputer.
When compared to the previous system, Jetson AGX Xavier, Orin provides an 8x improvement, with 275 TOPS. Improvements in deep learning, vision acceleration, memory bandwidth, and support for several sensor types are also included.
While current AI algorithms often run on cloud-based services, the development of AI chipsets that can perform the necessary processing at the edge will lead to an increase in the number of systems able to handle such workloads.
Privacy and security concerns
Concerns about the safety of data at the edge arise when its processing is delegated to a wide variety of endpoints that may not have the same level of security as more centralised or cloud-based options. IT must be aware of potential security threats and make sure that IoT systems are secure as the number of connected devices increases. Data encryption, access control measures, and VPN tunnelling are all part of this.
In addition, an edge device’s dependability might be affected by its need for different amounts of computing power, electricity, and network connectivity. This emphasises the significance of redundancy and failover management for edge devices processing data to guarantee proper data delivery and processing in the event of a node failure.