Page 20 - Kaleidoscope Academic Conference Proceedings 2021
P. 20
Technologies for cloud-native systems are well established and widely deployed in today’s data centers and
cloud. To make 6G system a native part of the cloud, new technology capabilities and design extensions are
required, including server/service discovery for device mobility, cross-domain computing with robust security,
design enhancements for radio link performance, end-to-end orchestration and resource management across
devices, network nodes and data centers, to name a few. As Artificial Intelligence (AI) and Machine Learning
(ML) become mainstream, a 6G system is expected to be designed to natively integrate AI technologies, which
is now often referred to as “AI-native”. On the one hand, many aspects of 6G system design can directly benefit
from AI/ML capabilities for better performance; on the other hand, 6G systems need to be architected for
efficient in-network data management and access for machine learning and inference.
Making 6G cloud-native and AI-native touches on every part of the system, from air interface design, to system
and network architecture, to service manager, to the software and hardware platform as well as the need for a
programmable high-performance transport network. As 6G becomes further embedded in every aspect of
human life and into our society, system reliability, trustworthiness, privacy, and information security are of
paramount importance to our new system. The picture below offers a high-level view on the key features that
are currently under research and development. A cloud-native design establishes a solid platform foundation
for the next generation system. As ML/AI continue to evolve, from algorithms to data processing to
applications, a cloud-native environment needs to be a highly scalable design to enable and to support existing
and future AI models and workload.
As 6G expands its technology dimensions beyond wireless connectivity, selecting and prioritizing a correct
set of fundamental capabilities, technology building blocks and network architectures become more critical
and challenging than for earlier generations. This increased system complexity and the interdependency
between communication, computing and data management also add new requirements to the standardization
process, where open architecture, open implementation/open source will become an integral part of the 6G
industry standards in addition to the traditional normative specifications.
Every generation of wireless technologies needs to bring a new level of experience and a set of new capabilities
to enable future applications and services, 6G is no exception. This talk provides an overview of 6G cloud-
native and AI-native design considerations, which are the fundamental new capabilities for next generation
systems.
– xvi –