ETRI-Knowledge Sharing Plaform

KOREAN
논문 검색
Type SCI
Year ~ Keyword

Detail

Journal Article AI-native 크로스-도메인 6G 네트워크 기술동향
Cited - time in scopus Download 19 time Share share facebook twitter linkedin kakaostory
Authors
이도영, 홍성태, 김태연
Issue Date
2026-02
Citation
전자통신동향분석, v.41, no.1, pp.11-21
ISSN
1225-6455
Publisher
한국전자통신연구원
Language
Korean
Type
Journal Article
DOI
https://dx.doi.org/10.22648/ETRI.2025.J.410102
Abstract
As 6G networks evolve beyond the capabilities of 5G, artificial intelligence (AI) is expected to become intrinsic to the network҆s design and operation rather than merely a supplementary tool. Such networks are called “AI-native” networks. This paper investigates the technical trends and architectural evolution of AI-native cross-domain 6G networks. We analyze the paradigm shift towards end-to-end network automation, in which AI is deeply integrated across all domains, ranging from the radio access network (RAN) and core network to the transport and service management and orchestration layers. The key enabling technologies driving this transformation, including the 3GPP network data analytics function, O-RAN RAN intelligent controller, and emerging application of large language model-based agents for intent-driven automation, are examined in detail. Furthermore, this study emphasizes the importance of unified frameworks for data and model lifecycle management (DataOps/MLOps) and the critical role of network digital twins in ensuring the stability and reliability of AI-driven policies. Finally, we identify significant technical challenges, such as real-time inference latency, trustworthiness, and cross-domain conflict resolution, and outline the future research directions required to realize fully autonomous 6G networks.
Keyword
6G, AI-native Network, Cross-domain Optimization, End-to-End Automation, Network Digital Twin, O-RAN
KSP Keywords
AND operation, Architectural Evolution, Core Network, Cross Domain, Data Analytics, Design and operation, Digital Twin, Domain optimization, Enabling technologies, End to End(E2E), Future research directions
This work is distributed under the term of Korea Open Government License (KOGL)
(Type 4: : Type 1 + Commercial Use Prohibition+Change Prohibition)
Type 4: