The shape of edge computing in 2021

In 2019 and 2020, edge computing received a lot of attention because most centralized cloud computing challenges involve latency issues that are difficult to mediate. You can tell the cloud providers to build points of presence closer to your users—and new points of presence continue to show up all the time and all over the world—but latency continues to be a common problem. Why?

A better question is, what’s not happening? New edge computing projects tend to stick processing and data collection closer to the sources and users of the data. However, architects need to do so in a way that is scalable and manageable. Operated individually, the number of edge systems under management will have a clear tipping point where operations and security can not keep up with the complexity of thousands of edge-based systems. The focus must shift from operating each edge-based system individually to setting up management and configuration systems for the whole. 

The irony is that many consider edge computing to be the antithesis of public cloud computing, a tool to remove many of the complexity problems that cloud computing generates. In real-life applications, however, public cloud computing is becoming the jumping-off point for all successful edge computing strategies. Ops teams manage edge computing systems using centralized command-and-control where configurations and security profiles are centrally kept and managed within public cloud providers.

Over time, we’ll see new approaches, such as edge ops and edge secops, that exist independently of the larger public cloud providers. Considering the number of resources you need to be successful with edge ops management, it’s unlikely that those services or technologies will be as compelling as the ones offered by the public cloud providers that will host cloud-based and edge-based systems. Those who still view this market segment as edge versus cloud are missing a major point. The core trends we saw last year are the same ones we will likely see again this year.

What does this mean to ordinary enterprises that want to understand the potential of edge computing? What trends will we see this year? Here are a few concepts for your consideration:  

• Security and governance are problems yet to be solved in the emerging world of edge computing. Hackers will likely infiltrate edge-based systems as points of access because the use of edge computing itself is well ahead of the use of edge computing security. 
• We will move to edge computing for performance reasons because it can remove much of the latency. However, we’re still operating widely distributed systems. Latency is pretty much nonexistent between the edge-based system and the data being gathered, but most edge systems still depend on centrally stored data. The system needs that data for most of the processing at the edge, and that data can’t be accessed without some degree of latency.
• In many instances, edge computing requires us to give up productive features of public clouds, such as serverless and managed containers. These features are likely to show up with cloud-based edge computing processing, but their capabilities will probably lag behind the core features of central public clouds. 

These forecasts should not turn you away from edge computing but should open your eyes to what’s realistic. As enterprises widely adopt edge computing and it moves from the hyped stage to real-life operations, there should be few surprises.