Edge computing is the next step forward beyond cloud computing. More organizations are looking to bring the cloud closer or distribute their computing resources across broader networks as data demands rise. This shift to the edge has many advantages, but it can also be a complicated transition.
The first step to making the most of the edge is understanding the different types of edge computing. Edge networks come in various shapes and sizes, each with unique advantages and disadvantages. Global spending on edge computing is expected to reach $176 billion this year, so learning which of these best fits your needs and goals is critical.
Also see: Best IoT Platforms for Device Management
Edge Types by Technology
There are many ways to classify edge technologies and networks. As a result, there’s no conclusive answer to how many types of edge computing exist or what those categories are.
One of the easiest ways to divide edge is by the systems and platforms on which the computing happens. Here are some of the most common categories by that definition.
The cloud edge is the first type of computing under this classification system. As the name implies, these networks more closely resemble traditional cloud computing. They use large data centers, but unlike conventional approaches, these centers are relatively close to end-users and often serve purpose-built applications.
These systems offer impressive latency improvements while maintaining the conventional cloud’s capacity. While 58% of cloud users in some scenarios could achieve latencies under 10 milliseconds with nearby edge data centers, only 29% can do the same with a traditional cloud data center.
However, these deployments can be more expensive and aren’t accessible to everyone, as they rely on available data centers in your area. Consequently, they’re best suited for larger enterprises or operations with high data demands nearby existing infrastructure.
The device edge is likely what most people think of when they think of edge computing. More organizations have noticed delayed long-distance transmission between colocation sites and have embraced these networks that bring computing processes closer to the data’s source.
Device edge networks distribute computing tasks across local devices such as phones, smart gadgets, and routers. This offers minimal latency while sacrificing capacity from these devices’ limited processing power. As a result, these networks typically serve specific use cases instead of being a general replacement for cloud computing.
Successfully implementing device edge networks requires understanding your goals and the capabilities of your accessible devices. These systems are best for simple, highly specialized applications like predictive maintenance, and these applications depend on the machines at hand.
Compute edge environments provide a middle ground between device and cloud edge networks. These may distribute computing tasks across small, purpose-built machines like the device edge but also take advantage of micro data centers (MDCs).
MDCs can come in various sizes, typically ranging between 50 to 400 kWh, and can have just one rack or host a few. They’re also often modular, making them more scalable and adaptable than conventional data centers. These technologies give compute edge systems a smaller footprint and more flexibility than the cloud edge but higher power than the device edge.
Compute edge networks are ideal for companies with various edge computing needs but may not have access to larger nearby data centers. Installing MDCs will cost more than device edge environments, but they’ll enable a broader range of use cases.
Also see: 9 Cloud Cost Optimization Strategies
The sensor edge is on the other end of the spectrum from the cloud edge. These systems perform computing tasks as close to the source of the data as possible, using Internet of Things (IoT) devices to execute basic computations at the sensor level.
There are more than 12 billion active IoT endpoints in the world today. Each of these devices represents a data collection point, and sensor edge computing moves some analytics processes to these endpoints to maximize IoT performance. However, since these endpoints typically have minimal hardware, these computations are far less complex than what you might see in cloud edge networks.
Sensor edge computing offers some of the lowest possible latencies but the least computing power. As a result, these environments are best for simple, device-specific tasks, such as using motion detection to trigger connections to other, higher-capacity systems.
Edge Types by Location
You can also divide different types of edge computing by the deployment’s physical location. These could fall into any of the technology-based categories or use a mixture of them, with their distinguishing factor being how computing tasks are distributed across various areas.
A branch edge environment features a dedicated edge network for each branch within an organization. These networks are also called local area network edge systems and aim to provide low-latency support for each office’s specific needs and goals. Consequently, they’re ideal for organizations with multiple branches with location-specific operations.
While 99.9% of all firms are small businesses, having multiple branches is still common for companies today. Adopting a branch edge approach may be ideal for these organizations, though those with less specialized departments may need something else. A business with one location and centralized edge environment could technically classify as a branch edge, just with a single branch.
An enterprise edge distributes computing tasks across multiple branches and locations. These systems typically use a combination of device and compute edge setups to make the most of all available resources within the company.
Each location hosts edge computing infrastructure in an enterprise setup, but they don’t necessarily perform all functions on-premise. The idea of these environments is to enable more flexibility, adapting how you distribute computing resources according to changing demand. As a result, they’re better for large organizations with many different needs but less location-specific specialization.
The most flexible and dynamic of these edge environments is the mobile edge. As you may assume from the name, these setups distribute computing across mobile devices, including non-fixed IoT gadgets and smartphones.
Roughly 85% of American adults own a smartphone, and more than half have a tablet. This vast array of mobile devices provides considerable distribution and computing power for the mobile edge. However, these environments are still less powerful than those using data centers.
Mobile edges are ideal for companies that do most of their work on mobile devices. Operations involving more travel than staying in the same office building can benefit the most from these environments.
When to Use Various Types of Edge Computing
Determining which type of edge computing to use starts with knowing what each offers. Cloud and compute edge deployments lean toward providing more capacity, while device and sensor edge deployment offer lower latency and costs. Consequently, the former two are better for resource-intensive processes or those with many disparate needs, while the latter are better for highly specialized tasks.
Similarly, you should also consider your location needs. More centralized businesses should use branch or enterprise edges, opting for the latter if they have higher computing demands. By contrast, companies with more flexible workflows and locations may prefer a mobile edge environment.
Security is another concern, with 85% of edge adopters citing it as their leading challenge with these environments. Cloud and compute edges are typically easier to secure due to their less distributed nature, so they’re ideal for more sensitive applications. Branch edges may be better if different locations have varying security needs, while enterprise edges are sufficient when these demands are consistent across branches.
Choose the Right Edge Approach for Your Applications
Edge computing comes in many forms, possibly more than many people realize. Understanding these different types and their unique strengths and weaknesses is the first step to capitalizing on this technology. When you know what to look for and how it can benefit you, you can find the best option for your operations.