Slicing Up the Network with 5G SA: An Interview with Telit Cinterion's Stan Gray

I had the privilege of interviewing Stan Gray, who helped me understand how the quality of service in network slicing is an end-to-end to service.

First, a little about him: Stan Gray is a seasoned professional in the technology sector, currently serving as the Senior Vice President of IoT Broadband and High Cat Vertical Sales at Telit Cinterion. His career began at Motorola in 1988 as a Corporate Sales Manager, followed by a decade at Panasonic Canada as Director of Marketing. Gray then held various leadership roles at Sierra Wireless, including Vice President positions overseeing global sales and OEM solutions. He holds a Bachelor of Arts in Political Science and Government from McMaster University

Stan Gray: The rollout of 5G Standalone (SA) represents a significant technological leap, but its global adoption varies widely.

Here at home, T-Mobile is leading with the first nationwide 5G SA network. Dish and Rogers Communications in Canada have also begun implementing SA in their networks nationwide. Others have been slower to launch SA networks. This is largely because of the success and maturity of its 4G LTE network, where carriers have been incentivized to leverage 5G Non-Standalone (NSA) (which piggybacks on 4G LTE cores) to deliver faster 5G speeds while minimizing immediate capital expenditures.

Abroad, countries like China, South Korea, and Germany are leading the global push. China’s strong government support has rapidly scaled nationwide coverage, South Korea’s competitive telecom landscape has driven aggressive innovation, while Germany has focused on industrial applications and private 5G networks to support manufacturing and automation.

National strategies, early and flexible spectrum allocation, strong industrial demand, and close collaboration between governments, operators, and enterprises are driving their successes. These countries are leveraging 5G SA not just as a consumer network, but as critical infrastructure for their digital economies.

CF: As I understand it, fundamentally, wireless is a last-mile service that connects to the fiber network. Does this mean the fiber network needs to do network slicing as well?

SG: To enable true end-to-end 5G capabilities — especially with features like network slicing — the fiber network must support or coordinate with slicing. Wireless serves as the last-mile connection, but most 5G data is transported over fiber backhaul and core networks. If slicing is only enforced at the wireless edge, the guaranteed latency, bandwidth, and security can be lost as traffic moves deeper into the network.

The fiber network must be capable of managing traffic in a slice-aware way to maintain performance and isolation for different network slices. This can be done through technologies and tools that allow operators to logically separate and manage traffic from different slices all the way from the radio access network through to the core and cloud.

Network slicing isn’t just a radio feature. To ensure consistent service levels and enable application-specific networking that 5G promises, it requires the entire infrastructure, including fiber transport, to be integrated into slice-aware orchestration and management.

CF: If I were still a network planner, I would recognize that I have many different areas that have many different needs, such as rural, suburban, inner cities and high-density business centers. Are the slices of the network different in each jurisdiction? And for high density areas – or, for that matter, hybrid private networks – will slicing the network have dedicated slices to a specific customer, like fiber has been?

SG: Network slices are customizable and designed to serve different use cases and geographies. For instance, different slices can support ultra-low latency use cases, from autonomous vehicles in cities to massive IoT applications in rural and agricultural settings. We sometimes say slices can have "different personalities," meaning slices can be tailored to match the specific needs of various jurisdictions, whether rural, suburban, or urban.

Slices can be dedicated, much like dedicated fiber lines. They ensure consistent performance even in high-demand scenarios (e.g., crowded stadiums or urban centers). Slicing can offer dedicated service per customer, similar to how fiber networks operate for premium or enterprise customers.

CF: The U.S. government is about to clear the way for another spectrum auction. How important is the auction to solving the requirements for 5G SA, particularly in the higher bands?

SG: The upcoming auction is very important for support of 5G SA, ultra-low latency, high capacity for dense urban areas and network slicing. While the auction presents opportunities to enhance network capabilities, it also brings forth challenges that require careful consideration to ensure the secure and effective deployment of 5G technologies.

As a side note, the U.S. government is hoping to use the funds from an auction for the announced infrastructure replacement program that has been proposed to the major MNOs. The fund is to help offset some of the costs associated with replacing Huawei and ZTE infrastructure.

