Edge computing: 6 questions to ask before you start

Before you start that edge computing project or strategy, experts advise you ask some key questions about fit, security, maintenance, and lock-in
499 readers like this.
CIO Cloud 2

What questions should you ask before venturing into edge computing? Asking all the same questions you do for cloud computing would be a good start – but only a start.

Where cloud computing is largely about the benefits of massive scale, edge computing is also about proximity.

Where cloud computing is largely about the benefits of massive scale, edge computing is also about proximity – placing computing resources as close as possible to the user or device that needs immediate access to them, with minimal network latency. Think of critical applications where the autonomous car will crash, or the malfunctioning robot will wreak havoc on the manufacturing shop floor if it has to wait for a response from a cloud system in a remote data center.

Edge can also be helpful for delivering a better user experience for virtual reality, augmented reality, and other computing experiences that benefit from the processing power of resources in the network – when not far away. Telecommunications service providers, in another prominent use case example, are utilizing edge as they modernize their networks and shift workloads out of data centers.

Early content delivery networks (CDNs), which were used to distribute images and videos to multiple locations outside the location of a publisher’s web servers in pursuit of more responsive media experiences, were a precursor to edge computing. The difference is that edge computing is about serving up processing power, not just static resources.

6 key edge computing questions to ask

Enterprises should be thinking now about how they might put edge computing to work in their own operations or to enhance the digital capabilities or products and services they will provide to customers. There could be first-mover advantages for early adopters.

On the other hand, as with any technology approach, there could be flops and embarrassments ahead for those who rush in unprepared. In a recent paper titled How to Overcome Four Major Challenges in Edge Computing, Gartner suggests, “Through 2022, 50 percent of edge computing solutions that worked as proofs of concept (POCs) will fail to scale for production use.”

[ Why does edge computing matter to IT leaders – and what's next? Learn more about Red Hat's point of view. ]

Here are some of the edge computing questions experts recommend you ask:

1. Exactly which edge are we talking about?

As often happens with IT buzzwords, “edge” is being used to mean several different things. By definition, edge is about bringing resources closer to the computing endpoint. But how close? Does the edge server need to be in the same building, in the same city, or just at the closest Internet peering point (where CDNs typically operate)?

“Edge computing can apply to anything that involves placing service provisioning, data, and intelligence closer to users and devices,” says Red Hat technology evangelist Gordon Haff. The term “edge computing” often covers a bit too much territory right now.

"For gamers, the edge is the end-device or the console; for manufacturers, it is more likely to be on-site within the production facility."

“The question of ‘where is the edge?’ is difficult to answer and is often lost in translation depending on who you are talking to,” Dalia Adib, principal consultant and leader of the edge computing practice at STL Partners, explains in a blog post on determining the latency reduction requirements for different applications. “For gamers, the edge is the end-device or the console; for manufacturers, it is more likely to be on-site within the production facility; for a CDN provider, this may be at internet exchange points, yet for the telco community, it is within the operator’s network.”

Adib says the telecom providers she consults with on strategy are excited about turning potentially every location where they now have networking equipment into a miniature data center, and the big ones have some advantages in terms of geographic coverage. They will be competing with, as well as cooperating with, the enterprise software and systems integration companies that bring their own expertise to this market.

[ Get a shareable primer: How to explain edge computing in plain English.] 

2. Will this edge approach meet my organization's privacy and security needs?

Privacy and security were among the challenges Gartner cited. The use of edge services could result in company data being stored in many more locations, potentially under looser management than with the cloud.

"It's a greater concern because now data is all over the place."

“It’s a greater concern because now data is all over the place,” says Adam Drobot, chairman of the board at OpenTechWorks, Inc. and a member of the FCC Technology Advisory Committee, which studies edge computing and its relationship to 5G and IoT, among other things.

Hyperscale cloud data centers are expected to be ISO-certified and conform to a long list of best practices for limiting both physical and network access to customer computing resources. In some edge computing scenarios, you might be deploying data to an unattended wiring closet in an office building or a cell tower equipment shack.

“The further we go into the edge layers, the less control one has on the physical security surrounding the hardware and the communications to that hardware,” says Nick Barcet, senior director technology strategy for Red Hat. “The edge is clearly bringing in new challenges which are requiring us to propose some radically new layers of security.” He points to two open source projects Red Hat is working on to tackle some of those new security challenges: Keylime, which helps ensure software has not been tampered with; and Enarx, a project using WebAssembly to provide a trusted run-time across multiple hardware platforms.

To be clear, there are security and privacy arguments in favor of edge computing. As an FCC white paper on edge, 5G, and IoT notes, “With software-defined networking, it’s possible to develop a multi-layered approach to security that takes the communication layer, hardware layer, and cloud security into consideration simultaneously.”

3. Does the footprint of this edge service match my requirements?

Different applications of edge computing could have wildly different requirements for geographic coverage and proximity. Think about your project’s needs. A manufacturer might want edge computing nodes within each factory, or close by, but for a limited number of locations.

The maker of an augmented reality app that consumers can carry into a store to get real-time product ratings and price comparisons might want to have edge nodes on every street corner in every city, or as close to that as possible.

Let’s explore three additional questions:

David F. Carr is a writer, speaker, student of digital business, and the author of "Social Collaboration for Dummies." He previously served as an editor for InformationWeek, Baseline Magazine, and Internet World and has written for Forbes, CIO Magazine, and Defense Systems.