What is edge calculation and why is it growing in importance? – Computer world

With the development of IoT devices and the advent of high-speed 5G wireless connectivity, the placement of computing, storage, and analytics close to where data is generated is becoming an argument in favor of external computing.

Edge computing changes the way you store, process, analyze, and transport data from billions of IoT and other devices.

The primary purpose of edge calculation was to reduce bandwidth costs associated with transporting raw data from the source to the enterprise data center or cloud. Recently, this concept is driven by an increase in real-time applications that require minimal delay, such as autonomous vehicles and multi-camera video analysis.

Also check:

The ongoing global application of the 5G wireless standard encompasses extraneous calculations, as 5G enables faster data processing in these new and lower latency applications.

What is edge calculation?

Gartner defines extraneous computing as “part of a distributed computational topology where information processing is located close to the edge – where objects and people produce or use that information.”

At the simplest level, computing and memory move closer to the devices on which it is stored, rather than relying on a central location that can be thousands of miles away. As a result, there are no problems with delays, especially in real-time data, which can affect the operation of the application. In addition, because data processing is done locally, businesses can save money by reducing the amount of data that needs to be sent to a central location or to the cloud.

Think of devices that control production equipment in a factory, or a video camera connected to the Internet that transmits live video from a remote office. Although a single device that produces data can send data over the network fairly easily, problems also arise when the number of devices that transmit data increases. Instead of just one live video camera, you can shoot this number on hundreds or thousands of devices. Not only will quality suffer from latency, but bandwidth costs can be astronomical.

Edge computing devices and services help solve this problem by providing a local source for data processing and storage for many of these systems. For example, an external gateway can process data from an external device and then send the relevant data back only through the cloud. It can also send data to an external device when real-time applications are needed.

What is the connection between 5G and external computing?

Although non-5G networks (eg 4G LTE) can perform external calculations, the opposite is not necessarily true. In other words, businesses can’t really take advantage of 5G without an external computing infrastructure.

“5G in itself reduces network latency between the endpoint and the mobile tower, but does not solve the data center problem, which can be a problem in delay-sensitive applications,” said Dave McCarthy, IDC’s director of external strategy research.

Mahadev Satyanarayanan, a professor of computer science at Carnegie Mellon University in 2009, is a co-author of an article that sets the direction for extraneous calculations. “If you have to go back to a data center in the country or all over the world, what’s the difference, even if it’s zero milliseconds in the last leap.”

As more 5G networks are implemented, the connection between remote computing and the 5G wireless network will continue, but companies can still deploy external infrastructure with different network models, including wired and even Wi-Fi, if needed. However, with the high speeds offered by 5G, especially in rural areas where wired networks are not available, external infrastructure is more likely to use 5G.

How does edge calculation work?

The physical architecture of external devices can be complex, but the basic idea is that client devices are connected to a nearby peripheral module for faster processing and smoother operation. Edge devices can include IoT sensors, work laptops, their latest smartphones, security cameras, and even a microwave connected to the Internet in an office conference room.

An external device in an industrial environment can be an autonomous mobile robot, a robotic arm in a car factory. In healthcare, it can be a high-level surgical system that allows doctors to work from a distance. External gateways themselves are considered external devices within the external computing infrastructure. Terminology is different, so you may come across the terms external servers or external gateways.

Although many gateways and external servers are hosted by service providers who want to have an outsider (for example, Verizon for a 5G network), businesses that want to host a private external network will also need to consider this device.

How to buy and apply edge systems How to buy and apply edge systems?

How you purchase and apply an edge system can be very different. On the one hand, the company may want to carry out most of the process on its side. This probably includes choosing devices from hardware vendors such as Dell, HPE, or IBM, designing a network that suits your needs, and purchasing management and analysis software.

This is a lot of work and requires significant IT experience, but it can still be an attractive option for large organizations that want to fully personalize outsourcing.

