Overcoming the core hurdle to edge computing adoption

Edge computing is picking up steam. According to this recent report by Turbonomic (requires registration), nearly 50 percent of organizations use or plan to use edge computing in the next 18 months.

For those of you watching this market, many existing development projects listed as “edge computing” barely qualify for the title. Still, considering the state of edge computing just a few years ago, this is a huge leap in growth.

The factors that drive enterprise movement to the edge include:

• Edge-based solutions in the public cloud. In essence, these are pared down, private cloud versions of public clouds, such as AWS Outpost and Microsoft Stack. They often serve as a jumping off point from legacy systems to public clouds—like a public cloud with training wheels.
• IoT-based projects. Data storage and compute that’s closer to the edge of the network and to the source of the data provides better performance because less data is sent back to the centralized public cloud server.
• Edge computing architectures. This architecture involves more substantial and traditional servers, such as traditional storage and compute servers housed in specific offices or branches. Consider a restaurant chain that needs to place storage and compute at all locations but also wants to use a centrally managed paradigm.

What stops the forward progress? No surprise here: It’s managing complexity without added cost and risk. According to the Turbonomic report: “Complexity, at 39 percent, is overwhelmingly considered the leading barrier to edge computing becoming conventional.” Complexity is almost double the second-place and third-place barriers: security (23 percent) and technology limitations in network/bandwidth throughput (22 percent). 

If this survey had been done a few years ago, I suspect that security and technology limitations would have been in the top two spots. What happened? In short, actual edge computing projects took the place of conceptual ones, with as many as 20 to 30 percent failing outright due to the inability to manage complexity.

It isn’t easy to manage widely distributed systems. There are challenges around configuration management, patching and software updates, CI/CD (continuous integration/continuous delivery), acceptance testing, distributed data storage, and security operations within edge-based implementations. This list is only a fraction of the complexity issues that must be managed at the edge.

For now, these problems are difficult but not impossible to manage. There are any number of AIops, governance, and configuration management tools for cloud computing; however, very few tools are focused at the edge.

Why? It’s difficult to nail down a repeatable approach and a technology stack for edge solutions. Edge-based systems can include almost any hardware and software, with a wide range of capabilities and limitations. In contrast, developers can depend on the consistencies of public cloud platforms.

Edge computing will need sound, repeatable approaches to mediate complexity, as well as tools that provide consistent ways to approach the problem. We are just not there yet.