What Is Edge Computing?

01What Is Edge Computing?

Edge computing means moving computation to the edge of the network, to the point where data is born or consumed. That point can be a small computer next to a sensor in a factory, a store's register, a cell tower or a server in the user's city.

In the classic cloud model, data leaves the device, travels a long way to a large data center, is processed there and the result comes back. Edge computing shortens this trip: part of the work is done on the spot, and only the necessary summary or result is sent to the cloud.

The concept is not new; content delivery networks have worked on a similar logic for years. What is new is that not only files but also code and AI models can now run close to the user.

02How Does Edge Computing Work?

It helps to think of an edge architecture as three layers:

• Device layer: Sensors, cameras, smartphones and registers produce the data. Some devices can handle simple processing themselves.
• Edge layer: An edge server or gateway located close to the devices filters and analyzes the data and makes instant decisions.
• Cloud layer: Long-term storage, large-scale analysis, model training and central management happen here.

A simple example

A camera on a production line captures dozens of images per second. Sending all the images to the cloud is both slow and expensive. A device at the edge inspects the images on the spot and sends a record and an alert to the cloud only when it detects a defective product. The line can be stopped instantly without waiting for a response from the cloud.

03Edge vs. Cloud

The difference between edge and cloud is not a difference between two rival technologies but a division of labor. Where each job is done is decided by latency, data volume and privacy needs. We covered the basics of the cloud separately in our article on what cloud computing is.

04Advantages of Edge Computing

Speed and low latency

The laws of physics do not change: the farther data travels, the later it comes back. Low latency is critical in areas such as autonomous systems, industrial control, live streaming and gaming. On the web side, serving a page from a point close to the user shortens the time to first response.

Cost and bandwidth

Sending only the meaningful part instead of moving all raw data to the cloud lowers data transfer and storage costs. For businesses that use many cameras or sensors, the difference can be significant.

Privacy and data locality

Sensitive data, such as camera footage or health data, can be processed on site and only an anonymous result sent out. This approach makes it easier to meet privacy and data locality requirements; however, it is not enough on its own for legal compliance, so check the current requirements of regulations such as KVKK (Turkish data-protection law, comparable to GDPR) separately.

Resilience

When the internet connection drops, a cloud-based system may stop working. A system running at the edge keeps its basic functions going and syncs the data to the cloud when the connection returns.

05The Internet of Things and Edge

The Internet of Things (IoT) is the most natural use case for edge computing. Moving all the data produced by millions of small devices to the cloud is impractical; the data is pre-processed close to the device.

In agriculture, soil moisture and weather sensors can manage irrigation on site. In cold chain logistics, a device in the vehicle notices a temperature deviation instantly and raises an alert. In hotels, room sensors manage energy consumption; integrating this data with hotel software lets operations be monitored from a single screen.

In retail, in-store cameras can analyze shelf stock on the spot, and checkout systems can keep selling even when the internet is down. What all these scenarios have in common is that the instant decision is made on site and the long-term analysis is done in the cloud.

06Edge for Websites

In the web world, the best-known form of edge computing is the content delivery network (CDN). Images, style files and scripts are copied to servers at different points around the world, and the visitor is served from the nearest one. We described how it works in our article on what a CDN is.

In recent years, CDN providers have started to offer not only files but also code execution. Routing rules, language and region detection, simple personalization or security checks let a request be processed at the edge without ever reaching the origin server.

Even so, an edge layer is not essential for every site. For a corporate site whose visitors mostly come from one region, a well-configured server, correct cache settings and optimized images are often enough. You can find the differences between hosting options in our article on what hosting is.

07Challenges of Edge Computing

Edge computing is not a cure-all. Before deciding on an architecture, you need to take these challenges into account:

• Scattered management: Instead of a single data center, dozens or hundreds of points must be updated and monitored.
• Physical security: Edge devices often sit in the field, not in a protected data center; there is a risk of theft and tampering.
• Limited resources: Edge hardware has limited processing power and storage; heavy jobs still go to the cloud.
• Data consistency: Merging data processed at different points correctly at the center takes careful design.

08Does Your Business Need Edge?

Ask three questions to decide: Do you have a job that needs a response in under a second? Is the volume of data you move to the cloud creating a cost problem? Do you have a process that cannot be allowed to stop when the connection drops? If you answer “yes” to none of these, do not rush to invest in edge computing.

For most small and mid-sized businesses, the first contact with edge computing is the CDN and cache layers they already use without even realizing it. Measuring the speed of your site and your e-commerce infrastructure and improving the server, cache and CDN layers together is the subject of our speed and performance management service.

For a level-headed look at future technologies, you can also read our article on what a quantum computer is. If you need a custom system that collects data from devices in the field, we can plan the right architecture together with the Aurora solutions and mobile-friendly design team.