On the other hand, retailers in the private sector are increasingly offering advanced services that they will manage for their customers. An organization that wants to go this route can simply ask the seller to install its equipment, software and network and pay for regular use and maintenance. IIoT offers from companies such as GE and Siemens fall into this category.

The advantage of this approach is that it is easy to implement and relatively hassle-free, but this type of high-level management services may not be available in every use.

What are some examples of edge calculation?

As the number of devices connected to the Internet increases, so does the number of cases where external computing can save a business or use it with extremely low latency.

For example, Verizon Business describes several boundary scenarios, including expired quality control processes for production equipment; use of 5G external networks to create pop-up network ecosystems that change the way live content is delivered with delays of less than a second; use of external sensors for detailed description of the crowd in public places to improve health and safety; automated production security that uses real-time monitoring to send alerts about changing conditions to prevent accidents; production logistics aimed at increasing efficiency in the whole process from production to transportation of the finished product; and creating accurate product quality models using digital twin technologies to gain insight into production processes.

The hardware required for different types of accommodation will be significantly different. Industrial users, for example, will focus on reliability and low latency, which require robust edge nodes and dedicated communication connections (private 5G networks, dedicated Wi-Fi networks, or even wired connections) that can operate in difficult factory floor conditions. ) to achieve your goals.

On the other hand, agricultural-related users will still need solid external devices that can handle outdoor placement, but the connection component may look completely different – low latency may still be required to coordinate the movement of heavy equipment, but environmental sensors are more likely. , both wider coverage and lower data requirements. Your best choice may be LP-WAN, Sigfox or a similar connection.

Other uses come with completely different challenges. Retailers can use external nodes, such as in-store data memory, to track pedestrian traffic and more functions by combining point-of-sale information with targeted promotions in a single store management application.

The connectivity component can be simple – built-in Wi-Fi for each device – or more sophisticated, dedicated to Bluetooth or other low-power connectivity and Wi-Fi hotspots and self-service checks that support traffic tracking and promotion services.

What are the advantages of edge calculation?

For many companies, cost savings alone may be the driving force behind the use of external calculations. Companies that initially use the cloud for many applications may have found bandwidth costs higher than they expected and are looking for a cheaper alternative. Edge calculation may be a suitable solution.

However, the biggest benefit of increasingly external computing is the ability to process and store data faster, allowing for more efficient real-time applications, which is important for businesses. Before applying external processing, a smartphone that scans a person’s face for facial recognition had to run a face recognition algorithm through a cloud-based service, which took a long time to process. In the Edge model, the algorithm can be run locally on an external server or gateway, or even on the smartphone itself.

Applications such as virtual and augmented reality, self-driving cars, smart cities, and even building automation systems require rapid processing and response at this level.

Edge computing and artificial intelligence

Companies such as Nvidia continue to develop devices that take into account the growing need for external processing, including modules with built-in artificial intelligence. The company’s latest product in this field is the Jetson AGX Orin development kit, a compact and energy-saving AI supercomputer designed for robots, autonomous machines and next-generation developers of external systems.

Orin provides 275 trillion operations per second (TOPS), which is eight times better than the company’s previous system, the Jetson AGX Xavier. It also includes updates for in-depth learning, vision acceleration, memory transfer capability, and support for multimodal sensors.

Artificial intelligence algorithms require a lot of computing power used in cloud-based services, but with the development of AI chips that can do external work, more systems will be created to perform these tasks.

Privacy and security issues

From a security perspective, peripheral data can be difficult to process, especially when processed by a variety of devices that are not as secure as centralized or cloud-based systems. As the number of IoT devices increases, IT needs to understand potential security issues and ensure the security of these systems. This includes data encryption, the use of access control methods, and possibly VPN tunneling.

In addition, different computing power, power, and network connection requirements of devices can affect the reliability of external devices. Therefore, it is important for external computing devices to manage backup and failure in order to ensure the correct transmission and processing of data in the event of a node failure.

Source: NetworkWorld

Leave a Comment