Organizations use the cloud for scalable computing and storage solutions to meet their changing business needs. However, another deployment model is quickly growing in popularity: edge computing. In fact, Gartner estimates that 75% of enterprise-generated data will originate and be processed at the edge by 2025.
As companies look for ways to power new use cases, especially artificial intelligence, automation, and machine learning applications, edge computing is a critical resource to lower latency and improve overall performance. It enables organizations to innovate around:
- New business models
- Software service models
- Revenue streams
- Technological use cases
Not all edge computing is the same, though. There are various types to support different business structures, company goals and use cases. Here is what you need to know about different edge computing types to help you find one that works for your organization.
Briefly, what is edge computing?
Edge computing is a computing model that occurs at or near the source of data. To be clear, the “edge” in “edge computing” doesn’t refer to any sort of physical edge.
Traditionally, most data is processed using centralized computing, through major cloud vendors like AWS and Microsoft Azure. If you’re not at the center, then, you’re at the edge of the network — hence “edge computing.”
When computing happens closer to data sources, services become faster and more reliable. Organizations also benefit from the flexibility, given that edge computing allows organizations to use and distribute their systems and resources across multiple locations.
Today, edge computing is critical in many (but certainly not all) cloud environments and here’s why: Cloud infrastructure is often pushed to the limits thanks to the abundance of cloud services and applications it supports. That means it might struggle to process data from connected devices, making it a lot more difficult to generate insights and action in near-real time in response. Cloud computing struggles to keep pace with this explosion of services and applications due to latency, often caused by network distance from the data source.
So, when you have applications (whether internal-facing or customer-facing) that require near-instant analysis and response, edge computing is for you!
Though most data computing today still occurs at centralized data centers, several industries in particular can benefit from edge computing:
How to classify edge computing types
The reason that most companies turn to edge computing is that it supports computing nearer to end-user devices. This reduces how long it takes to process and react to data. Edge has to be highly separated into multiple layers of access points. Some of the ways it is disaggregated include:
- Physical location
- Round trip latency
- Security requirements
- Scale needs
- Data locality
- Connectivity management
- On-premises requirements
- End-user proximity
- Real-time communication needs
- Service capabilities
These factors influence the design, architecture and deployment decisions for practical application. It’s critical to consider what your requirements are first before upgrading your current infrastructure.
Edge computing types by technology
Because there are so many ways to classify edge networks and technologies, there’s no one set answer to how many edge computing types exist or what their categories are. However, one of the easiest ways to separate different types is by their systems and platforms.
Some of the most common edge categories by technology include device, cloud, compute and sensor.
Device edge is the most traditional form of edge computing. Device edge networks use local devices to distribute their computing tasks. These networks have low latency, but they sacrifice capacity because they use devices with minimal power, such as smart gadgets, phones, and routers. As companies over time noticed latency in long-distance transmissions between their colocation sites, they have embraced device edge to bring their computing processes closer to the source of their data.
Device edge networks are ideal for simple and highly specialized applications, so it’s vital to understand your organizational goals and your devices’ capabilities before implementing them.
The cloud edge most closely resembles cloud computing of all types because it relies on large data centers. However, cloud edge computing keeps the centers close to end-users and uses purpose-built applications. They provide significant latency improvements but still maintain the capacity of conventional cloud computing. In fact, 58% of cloud edge users have latencies under ten milliseconds compared to 29% of traditional cloud data centers — a significant reduction that is critical in certain industries.
Because they are more expensive and not available everywhere, cloud edge is best suited for enterprises and large operations with high data demands and nearby existing infrastructures.
Compute edges provide a compromise between the device and the cloud edge. It uses small, purpose-built machines to distribute tasks while also leveraging micro data centers (MDCs) for greater capabilities.
MDCs usually range between 50 to 400 kWh in size with up to a few racks. There are more scalable and adaptable than traditional centers because they are modular — giving compute edge systems more flexibility, a smaller footprint and more power than conventional device edge.
Compute edge is best for companies that don’t have access to nearby data centers and have various edge computing needs. While MDCs cost more than device edge networks, they also serve a wider variety of use cases.
The rise of the Internet of Things (IoT) and automation has required organizations to leverage networks as close to data sources as possible to minimize latency. Because they are near the data, they can perform basic computations using sensors.
IoT endpoints are data collection points. Sensor computing enables organizations to move some of their analytics processes to the endpoints to minimize latency and maximize performance. However, the computations are less complex because the endpoints have minimal hardware.
Sensor edge provides the lowest latencies possible but also has the least computing power. It is ideal for device-specific and simple tasks.
(Learn how to monitor IoT.)
Edge types by location
Your deployment’s physical location is another common way to organize edge types. While these could be a part of any of the technology-based categories — or even a mixture of them — how they compute tasks across various areas are their distinctive factors.
Many organizations use edge computing to complete tasks across multiple locations and branches. Enterprise edge helps you distribute your computing tasks with a combination of compute edge setups and devices to use all available resources effectively.
Enterprise edge provides flexibility to adapt to distributed computing resources and changing demands. It’s best suited for large companies with different needs but limited location-specific specialization.
Branch edge networks are dedicated edge networks specialized for each organizational branch. Sometimes called local area network edge systems, they provide specific, low-latency solutions suited for each office’s goals, challenges and needs. They’re best suited for companies that have location-specific operations and multiple branches.
Mobile edge computing provides the most flexible and dynamic environment. Its network is set up across mobile devices, including smartphones and non-fixed IoT gadgets.
Considering that most Americans own smartphones and many have tablets, mobile edge harnesses this vast array of mobile devices to offer significant distribution and computing power. However, it is also limited because mobile devices lack the capability of more traditional data centers and devices.
Mobile edge provides an effective network for organizations that rely on mobile devices for most of their work. Also, those whose business relies more on travel — as opposed to being anchored to an office building — might best benefit from the mobile edge network.
Choosing your edge computing type
With the number of options available for edge computing today, it can be challenging to know which one is right for your organization's needs.
Know your goals
It is first critical to understand what each type offers and how they fit your company's goals. For example:
- Compute and cloud edge deployments provide users with more capacity.
- Sensor and device edge deployments have higher speeds and lower costs.
That means the first two are more suited for resource-intensive processes and organizations with multiple different needs, while the second two are best suited for highly specialized tasks.
Understand your location(s)
Location is another important consideration when finding the right edge type. If your company is more centralized, branch or enterprise edges will be the better option, depending on your computing demands. On the other hand, if you have more flexible locations and workflows, consider a mobile edge environment.
Don’t forget security
Security is another concern regarding edge computing: 52% of IT teams said they had issues ensuring complete security across all their edge devices. However, compute and cloud edges are easier to secure because they are less distributed in nature. That makes them an ideal type for sensitive data and applications.
Branch edges are best suited for various locations with different security needs, and enterprise edges are adequate for consistent demand across branches.
(See how security frameworks can support edge use cases.)
Find the right technology for your needs
Each edge computing model comes with strengths and setbacks. Determining which is best depends on the size and type of business, your data needs and your company goals. The industry constantly evolves, leading the way for innovation edge models and use cases. When you know the benefits of each and what to look for, you can find the best fit for your needs and operations.
What is Splunk?
This posting does not necessarily represent Splunk's position, strategies or opinion.