Picture a highway. Today, every vehicle — freight trucks, ambulances, bicycles, sports cars — shares the same lanes. Now imagine you could carve out dedicated lanes for each type of traffic: an express lane for emergency vehicles, a steady one for freight, a wide one for mass transit. That is exactly what network slicing does for 5G networks — it partitions a single physical network into multiple virtual ones, each optimized for entirely different requirements. Ericsson calls it “differentiated connectivity” — and it is the technology poised to unlock a multi-billion-dollar market.
What Is Network Slicing?
Network slicing is a network architecture that enables the creation of multiple virtualized, logically independent networks (slices) on top of the same physical infrastructure. Each slice operates as an isolated, self-contained end-to-end network, tailored to the specific requirements of a given application or service.
The concept evolved from overlay networks and the PlanetLab research project, but only became practically viable with the arrival of 5G. The technology rests on two foundational pillars — SDN and NFV — combined with the new 5G Core architecture:
The Three Key Building Blocks
- SDN (Software-Defined Networking): Decouples the control plane from the data plane, enabling centralized, programmable management of the entire network
- NFV (Network Function Virtualization): Transforms network functions into software running on commodity hardware, replacing expensive, purpose-built appliances
- Orchestration: A central orchestrator automatically coordinates the creation, scaling, and teardown of each slice in real time, ensuring optimal resource allocation
Each slice is fully isolated: if one slice becomes overloaded or suffers a cyberattack, the others remain unaffected. The ITU (International Telecommunication Union) has defined three primary slice categories that map to the three major 5G use-case families:
- eMBB (enhanced Mobile Broadband): Maximum bandwidth for streaming, VR/AR, and cloud gaming — speeds up to 10 Gbps
- URLLC (Ultra-Reliable Low-Latency Communication): Sub-millisecond latency (<1 ms) and 99.999% reliability for autonomous vehicles, robotic surgery, and industrial control
- mMTC (massive Machine-Type Communication): Connecting up to 1 million devices per km² with ultra-low power consumption — IoT sensors, smart cities, agricultural networks
How Does It Work?
The network slicing architecture is built on three layers coordinated by a central orchestrator. Each slice is created dynamically, receives exactly the resources it needs, and is torn down automatically when no longer required:
The 3 Layers + Orchestrator
| Layer | Role | Examples |
|---|---|---|
| Service Layer | Defines business services and their requirements (SLAs) | Remote surgery, vehicle fleet, live broadcasting |
| Network Function Layer | Creates virtualized network functions (VNFs) for each slice | AMF, SMF, UPF — virtualized via NFV |
| Infrastructure Layer | Provides physical resources (compute, storage, networking) | 5G antennas, data centers, fiber optics |
| Slice Orchestrator | Coordinates creation, scaling, and teardown of slices automatically | NSSF, CSMF — automated through AIOps |
Key 5G Core functions that make slicing possible: the NSSF (Network Slice Selection Function) selects the right slice for each device, the CSMF translates business requirements into technical slice specifications, and the UPF (User Plane Function) routes user data to the correct slice. The technology reaches its full potential only with 5G Standalone (SA) and advanced 5G Core automation.
Critically, slice isolation is a foundational principle. If one slice suffers a DDoS attack or becomes congested during peak traffic, the remaining slices are unaffected — each maintains its guaranteed resources. Security is reinforced through end-to-end encryption per slice, zero-trust architecture within each slice (micro-segmentation), and AIOps systems that use machine learning to detect anomalies in real time.
Use Cases & Applications
The true potential of network slicing becomes clear in specialized industry applications. Each sector has radically different connectivity requirements — and now a single physical network can serve them all simultaneously:
Remote Broadcasting
Dedicated eMBB slices for live sports coverage — 4K/8K video feeds with guaranteed ultra-high bandwidth. Separate slices for spectator engagement apps and venue security cameras, all running on the same stadium infrastructure.
Mobile Cloud Gaming
Slices optimized for low latency (<10 ms) and high bandwidth simultaneously. AR/VR experiences on dedicated slices with zero buffering or lag, even in crowded venues like concerts or conventions.
Smart Manufacturing
URLLC slices for AGVs (Automated Guided Vehicles) on the factory floor with sub-millisecond latency. Real-time industrial control, predictive maintenance via IoT sensors on a parallel mMTC slice.
Healthcare & Remote Surgery
Exclusive URLLC slices for robotic surgery with 99.999% reliability. Isolated slices for medical data ensuring strict GDPR compliance. Patient monitoring through wearable devices on mMTC slices.
Autonomous Vehicles (C-V2X)
URLLC slices for Vehicle-to-Everything (V2X) communication with guaranteed QoS. Real-time decisions: emergency braking, obstacle avoidance, fleet coordination. Critical for Level 4+ autonomy.
Smart Cities
Millions of sensors on mMTC slices: street lighting, parking, air quality, waste management. Security cameras on eMBB slices. Emergency response systems on dedicated URLLC slices.
