Originally published May 2019. Based on analysis from the CEO of Packet (now part of Equinix Metal), this article outlines five foundational shifts in wireless technology infrastructure that were emerging at the time and have continued to develop in subsequent years.
The wireless world entered 2019 at an inflection point. Landmark developments in 5G, cloud-native software, edge computing, and custom hardware were converging to reshape how enterprises think about connectivity and compute infrastructure. From sports stadiums and retail buildings to airports and urban street corners, no physical environment was left untouched by these rapid changes.
Below are five trends that defined the wireless technology evolution of that era — and whose implications continue to play out today.
1. Software Will Eat the Wireless World
The assumption that “technology equals software” is a simplification, but it captures a real market dynamic. Even cloud computing — itself the engine of the most recent software wave — was being disrupted from within. Cloud-native ecosystems such as Docker and Kubernetes began eroding the proprietary value of vertically integrated cloud providers by making workloads portable across any infrastructure substrate, spanning multiple architectures, providers, and form factors.
For wireless infrastructure, this meant that the intelligence governing how networks are managed, sliced, and optimized was increasingly delivered through software layers rather than proprietary hardware stacks. Network functions virtualization (NFV) and software-defined networking (SDN) accelerated this shift, enabling carriers and enterprises to program their networks rather than re-cable them.
2. Hardware Is Not a Commodity
Despite the software narrative, hardware was staging a quiet comeback. The limits of general-purpose CPUs for workloads such as machine learning inference, packet processing, and real-time video analytics were driving significant investment in specialized silicon: GPUs, FPGAs, SmartNICs, and purpose-built accelerator cards.
The insight here is that software and hardware are not in opposition — the most capable infrastructure stacks require both to be co-designed. Customizable, bare-metal hardware in combination with cloud-native software represents a powerful deployment model, particularly at the network edge where latency budgets are tight and compute resources are constrained.
This trend has continued to accelerate. Custom silicon efforts from major cloud providers and chip vendors underline that differentiated hardware remains a genuine competitive moat.
3. Real Estate Is Being Digitized
Enterprises deploying wireless and compute infrastructure at scale quickly discovered that traditional data centers are not the right model for every use case. Applications such as 5G radio access, mobile gaming, augmented reality, and real-time analytics require compute resources that are geographically distributed and physically close to end users.
This drove a new category of infrastructure deployment: digitizing physical real estate. Large venue operators — stadiums, airports, co-working spaces, shopping centers — became candidate locations for micro data centers and edge compute nodes. Rather than routing traffic to a centralized cloud region hundreds of miles away, operators could process data on-site, reducing latency and improving the user experience for high-density environments.
For DevOps teams, this introduced new operational complexity: infrastructure is no longer confined to a handful of known data center locations, but distributed across dozens or hundreds of edge sites that must be managed consistently.
4. 5G Is Really About Private Networks
Much of the public conversation around 5G focused on consumer smartphones and headline download speeds. But for enterprises, the more strategically significant development was the emergence of private 5G and private LTE networks.
Industrial companies — manufacturers, logistics operators, port authorities, healthcare systems — have connectivity requirements that public carrier networks are poorly suited to meet. These requirements include deterministic low latency, network isolation for security, and the ability to operate independently of public infrastructure. Private LTE and 5G networks, enabled in part by spectrum access frameworks such as Citizens Broadband Radio Service (CBRS) in the United States, made it feasible for enterprises to own and operate their own wireless networks on licensed or shared spectrum.
Private 5G deployments allow organizations to apply the same security and access-control policies to wireless connectivity that they apply to their wired networks, closing a significant gap in enterprise network architecture.
5. Edge Computing: The Next Frontier in Cloud Infrastructure
Edge computing refers to processing data at or near the source — on a device, at a base station, inside a venue, or at a regional hub — rather than routing everything to a centralized cloud data center. For latency-sensitive applications such as autonomous vehicles, industrial automation, augmented reality, and real-time video analytics, the physics of light-speed transmission mean that centralized cloud architectures simply cannot meet the required response times.
Edge computing is not a single technology but a deployment philosophy. It encompasses a spectrum of locations and compute tiers:
- Device edge: Processing on the end device itself (smartphones, sensors, cameras).
- On-premises edge: Compute located within the enterprise facility or venue.
- Network edge: Compute co-located with carrier infrastructure such as base stations or central offices.
- Regional edge: Smaller cloud regions positioned closer to population centers than hyperscaler availability zones.
For DevOps engineers, edge computing extends familiar automation and orchestration challenges into environments that are physically distributed, resource-constrained, and sometimes intermittently connected. Tooling choices — container runtimes, orchestration platforms, observability stacks — must account for these constraints.
What These Trends Mean for DevOps and Infrastructure Teams
Taken together, these five shifts have material implications for how engineering teams plan, build, and operate infrastructure:
- Infrastructure-as-code must scale to edge sites. Managing configuration drift across hundreds of distributed edge nodes requires mature automation practices and centralized control planes.
- Hardware selection matters again. Workload-specific silicon choices — accelerators, SmartNICs, ARM-based processors — affect both performance and operational complexity.
- Network and compute teams are converging. 5G private networks and edge deployments blur the traditional boundary between network engineering and cloud infrastructure. Teams that can work across both domains are increasingly valuable.
- Security perimeters are dissolving. When compute nodes exist in co-working spaces, stadiums, and street-level cabinets, physical and logical security models must be redesigned from first principles.
- Observability becomes harder and more important. Distributed edge deployments generate more failure modes and make centralized logging and metrics collection more architecturally challenging.
The trends identified in 2019 were early signals of a structural shift in how compute and connectivity infrastructure is designed. Understanding their trajectory helps engineering teams anticipate where investment and skill development will be most valuable in the years ahead.