Once upon a time, the amount of data storage was a big problem. Not anymore. Instead of just aiming for cloud-based business practices every time, companies are thinking long and hard about data speed and the ability to get critical info when it’s needed most.
Think file drawers, from our parents’ and grandparents’ offices years ago. You had a Rolodex on the desk, folders in drawers behind you, and for extreme situations—the file room down the hall.
Now, skip to 2018, and we’re having the same proximity issues. We don’t want to walk to the next room or building to get our data when it’s an item that should really stay here in the cubicle.
Edge computing, and fog computing, are conceptually the answer.
LOCATION. LOCATION. LOCATION.
So where is your data housed?—This is the question. It’s actually a place issue that’s also a time issue. Where we store data depends not just on space, but on how fast we’re going to need to access it.
When data is far away—way up in the cloud—the challenge is not so much storage as it is getting to the data and working with it. But wireless data transmission eats up bandwidth. And if a company doesn’t have much of that to start with, cloud-based work can take forever. And we don’t need time-wasters in our era of insta-everything.
Download and upload speeds are everything, and most enterprises don’t have seconds to waste on data processing. So if “up in the cloud” means “twiddle your thumbs,” then we need a data center solution that’s closer to home. Literally.
Edge computing keeps the data in-house—actually, in-device. If an employee’s desktop or iPad or laptop has to send and receive info, not just to and from the office network, but to and from the cloud, performance will slow—likely not just for that device, but for company productivity.
The edge of the individual computer—be it a desktop, laptop, iPad, or smartphone—is a good place to stop that data travel. Park it in the device itself when you can. Keep the data in the mobile or desktop item, and don’t store it in the cloud so you don’t have to wait to access it. This is the gist of “edge.”
Edge computing is brilliant, really. Usage shouldn’t depend on device-to-cloud speed, especially when time-sensitive performance is of the essence. But if we’re not storing data in the cloud, what happens when people and devices need to share vital data? Is keeping everything on the device, or even on the company network, the best cloud-alternative solution?
Here’s where fog computing comes in . . . the mid-step between the device and the cloud, between right here and remote.
Fog computing takes the ideas of edge data storage and access proximity and operates like cloud architecture. The term is perfect: weather-wise, fog is clouds that rest at street level. So, picture your data, not in the stratosphere, but in the office, on your network, and on the individual devices that are creating and using it. For everyone in the business to have quick access, the devices still need to communicate, but not via a faraway cloud.
Fog computing is your answer—close connectivity. Keep the data on devices and in the office network, on the building server, at the company, across the domain—in a cloud-like fog. Fog experts Nebbiolo Technologies say in a defining whitepaper, “Fog computing interconnects different IoT verticals for resource sharing, data sharing and service sharing for productivity, efficiency and other business factors improvement.” It’s flexible, it’s fast, and it’s functional for B2B or B2C companies that aren’t ready to put everything in a cloud, or who’ve tried that and experienced delayed performance.
Part B of the fog computing idea is that it’s still cloud-ready. David Linthicum of Cisco—the organization that brought fog computing to life—emphasizes the crucial need for encryption and protection in fog processing. “Security is systemic to both edge and fog computing architectures and centralized processing. Security needs to span both and use mechanisms such as identity and access management.”
And then with security measures in place, the data has potential to shift if needed. The fog provides “a layer of computing at the edge of the network that could allow pre-processed data to be quickly and securely transported to the cloud,” says David Greenfield of Automation World.
WHAT’S THE DIFFERENCE? FOG VS. EDGE COMPUTING
Edge and fog computing are clearly related, but not truly interchangeable terms.
The bottom line: Edge computing signifies device—locating data processes near their source for ease and speed of access. Fog computing is network connectivity at the edge—allowing secure wireless, in-house, on-server communication between devices, ready for the cloud but not transmitted.
You can only get so far with edge computing. The data is pushed to the edge of the network. It’s on the device—stored there, processed there, available there. This is intentional because the basic idea is to keep the data processing as close to its source as possible.
But fog computing allows it to develop a little farther out by sharing it. David King of Foghorn clarifies: “Edge computing is not scalable and you can't see across multiple machines or processes with it. Fog computing,” he argues, “extends that edge so you’re not trapped at the individual machine level and connects the devices that need to share that process.”
The sensor-based data processing concept of the Internet of Things (IoT)—data connections between our real-life devices, essentially considering our human activity as a large computer-system-driven experience—is where the edge and fog ideas start to matter more. Sometimes the process can stay on device, where the IoT data originates. The ability of individual devices to use non-cloud-connected sensors to process data right then and there is huge, and for some companies, it looks like enough. But add a mini-server nearby, and you’ll expand your industrial IoT possibilities immediately. This is fog at work. Linthicum states, “Placing a micro-platform at the IoT device, as well as providing tools and approaches to leverage this platform, will likely expand the capabilities of IoT systems.”
So you’re probably questioning the significance of the cloud already. What’s the purpose of sending anything out of the building and into wireless cloudland if we can keep it close by, at the edge?
Back to the 1960s office image: Only so much fits on the desk and in the drawers.
When there’s a big job, taking place over time, requiring access to large amounts of data by many stakeholders, choose the cloud. Not even fog computing can handle some applications at that level of connectivity. “These strategic processes are better placed at centralized servers that can store and process petabytes of data, such as a public cloud,” Linthicum recommends. But when smaller software tasks can stay close—edge computing or fog computing should be able to handle it.
Marin Ivezic of PwC suggests telecom industries, for example, may find fog computing a perfect solution in their struggle to compete: “Fog computing offers a middle ground by mimicking cloud capabilities. . . . Telecom operators have networks that more closely echo this need for ‘small data centers’ that exist at the level of a city block.”
IT scholars Khalid Alqahtani, Saeed Alqahtani, and Yousef Aleidi describe benefits of fog computing in the world of E-commerce as well (International Research Journal of Electronics and Computer Engineering). “The wide range of fog-based E-commerce as a service and integration platforms offers IT leaders great flexibility when choosing deployment models and suppliers. . . . E-commerce companies can start with the basic requirements and upgrade their computing resources as their customer base grows with time.”
CHOOSING AND RE-CHOOSING
So initially, the edge-fog decision is a space decision. Is this data process small enough to stay at the edge, allowing fog computing to work well enough within the business, or does it need more room than what’s available on the device or the network?
Regular capability analysis is not unimportant here, too. For example, your business may decide that a certain application should and could handle most of its data storage and processing at the edge, but later you realize the amount of data is maxing out the network, and you need to switch platforms or take that process back out to cloud. Revisit and choose again. This is not a lock-in. In general, instead of assuming cloud is best, businesses would do best to regularly explore the value of edge and fog alternatives.