Edge computing

Data generated by IoT and other millions of distributed devices is processed, analyzed, and stored near the original point of generation. Latency, fluctuating bandwidth, and need for secure connectivity are the key motivators of edge computing

When cloud computing made an entry into the IT world, it became the holy grail for most organizations. It still continues to enchant people around the world. However, as the cloud craze is slowly wearing down, experts in the field are noticing the flaws with this cloud technology. There’s a new computing style that is complementing the cloud. It’s not hard to guess which is this new style: It’s an Edge Computing.

This computing style preaches that the data generated by IoT and other millions of distributed devices should be processed, analyzed, and stored near the original point of generation. Does this sound like the function of an on-premises data center to you? But it’s not. The data ultimately reaches the cloud due to its massive storage capabilities. However, every single byte need not be sent.

Why Edge is about to flip things on its head

Edge certainly is a very topical area at the moment and is going to be for the foreseeable future. Most articles and conversation talk about the benefits that Edge will bring from low latency 5G networks, AI, cloud computing, storage and so on, but not many of them talk about what it is going to take to get there.

If you take a look at what is being proposed, and I mean a really good look, you may notice that a lot of the “traditional” methods we use today for cloud, networking. And security may not be up to the job and therefore, this need to be updated. Thus we need a new approach.

So why edge computing? Well at high-level, it’s a huge undertaking of decentralized resources. Its going to take a number of vendors, providers and solutions to deliver the end results .We need to have open and interoperable platforms with security and collaboration at its heart.

Edge locations are not just a scaled down version of a data center, they are critical infrastructure that will support a million connected devices and IoT censors. As more and more Edge locations are rolled out the attack surface is increased.

With each Edge location supporting connected devices providing services for health, education, local authorities, businesses, latency sensitive vehicle control and so on, any impact to the infrastructure would have a dramatic effect on service delivery and could potentially be life threatening. Many Government agencies see this as a threat. We need to rethink about Edge scale.

There is no doubt that edge computing and cloud computing are intimately tied.

Is an edge part of a cloud?

Edge devices can contribute to a cloud, if the storage and computing capabilities provided by those devices at the endpoints of a network are abstracted, pooled, and shared across a network — essentially becoming part of a larger cloud infrastructure.

Edge computing is not part of a cloud. What makes edge computing so useful is that it is purposefully separate from clouds and cloud computing.

Here’s how we see it:

• Clouds are places where data can be stored or applications can run. They are software-defined environments created by datacenters or server farms.
• Edges are also places where data is collected. They are physical environments made up of hardware outside a datacenter.
• Cloud computing is an act; the act of running workloads in a cloud.
• Edge computing is also an act; the act of running workloads on edge devices.

An edge (location) is not the same thing as edge computing (action). Collecting data at the edge of a network and transferring it to a cloud with minimal (if any) modification is not edge computing — it’s just networking.

But, if that data is collected and processed at the edge, then it’s edge computing.

Edge computing is separate from clouds for 2 main reasons:

1. Time sensitivity. The rate at which a decision needs to be made doesn’t allow for the lag that would normally take place as data is collected by an edge device, transferred to a central cloud without modification, and then processed before a decision is sent back to the edge device for execution.
2. Data volume. The sheer volume of data collected is too much to send — unaltered — to a cloud.

Why is it Popular?

Edge computing is becoming more popular for a variety of reasons:

• The use of mobile computers and “internet of things” devices is growing, as is the cost of hardware.
• The correct operation of Internet of Things devices necessitates a fast response time and a large amount of bandwidth.
• Cloud computing is a centralised method of computing. Massive amounts of raw data must be transmitted and processed, putting a strain on the network’s bandwidth.
• Furthermore, the ongoing transfer of vast amounts of data back and forth is beyond realistic cost-effectiveness.
• Processing data on the spot and then transferring valuable information to the centre, on the other hand, is a significantly more efficient method.

Types Of Edge Computing

1. Device Edge

It is also known as a nano DC and comprises one or more microservers. It would be limited in processing power and would only have one or a few customizations. This segment’s databases are unlikely to be installed on a rack. We should be able to operate without the use of refrigeration. They’re also in locations that aren’t normally associated with data centres. Warehouses, wind generators, and weather-resistant structures are examples. They’re directly present near IoT sensors, so latency, bandwidth, and communication concerns aren’t an issue. The disadvantage is that these little gadgets can only use a limited amount of power and have limited capabilities.

