Top three problems to solve around edge computing

February 28, 2019

A blog post by David Linthicum, managing director, chief cloud strategy officer, Deloitte Consulting LLP

Edge computing is often confused with cloud computing, but they are very different and do very different jobs. Indeed, I’m not sure edge computing would have been created were it not for issues that business found in transmitting everything back to a centralized cloud.

As with any new technology, there are specific problems related to edge computing that are still left to be solved and the solutions approved. Here is my take:

Data tiering. While we know that edge-based devices need to store data, and we do so on the cloud tier as well, how do you divide things up? It’s really a matter of what needs to return to the data provider in near real time, what can move to the back-end systems, and finally what data should be replicated.

No easy answer for this one and indeed that’s the problem. As time progresses it’s a matter of planning and deciding which data should exist at the edge tier and which data should exist at the cloud tier. For now though, it’s still a case-by-case basis.

Security. While there has been a lot of thought around what edge means and how it’s implemented, there is still no standard approach to formal security planning for edge computing.

Core to these issues is one of planning standards that are not just extension of existing security systems on the cloud, or on premises. This means that even with developing industry standards, there is currently no widely-accepted best practice.

The trouble with security and edge-based systems is that they are typically lower-powered computers that can’t handle much overhead other than what they currently do as an edge device. Therefore if the security systems are implemented there, they need to keep a low profile. And that is something that security systems have a hard time doing.

Management. While businesses are getting better at managing and monitoring, much like the two issues listed above there is no one-size-fits-all approach or tooling. The specific features of edge computing make “edge-ops” solutions unique to each situation. This will be the case for the near future.