CF: I have been troubled by the edge compute market surpassing the small cell market and thought that small cells would join the edge compute industry. Will either product be able to support network slicing? 

SG: While both small cells and edge compute can support network slicing, they play different roles in the 5G ecosystem. It may seem surprising that edge computing is outpacing small cells in market growth, but this trend reflects demand for low-latency data processing close to users and devices, which edge compute is better positioned to address at scale.

Small cells are fully capable of supporting network slicing in 5G SA deployments. When integrated with a slice-aware RAN and coordinated with centralized orchestration, small cells can allocate resources based on the needs of each slice.

Edge compute, on the other hand, complements this by enabling slice-specific application processing and data handling closer to the user. For example, a slice designed for autonomous vehicles could be paired with edge compute nodes that process sensor data locally, reducing latency and network load. As a result, edge compute and small cells are complementary parts of the same network slicing ecosystem. Together, they enable tailored, end-to-end experiences that are central to 5G’s promise.

CF: These efforts to deploy 5G SA are laying the foundations for a new generation of fully cloud/AI-based platforms, and a world in which the boundaries between small and macro cells become irrelevant, as all kinds of connectivity and cloud infrastructure converge to support a vast range of future services and experiences. This could be a chicken vs. the egg problem: There are few devices for the mid (eMBB) and high (URLLC) spectrums. Is this a case of "if you build it, they will come" or is there pent-up demand already on the device side? If so, what vertical markets are waiting?

SG: Should networks invest heavily in advanced capabilities like URLLC and eMBB without widespread device support (the chicken), or will demand naturally follow as the infrastructure matures (the egg)?

In many ways, this is a "build it and they will come" scenario, where the use cases are still emerging and require a mature, low-latency network environment before device ecosystems can flourish. That said, there is already pent-up demand in several verticals waiting on more consistent mid- and high band 5G coverage, low latency, and better edge integration.

While these industries are actively testing or piloting 5G SA use cases, broad-scale deployment truly hinges on the maturity of both the network and the device ecosystem. As cloud-native network slicing and AI orchestration become more mainstream, the synergy between connectivity and compute will unlock scalable, application-specific services. This will ultimately fulfill the long-promised potential of 5G.

CF: The auto industry is working hard to enable V2X. Will full implementation of V2X require network slicing?

SG: The full implementation of V2X will greatly benefit from – and in many cases, require – network slicing to meet its diverse and stringent requirements. V2X encompasses various communication types, each with different performance needs (ultra-low latency, high reliability, and guaranteed bandwidth). Network slicing allows operators to create dedicated virtual networks tailored specifically for V2X services, ensuring that critical safety messages and autonomous driving data get priority and are isolated from less time-sensitive traffic.

For example, a slice can guarantee the URLLC needed for collision avoidance, while another slice might support infotainment or over-the-air updates without impacting safety-critical functions. Network slicing provides the key flexibility, and performance guarantees essential for the complex V2X ecosystem, making it a paramount enabler for the safe, efficient, and scalable deployment of connected and autonomous vehicles.

CF: Is network slicing going to have incompatibility issues? Will there be an expectation for handoffs between carriers? Does ONAP support network slicing? 

SG: While some interoperability challenges remain, ongoing standardization efforts and platforms like ONAP are paving the way for seamless network slicing across carriers and vendors in the near future.

Since slices are virtual networks tailored for specific services, coordinating slices across multiple operators’ infrastructures requires standardized interfaces and agreements to ensure seamless service continuity. This includes handoffs when users move between coverage areas managed by different carriers.

Currently, industry bodies like 3GPP and ETSI are working on defining standards to support multi-operator slicing and inter-slice handovers, but widespread real-world implementation is still evolving. This means that, initially, network slicing may be mostly confined within single operators’ domains, with multi-carrier handoffs becoming more common as ecosystems mature.

ONAP provides an open-source framework for automating the design, deployment, and management of network services, including slice lifecycle management. It enables operators to orchestrate and coordinate resources across the RAN, transport, and core networks, making it a key enabler for operationalizing network slicing in complex, multi-vendor environments.

CF: I want to thank Stan and his team for organizing our conversation, answering my questions, and his willingness to participate in future dialogues.