Edge computing is a model of distributed computing that relocates computation and data storage near the data sources. This should reduce bandwidth usage and speed up reaction times. Edge computing, a type of distributed computing that is topology and location-sensitive, is an architecture rather than a particular technology.
The concept of edge computing was first introduced in the late 1990s when content distributed networks (CDNs) were developed to deliver web and video material from edge servers placed near users.
The first commercial edge computing services appeared early in the 2000s as a result of these networks’ evolution to host applications and application components on edge servers. These applications included dealer locators, shopping carts, real-time data aggregators, and ad insertion engines. In an episode of ACM SIGCOMM’s Networking Channel commemorating the 25th anniversary of the creation of CDNs.
The story of how CDNs established the first edge networks that later developed into the first edge computing service in 2002, the name itself coined as a trademark, was told.
Types of Edge Computing
The newest step in distributed computing is edge computing. However, only some are aware of what it is or that there are various kinds of Edges.
In a nutshell, edge computing facilitates connectivity between a wide range of devices (typically referred to as User Equipment or UE in 5G terminology) and the telecom data center’s core by bringing computing closer to the end user. The main enabler is the round-trip communication latency between the UE and computer servers, necessitating a highly distributed design.
The closeness to the end device and round-trip latency to the data center’s core provide a simple way to comprehend the various kinds of edge computing. Since latency is the main determining element, edge computing can be broadly divided into the following categories:
- Internet of Things Edge
- Network Edge
- Access Edge
- On-Premise Edge
Internet of Things Edge
There is typically less than 1 nanosecond of expected latency at this Edge. This includes almost every gadget with internet access or a link to a private or public network. The gadget could be a basic sensor that transmits information about its surroundings or a smart device with data processing capabilities, like a smartphone.
Retail kiosks, factory sensors, cameras, connected autos, connected streetlamps, remote surgery equipment, drones, smart parking meters, etc. are a few examples.
Network Edge
The normally expected latency at this Edge is between 10 and 40 milliseconds or less. Before connecting to the centralized data center, which can cover a wide range of areas, On-premises edge, data from various IoT edges, and catering to the individual region, Access edge, needs to be aggregated.
The Next Generation Central Office (NGCO) design seeks to meet Network Edge’s needs. Wireline Fixed Access Edge, which offers broadband services like internet services, VoIP, IPTV, etc., is another illustration of this type of edge. The line dividing different device types at this Edge is blurring with the introduction of Fixed Mobile Convergence due to architectural convergence.
Access Edge
The typical latency expectation at this Edge is between 10 and 40 milliseconds. Previously available only as a fixed-function device, the classic Radio Access Network (RAN) has now decomposed and operates as a collection of virtual functions in software on commercially available servers. The key link connecting wireless devices to the main telecom operator network is the radio access network or RAN.
Ideas like virtual RAN (vRAN) and industry projects like O-RAN and Open-RAN have enabled managing these virtualized deployments using interfaces comparable to managing any other edge device. By utilizing Continuous Integration (CI)/Continuous Deployment (CD) frameworks that leverage DevOps paradigms, the move towards cloud-native instantiation of RAN function is simplifying the life cycle management of a large number of RAN deployments.
On-Premise Edge
At this Edge, latency expectations are typically lower than 5 milliseconds. To store, analyze, process, and reply to pertinent requests utilizing the data, one needs a method to aggregate data from several devices at the IoT Edge. On-Premise Edge helps localize data processing and reduces processing time by delivering computational resources locally. For additional requests, the devices at this Edge typically connect to a network edge or data center.
Huge businesses, huge retail operations, industrial production facilities, etc. benefit from these types of installations since they may process data close to its source while maintaining their unique rights to the required gear.
One type of equipment that combines a firewall, WAN optimizer, and router into a single piece of installed Premise equipment is Universal Customer Premise Equipment (u-CPE).
Use Cases of Edge Computing
Every industry can use edge computing. Now let’s look at some of the sectors where edge computing can be most helpful.
Healthcare
Regardless of the caliber of the Internet connection, healthcare software needs real-time data processing. The system must be able to quickly and accurately obtain a patient’s medical history. Edge computing, like autonomous vehicles, delivers a quick response from the server. Because it is located immediately on the local network and is, therefore, able to operate online.
