As the use of Internet of Things (IoT) devices expands and the need to process IoT data quickly increases, many IT leaders are considering or beginning to employ  edge computing options. As Forrester vice president and principal analyst Brian Hopkins has also pointed out, advanced artificial intelligence capabilities like deep learning and neural nets also benefit from an edge computing architecture.

[ What is edge? For a primer, see How to explain edge computing in plain English.] 

The upside is apparent for such use cases. Pushing processing power to the edge of the network – and closer to data sources – can reduce latency, enable near real-time analytics, enhance customer experiences, improve operations, and enable better or faster decision-making. However, IT leaders need to work through a number of issues when adopting edge computing.

The Enterprisers’ Project talked to edge computing experts about the key questions CIOs and IT leaders should ask as they plot their edge computing initiatives.

Edge computing: Getting started

1. What applications, services, or business strategies will be delivered from your edge computing platform?

If you haven’t already, determine which workloads should really run on the edge versus a central location, advises Yugal Joshi, vice president with Everest Group. IT leaders should also explore whether there are any existing intiaitives (IoT or AI) that would benefit from edge processing.

2. How far out do we need to go?

Latency decreases as you push computing closer to the edge. However, costs increase. “If you have a low bandwidth, latency tolerant app, it would make sense to host the edge in a data center or regional point of presence where the cost is lower,” says Julius Francis, director of product management and marketing at Juniper Networks. “However, if you are doing 4K/8K gaming with ultra-low latency and high bandwidth, you would want to position the edge as close to the consumer as possible.”

3. Are there proven use cases which could inform our approach?

“Although over focus on peer adoption may bog down innovation and the disruption an enterprise can bring, it can help in building more confidence in the business case and defining the charter,” Joshi says.

4. Can we recycle any code assets from our on-premises environments?

“Look for synergies,” says Holger Mueller, vice president and principal analyst at Constellation Research. That will reduce the time it takes to develop the edge environment and also help contain costs.

5. Will edge computing require any changes to the organization’s operating model?

Edge computing is often used to support operational technology. In such cases, technology leaders should decide who will own and manage the edge environment, whether greater alignment is required between operating technology and information technology groups, and how performance will be measured, says Joshi.