Edge computing is computing that is handled away from central data centers -- at or near the source of data. This computing could be on a manufacturing shop floor, in a finance department, at a retail store, or in a field service office far away from headquarters.
In almost every case, the method for handling data at the edge is the same. You process the data locally at the site; you store the data locally; and at an appropriate time -- whether it's real time, near real-time, or in a batch data delivery mode -- you upload this data to a more centralized cloud destination or to your own data center.
Legacy edge computing was known as distributed computing. It consisted of servers and printers and storage that were remotely positioned in company departments, away from central IT. This is the kind of edge computing that companies deploy with confidence because they already know how it works, and they have many examples of how it can be used successfully.
A bigger challenge, however, is getting into newer edge technology -- like IoT.
"IoT projects often stagnate because decision-making is hard; because the dials, knobs, and switches are unseen or poorly understood, and decision-makers don't want to break anything," said a founder of an IoT integration company.
For CIOs and other IT decision makers, recommending these newer IoT edge projects, which still have so many unknowns, can be risky and even career-threatening.
What's the best way to make a case for an edge computing project? Here are 10 questions to consider.
SEE: IT leader's guide to edge computing (Tech Pro Research)
1. Does everyone see a business case?
If your industry is agribusiness and you can see how sensors placed on pallets of produce picked in the field can monitor the freshness of each pallet, so you can get produce that's prone to spoilage to market faster, there is a business case because it's measurable and you have the potential to reduce waste and improve profit margins. The rule of thumb is this: promote edge projects only where results are measurable and capture business value that everyone will understand and support.
2. Are you feeling confident in promoting an edge project?
One of the most important areas of edge computing is seeking help from vendors via proofs of concept and no-obligation pilot projects. IT leaders want to be convinced that a new technology works before they walk into the CEO's or CFO's office. Never approach a CEO or CFO with an edge computing funding request unless you're convinced that the project will deliver what you think it will. The best way to do this is to trial the technology first.
3. Will the technology pay off soon?
Gone are the days of waiting one or two years for a return on investment (ROI) from an edge computing project. Organizations now look for payoffs within three to six months. If you're not convinced that your project can show ROI within that time period, reconsider your proposal.
4. Is the technology easy to use?
If technology is being implemented at the edge, it generally means that employees at the edge (not IT) are going to be the ones using and managing the technology. If users find that the technology is cumbersome and difficult, they'll give up. Adoption of new edge technology depends upon user acceptance. Don't even consider an edge project if you don't have enthusiastic users backing it and if a proof of concept can't demonstrate that the technology is easy to use.
SEE: IT pro's guide to the evolution and impact of 5G technology (free TechRepublic PDF)
5. How supportive is the vendor?
Edge technology projects require strong vendor support that goes beyond helping you with your ROI formula or conducting proofs of concept. Take a hard look at your vendor's SLAs before you sign on any dotted lines. Does the vendor offer strong project implementation and support? How strong is the vendor's post implementation support? Avoid proposing an edge computing project until you know the answers to these questions and are comfortable with them.
6. Does the edge technology integrate with your existing technology base?
Most edge technology vendors offer open APIs (application programming interfaces) and will tell you that their solutions interface with any system, device, or network that uses a common industry interface or communications protocol. Unfortunately, glitches can still come up during installation and will require extra time to resolve. In other cases, there is IoT equipment (medical equipment, for example) that might use non-standard communications protocols. This equipment must be custom-integrated with your existing systems, and that effort will take time as well. If extra time and effort are needed, you'll want to include that in your project and in your projected ROI.
7. Can IT support the edge technology?
Before presenting a new edge technology for budget consideration, there should be a plan for how the technology will be supported and enhanced once it is installed. Is the technology going to be deployed as a totally outsourced solution that the vendor alone takes care of? Or will you expect IT to maintain and enhance it? If the choice is the latter, does IT have the knowledge base to support the technology?
SEE: Internet of Things policy (Tech Pro Research)
8. Will it be necessary to revise IT architecture?
In most cases, edge computing requires changes in IT architecture. Data is processed and stored locally at the edge. It is then transported to intermediary (or perhaps permanent ) clouds. From the cloud, it is often transported at an appropriate time to the data center so all data can be consolidated in a central repository. Of course, there are many variations to this theme that companies enact depending on their data needs. This impacts IT architectures, which will need to be redefined as more edge technology projects are deployed.
9. Do you have the communications pipelines needed to support edge computing?
If you're going to move payloads of data between devices, servers, and storage internally -- or ship them out to clouds or to your central data center -- you're going to require the bandwidth to accommodate these payloads. This will require deployment of data pipelines that can handle the traffic. Schedules will need to be made regarding the data that's transported over those pipelines so traffic contention is kept to a minimum. If in the process you discover that more bandwidth will be needed for your project, those costs should also be included as an upfront budget expense.
10. Can you bulletproof the technology for security?
If you deploy edge technology, and end users are using and maintaining it, security issues that are second nature to IT might not be followed. You still have to ensure security. One way is by installing a zero trust network that authenticates user identity. A second way is to make sure that physical IT assets are secured in cages, closets, and other lockdown areas.
- 3 tips to make implementing edge security easier (TechRepublic)
- Where's the "edge" in edge computing? Why it matters, and how we use it (ZDNet)
- Internet of Things (IoT): Cheat sheet (TechRepublic)
- Enterprise IoT research: Uses, strategy, and security(Tech Pro Research)