Remote Drilling Platforms For Oil
Some sectors make use of software that requires little or no bandwidth. In these circumstances, data synchronization is rather challenging. Edge computing offers a solution if external variables, geographic location, or accidents can interfere with the Internet connection.
Safety
When a prompt security response is required, edge computing architecture is a better choice than traditional cloud solutions. The queries are handled directly at the network, bypassing a data center. It enables security companies to quickly respond to threats and foresee risks in real time.
Manufacturing
Edge computing allows manufacturers to manage large networks and handle several data streams at once. Edge computing will offer quick connections between all devices at all points of the network if the industrial equipment is dispersed over several sites. Once more, the data stream is independent of the Internet connection’s quality.
Wireless Speakers
Speakers should begin processing user input as soon as possible to carry out specified tasks. Once more, they should be unaffected by the bandwidth quality. Edge computing offers reliable data storage and quick user command execution.
Advantages of Edge Computing
Enhanced Performance
When consumers attempt to use centralized hosting platforms or centers-hosted programs and data, delays may result. The process of requesting data from these data centers may take longer when there are issues with internet connectivity.
Lowers Operating Expenses
Moving data around on cloud hosting services is one of the things that businesses spend a lot of money on. Due to the reduced requirement to transfer data to the cloud, organizations using edge computing spend less on operating expenditures.
Increases Data Security And Privacy Protections
In the world of IT, privacy protection and data security are hot topics. Edge computing provides a higher level of data security and privacy protection because data is processed locally rather than originating from centralized servers.
Yet, this does not imply that edge computing devices are completely safe. In no way. In contrast, as edge computing only creates, processes, and analyses the set of data required at any given time, other data that would compromise privacy in the event of a hack are not tampered with.
Less Expensive Transmission
Edge computing can result in significant cost savings due to decreased bandwidth, in addition to the potential for streamlining cloud security methods. Edge computing consequently uses less bandwidth in the data center.
Data centers can conserve bandwidth capacity and avoid spending money on pricey upgrades to the current cloud storage capabilities by storing fewer files in the cloud and processing more tasks locally.
Versatility and Scalability
Data must be transmitted to a central data center in a cloud computing system. The cost of expanding or upgrading this data center may be high. On the other side, you can scale your own IoT network using the edge without worrying about storage.
Improves Resilience And Reliability
Edge computing allows data to be collected and processed with little to no difficulty even when there are issues with insufficient internet connections. Furthermore, the ecosystem’s edge devices will continue to operate normally even if one of them malfunctions, increasing the dependability of the entire connected system.
Helps In Compliance Requirements And Meeting Regulatory
Meeting regulatory and compliance criteria may be more difficult when data is kept and handled by multiple data centers or hosting providers. This is due to the unique privacy and regulatory requirements that each data center has.
Supports Applications for AI/ML
There is no doubting the growing importance of artificial intelligence (AI) and machine learning (ML) in contemporary computing. However, AI/ML applications operate by obtaining and analyzing massive amounts of data. Connectivity and latency issues may occur when the data is kept on a centralized server.
Disadvantages of Edge Computing
Security
Edge computing reduces the amount of data that needs to be safeguarded in data centers, which enhances security, but it also causes security issues at each localized point of the edge network. Furthermore, not all edge devices have the same internal authentication and security measures, which leaves certain data more vulnerable to hacks.
Loss of Data
Although sorting through all of the data in a cloud data center can be time-consuming, you can rest easy knowing that it will be accessible when you need it thanks to the data’s central storage. Although edge computing operations reduce storage costs and space requirements, by accident an edge device could misinterpret or even destroy sensitive data.
Costs of Infrastructure
Whether you choose to invest in massive, global clouds or dispersed edge devices for your computing requirements, networking technology is always a significant financial commitment. While spending more on an edge network can reduce the cost of data center bandwidth, the plan has its own set of startup and maintenance costs for edge devices.
For maximum performance and local storage requirements, edge devices may need additional hardware and software, and costs can rise quickly when they’re dispersed across several local geographies.
Maintaining Hardware
Edge computing gives you more control over how your data is processed and stored, on the one hand. On the other hand, the business must be in charge of keeping an eye on and fixing local servers, making maintenance purchases, and handling outages. With cloud computing, the service provider receives complete outsourcing of this work.