2. Cloud Edge

It primarily refers to huge data centres run by cloud providers like AWS, Azure etc. This might contain VMware Cloud on AWS as well as other cloud or service providers. The cloud’s main characteristics are that it is centralised and that it operates at a large scale. The disadvantage is that infrastructure availability is extremely high as there is no assurance that network connection to sensors or processors at the edge will be available, and there will be a lot of latency. Internet activity both to and from the cloud is almost certainly costly.

3. Compute Edge

It’s a modest data centre with anywhere from a few to a lot of server racks. They are frequently placed near or close to IoT equipment, and they may be required for local law enforcement purposes. The idea is that these data centres have standard servers installed on racks, as well as ventilation and other amenities. One benefit would be that network latency at the edges would be lower than in the cloud. As a result, network bandwidth should increase while remaining more efficient.

Edge computing vs cloud computing

Firstly, it’s important to clear one thing that the cloud and edge computing are two different technologies. They are non-interchangeable technologies that cannot replace one another. Edge computing is used in process of time-sensitive data, while cloud computing is used in process of data which is not time-driven.

Edge computing is beneficial to specialize and intelligent devices. While these devices are alike PCs, they are not meant to regular computing devices designed to perform multiple functions. These devices are intelligent and respond to particular machines in a very specific way. However, this specialization becomes a drawback for edge computing in certain industries that require immediate responses.

The process of edge computing is very different from cloud computing because it needs time. Sometimes it needs up to 2 seconds to relay the information to the centralized data center and it delays the decision-making process. The signal latency can lead to the organization incurring losses, hence organizations prefer edge computing to cloud computing.

Lets take an example of a machine whose functionality is very crucial for an organization. A delay in the machine’s decision-making process due to latency would result in losses for the organization. In such cases, organizations will prefer edge computing because smart devices with computation power are placed on the edge of the network. The device monitors a pre-defined metrics set for tolerance levels, if the metrics are outside of the prescribed tolerance, a warning signal is issued as soon as the machine reaches the failure level, resulting in the shutdown of the machine within microseconds to avoid further losses.

Real Life Examples

1.Fleet Management

Saltatory edge computers are often deployed in fleet vehicles, allowing organizations to intelligently management their vehicle fleets. The edge PCs can tap into the CANBus network of vehicles, collecting a variety of rich information, such as mileage per gallon, vehicle speed, on/off status of vehicle, engine speed, and many other relevant as well as important information. Moreover, edge computers can collect more data from cameras and sensors deployed on the vehicle.

All of this collected data can be leveraged by fleet companies to improve the performance of their fleet, as well as to reduce the operation costs of the fleet. Rugged edge computers are hardened to withstand exposure to challenging environmental conditions that are commonly found in vehicles. Such challenging conditions include exposure to shock, vibration, dust, and extreme temperatures.

2.Predictive Maintenance

Edge computers are often used by organizations because they can gather information from various sensors, cameras, and other devices. They can use the collected information to determine when components or certain machinery fails. Predicting when a machine will fail allows factory operators to perform maintenance on the machine or replace a component before a failure occurs during normal machine operation, it saves organizations costs from lost productivity and missing delivery times and expectations.

This is asundar from the traditional model where organizations conducted routine diagnosis and inspections, which is labor intensive and costly. Moreover, with the traditional model it is difficult to perform maintenance before a component or machine fails. With predictive maintenance, organizations can intervene and maintain machinery and equipment before the failure ever occurs.

3.Kiosk Machines

The computers are often used to power interactive kiosk machines such as the ones you often pass by or use while you’re at the airport or supermarket. Interactive kiosk machines utilized rugged edge computers because they have a rich I/O that allows the system to connect to the various peripherals found on kiosk machines.

As well as the hardened design of the system that enables the deployment of kiosk machines indoors and outdoors. Hardened edge computers keep kiosk machines only 24/7 regardless of challenging environmental conditions. Rugged edge computers deliver the performance necessary to power kiosk machines while maintaining power efficiency.

4.Remote Monitoring Of Oil & Gas Assets

The edge computers can be deployed in oil and gas fields where the temperatures often reach and slightly exceed 50°C without running into thermal issues such as thermal throttling. Rugged edge computing solutions can be used to monitor the large number of assets deployed in the fields to ensure that oil and gas production facility runs smoothly without interruptions to field operations.

Rugged edge computers are used to monitor the performance and control the oil production process. For example, rugged edge computers are used to monitor the fuel flow in pipelines, providing oil production facilities with invaluable insights into the flow metrics and pipeline performance so that operators can quickly uncover and respond to any critical issues that may arise.