Network Slicing vs Private 5G vs Public 5G
Many enterprises wonder: do I need a private 5G network, or is a network slice enough? The answer depends on specific requirements — network slicing blends features of both public and private network models. A key advantage is that slices can be provisioned in hours rather than the weeks required for private 5G deployments, while still offering enterprise-grade isolation and guaranteed QoS:
Network Model Comparison
| Feature | Network Slice | Private 5G | Public 5G |
|---|---|---|---|
| Coverage | Flexible (local or nationwide) | Local only | Nationwide |
| Isolation | Logical (software-level) | Physical (hardware-level) | None |
| Cost | Moderate (€500–5,000/month) | High (€50,000+ setup) | Low (standard plan) |
| Deployment Speed | Hours to days | Weeks to months | Instant |
| Security | High (encrypted, isolated) | Very high (dedicated) | Basic |
| Customization | High (SLA per slice) | Full | Minimal |
In practice, many large enterprises opt for a hybrid approach: private 5G on their premises for mission-critical operations (e.g., robotics production lines), combined with network slices on the public network for inter-site connectivity, logistics, and remote workforce. Major operators such as Singtel, Telia, Rogers, and T-Mobile have already launched enterprise slicing services, proving the technology is no longer theoretical but commercially viable.
Business Model & Monetization
Network slicing is not just a technological achievement — it is a new business model. Telecom operators (CSPs) can now offer customized services instead of a one-size-fits-all data package. The sport event example illustrates this perfectly: at a major stadium, one operator can simultaneously run a dedicated broadcasting slice for TV crews, a spectator engagement slice for fan apps, a security slice for surveillance cameras, and a public safety slice for emergency services — all on the same infrastructure, each with different SLAs and pricing.
Revenue Models by Industry
- Healthcare: Premium URLLC slices for hospitals — guaranteed reliability for telemedicine and remote diagnostics
- Government & Public Safety: Dedicated slices for first responders and secure government communications
- Transport & Logistics: URLLC slices for autonomous vehicles, mMTC slices for fleet tracking
- Energy: IoT slices for smart grids, monitoring wind and solar farms
- Manufacturing (Industry 4.0): Campus-wide slices for smart manufacturing, robotics, and predictive maintenance
- Media & Entertainment: eMBB slices for live broadcasting, AR experiences at stadiums and festivals
Pricing models include: as-a-service (monthly subscription per slice with fixed SLA), dynamic pricing (demand-based billing — higher costs during peak hours for premium slices), and MVNO leasing (infrastructure providers rent physical resources to virtual operators who create their own slices under their own branding). Ericsson estimates that the network slicing market represents a “multi-billion-dollar opportunity” globally — a new chapter for telecoms that transforms the network from a cost center into a revenue engine.
Network Slicing in Greece
Network slicing requires a 5G Standalone (SA) core — without it, slices cannot function to their full potential. Today, Greek 5G networks primarily operate in Non-Standalone (NSA) mode, relying on the existing 4G core. For full network slicing, a native 5G core is required — something the major operators are expected to roll out gradually throughout 2026 and 2027.
Status in Greece (2026)
- Cosmote: Testing 5G SA core in lab environments. Piloting network slicing with enterprise clients — ports and logistics. Commercial availability expected during 2026–2027
- Vodafone: Partnering with Ericsson for 5G SA deployment. Running pilots in smart manufacturing and IoT applications. Focusing on enterprise solutions before retail
- Nova: Upgrading its 5G network toward SA readiness. Initially focusing on eMBB applications
The potential applications in the Greek market are particularly compelling: tourism-focused slices on islands for premium connectivity during the summer season (imagine guaranteed 4K streaming on a Mykonos beach in August), port operations at Piraeus with URLLC slices for automated container handling and AGVs, smart city applications in Athens and Thessaloniki for traffic management, public safety, and energy optimization, and agricultural IoT networks in Thessaly for precision farming on mMTC slices.
The rollout will be gradual, starting with enterprise customers (ports, factories, hospitals) and expanding to consumer services. The role of EETT (the national telecommunications regulator) will be critical in governing the new slicing services, particularly around pricing, quality-of-service guarantees, and data protection.
What Does the Future Hold?
The evolution of network slicing follows a three-step transformation journey, according to the major vendors:
Phase 1: Pre-configured Slicing
Pre-defined slices for basic categories (eMBB, URLLC, mMTC). Manual management. This is where early deployments stand today.
Phase 2: Dynamic Slicing
AI-driven slice creation in real time, automatic scaling based on demand. Slices adapt dynamically to changing conditions. Expected 2027–2028.
Phase 3: Exposed Slicing
APIs that allow third-party developers and enterprises to create their own slices through self-service portals. Full democratization of the network.
Artificial intelligence will play a central role in this evolution: AIOps systems will use machine learning for automatic anomaly detection, predictive slice maintenance, dynamic resource reallocation, and real-time SLA optimization. In 6G (post-2030), slices will become even more granular — “sub-slices” within slices — with native AI orchestration, fully automated lifecycle management, and intent-based networking where the customer describes what they need in natural language and the network automatically creates the appropriate slice.