PLNOG23 - Marcin Aronowski

### Overview At PLNOG23, Marcin Aronowski delivers an introductory talk examining the realities of 5G standardization, deployment status, and architectural shifts. Drawing from over 15 years of telecommunications experience across network operators and equipment vendors, Aronowski contrasts the theoretical promises defined in high-level standards (such as ITU-R IMT-2020 and 3GPP specifications) against practical operational constraints. He outlines what 5G means today from both end-user and operator perspectives, debunking unrealistic expectations around immediate latency drops and multi-gigabit speeds.

### Key Topics - **Standardization Evolution & Pillars**: 5G high-level requirements evolved from initial ITU-R 2020 definitions (formalized around 2015) based on three primary pillars: massive machine-type communications (high device density), ultra-reliable low-latency communications (URLLC), and enhanced mobile broadband (higher throughput). 3GPP addressed this incrementally, splitting Release 15 into Non-Standalone (NSA) to quickly leverage existing 4G EPC cores with new 5G New Radio (NR), followed by Standalone (SA) 5G core specifications, while leaving full URLLC enhancements for Releases 16 and 17. - **Theoretical Targets vs. Real-World Deployments**: While 5G standards outline peak speeds up to 20 Gbps, latency under 1 ms, and densities of one million devices per square kilometer, current live production networks deliver more modest increments—typically 60 Mbps to near 2 Gbps depending on region and spectrum. Latency in current operational networks (such as SK Telecom trials) remains in the 18–19 millisecond range, comparable to optimized 4G LTE. - **Radio and Spectrum Realities**: Deploying millimeter-wave (mmWave) spectrum (such as 26–28 GHz) allows wider channel bandwidths but introduces severe physical attenuation and limited range (often under 100 meters, unable to penetrate solid walls). Building out nationwide coverage exclusively on mmWave would require impractical base station densities. - **Core Architectural Transformation**: Transitioning to full 5G requires operators to abandon legacy point-to-point hardware appliances in favor of service-based architectures (SBA), microservices, containers, network slicing, and edge computing to push processing closer to endpoints.

### Technical Details - **Non-Standalone (NSA) Option 3x**: Most commercial network operators deploy Option 3x, anchoring control in existing 4G Evolved Packet Core (EPC) infrastructure with minor software updates to interface parameters and radio equipment. - **Network Slicing & Edge Compute**: To physically reach sub-millisecond latencies, data processing cannot remain in centralized operator core networks; it must reside at the network edge. Network slicing enables dedicated radio and core resources per service rather than sharing a single pipe across all traffic types. - **Open APIs and Microservices**: The 5G core adopts cloud-native microservices architecture and exposed APIs, shifting telecommunications closer to modern IT and developer platforms, though requiring operators to modernize their operational deployment cycles.

### Takeaways - For end users, current 5G largely represents incremental throughput enhancements rather than a revolutionary leap in latency or novel services. - True low latency requires edge infrastructure deployment and Standalone (SA) 5G cores, which are still evolving. - Operators face major organizational and architectural hurdles, transitioning from rigid hardware-centric models to agile, software-defined containerized microservices.