It’s easy to view edge computing as a literal outlier. As compute workloads have migrated towards ever larger data centres, it can seem like the edge is home to a small number of applications that are either too niche or too legacy to move.
But that impression is no longer true, if it ever was. After all, what are humanoid robots or self-driving vehicles but moving examples of computing at the edge?
The edge encompasses a whole range of critical workloads, from retail networks and remote mining sites, to managing factories and supporting medical facilities.
Certainly, investment at the edge is on the rise. IDC earlier this year said global spending on edge solutions would hit $261 billion in 2025 and would hit $380 billion by 2028.
In an AI-driven, data-saturated world, this makes sense. Edge compute brings information storage and computing abilities close to the sources of the data and/or the users who consume it. This delivers key advantages, both for users and the technology teams supporting them.
Huge advantage
As Nate Melby, Dairyland Power Cooperative VP and CIO, told CIO.com: “AI makes edge computing more relevant to CIOs because it helps us reduce delays in processing data. And in situations where we’re aiming for real-time processing, this can be a huge advantage.”
Take manufacturing for example. An industrial or humanoid robot needs to carry out actions and make decisions in real time. Sending data back to a central data centre for key decisions can mean delays or even a critical failure.
The stakes are even higher when it comes to healthcare settings, where the smallest hold-up can have a life-changing impact.
Likewise, holding data at the edge can offer a security and governance advantage. Anything that must be transmitted can be encrypted, while local storage may make it easier to comply with data regulations.
High speed
All of this reduces the need for bandwidth, a key benefit in itself. Some real-world applications – autonomous operations in agriculture or the energy sector for example – may simply not have access to high-speed data links.
A major driver of the shift to the edge is the increasing importance of inference in AI.
Remote data centres are ideal for the training phase of model building. But putting those models to work in real time – for chat bots dealing with customers, or to deliver instant credit or onboarding decisions – means latency must be shaved to the absolute minimum.
But that doesn’t mean cutting the cord with a cloud provider and forgoing the advantages they bring in terms of security, scalability and breadth of platforms and tooling.
As more focus moves to the edge, major cloud providers are expanding their own offerings accordingly. IDC predicts that infrastructure investments by service providers in multi-access edge computing, content delivery and virtual network functions for enterprise customers will be close to $100 billion by 2028.
This covers services that reach right to company premises. AWS, for example, extends its cloud infrastructure, services and tooling to virtually any on-premises data centre, co-location space or edge facility. This means even “austere” locations can take advantage of AWS services and infrastructure.
It’s clear that AI is playing an increasing role in all aspects of day-to-day life, from delivering content and speeding interactions with finance institutions to underpinning healthcare. It is at the edge where computing really touches our real lives.