Edge Native: Concepts, Technologies, and Architecture for 5G MEC

With the rapid growth of 5G industry applications in 2020, existing edge computing capabilities show shortcomings that prevent full satisfaction of diverse industry scenarios; the MEC platform cannot operate completely independent of the carrier 5G network, requiring tight network‑compute collaboration, which motivates the Edge Native concept for future network evolution.

The paper first examines the development status of edge computing, the market demand of 5G MEC, and the pain points in 5G MEC deployments, arguing that a new Edge Native approach is needed.

Edge Native is presented as an industry paradigm that tightly couples connectivity and computation, offering capabilities such as edge orchestration, edge intelligence, and edge security, forming the mid‑term target architecture for operators serving industry markets.

Edge Computing 2.0 is described with three deployment forms—cloud‑edge, edge‑cloud, and edge‑gateway—focusing on “edge‑cloud collaboration” and “edge intelligence” as core development directions, and requiring heterogeneous hardware (x86, ARM, GPU, NPU) and cloud‑native software platforms.

Statistics show that over 5,000 industry pilots were launched worldwide in 2020, with more than 60% involving MEC; Chinese operators have conducted 100+ 5G MEC pilots across 40+ cities, covering smart factories, mining, energy, etc.

The analysis identifies remaining gaps in network deployment, device capabilities, and software development, emphasizing the need for richer heterogeneous device support, stronger AI‑driven data analysis, higher reliability, system isolation, and data‑security mechanisms.

A typical MEC application development workflow is outlined: selecting a service framework based on language skills, defining REST APIs, leveraging 5G network capabilities via SDKs or templates, adapting to both IaaS (e.g., OpenStack, Kubernetes) and PaaS layers, and employing DevSecOps toolchains for end‑to‑end orchestration.

Edge Native parallels Cloud Native but extends it with telecom‑specific features; the architecture includes enhanced telecom network support, mixed container/VM orchestration, edge networking (5G open APIs, ETSI MEC), and a robust edge orchestrator that enables edge‑edge, edge‑cloud, and edge‑end collaboration.

Edge AI capabilities are discussed, including unified edge‑intelligent frameworks, heterogeneous AI hardware abstraction, and scheduling layers that hide hardware differences.

Edge Security requirements cover end‑to‑end protection, network isolation, data integrity, and the use of blockchain to store immutable metadata for trustworthy edge data.

Edge Data storage considerations include distributed consistency, heterogeneous storage abstraction, and lightweight solutions for IoT scenarios such as time‑series databases.

The paper concludes that while 2020 MEC pilots made progress, Edge Native remains at an early stage; further research and collaboration among alliances and open‑source communities are needed to mature the technology and drive industry adoption.