By Erik Nordmark, co-founder and CTO of Zededa
While AI grabbed nearly all the tech industry headlines in 2023, another industry — edge computing — was gaining tremendous traction and maturity across many sectors.
Here are just a handful of predictions from the 2023 Gartner IT Infrastructure, Operations & Cloud Strategies Conference that drive the point home:
- By 2025, more than 50% of enterprise-managed data will be created and processed outside the data center or cloud.
- The number of IoT devices will triple from 2020 to 2030 — and the data they produce will grow even faster.
- By 2027, 20% of large enterprises will have deployed an edge management and orchestration solution, compared to less than 1% in 2023.
As more computation and data storage moves to distributed edges of the network, close to where data originates, exciting new capabilities become possible even as new challenges arise.
Here’s a look at three key trends to watch in 2024:
Centralized control: integrating policies from day one
Policy control, including security measures, has traditionally been applied after the fact, such as handling how people use their laptops or phones on the corporate network. For edge computing, however, it makes more sense to establish these controls from the outset. Even though there’s no central physical control over something sitting out on the factory floor or in a solar farm, there should still be a centralized policy control.
This involves implementing centralized policies governing applications, security postures and deployments globally while allowing edges to operate independently.
There’s also work to be done around governance and overseeing the provenance and security of data. Policy control and governance nuances become crucial, especially in edge computing, where data exists in various distributed locations. This dual-layered approach balances the need for control and visibility while grappling with challenges like data ownership and provenance.
Edge AI inferencing and privacy-preserving distributed learning
Enterprises sometimes refer to edge AI efforts as “boring AI” — the practical application of AI in industrial settings to unlock incremental improvements in predictive maintenance, defect detection and analytics.
Even with that in mind, it’s still hard to predict what will happen with AI. It appears that companies are less likely to train their models and will leverage something already trained. Edge devices will increasingly collaborate on privacy-preserving distributed learning to train and update models over time without exposing proprietary data. This situation will allow models to evolve and improve by learning from new data throughout deployments that can run for nearly a decade.
OT and IT each gain a broader understanding of edge security requirements
Successfully securing edge deployments requires bridging the divide between traditional corporate IT teams and operational technology (OT) teams running distributed infrastructure. These groups have not traditionally worked together, but taking an integrated view of requirements and threats is essential for safety and security.
Success requires collaboration between IT and OT experts from the start rather than tacking on security as an afterthought. The convergence of security and safety considerations is crucial, demanding a holistic approach rather than a segmented, adversarial one.
Circling back to the insights from the Gartner conference, it’s clear that most edge computing is being implemented without a strategy. In 2023, 19% of CIOs that Gartner surveyed said they had already deployed edge computing, while more than half said they expected to deploy edge computing by 2026 and 30% weren’t sure.
That leaves plenty of room for business and technical leaders to collaborate on an edge computing strategy and the opportunity for them to include these emerging trends in their plans.
About the author
Erik Nordmark is the co-founder and CTO of Zededa. Nordmark is widely considered to be one of the technical visionaries in the edge computing industry.
DISCLAIMER: Guest posts are submitted content. The views expressed in this post are that of the author, and don’t necessarily reflect the views of Edge Industry Review (EdgeIR.com).
data center | edge AI | edge computing | IoT | Zededa