This graphical abstract gives an overview of the systematic review approach taken in this paper to understand the state of the art in literature on the deployment and man...
Abstract:
Predictable and dynamic support of custom slices in 5G will be aided by integrating intelligence into the network using machine-learning techniques. However, this idea is...Show MoreMetadata
Abstract:
Predictable and dynamic support of custom slices in 5G will be aided by integrating intelligence into the network using machine-learning techniques. However, this idea is still in its conceptual and infancy stages due to the slow adoption and advancement of practical deployments of intelligent machine-learning techniques in the context of the life cycle of a 5G network slice. In this work, we considered the challenges that contribute to not achieving the vision of embedding intelligence in network slicing. These challenges are the lack of freely available end-to-end 5G networking datasets and the absence of easily replicable end-to-end implementations of 5G network slices using open-source software on commercial off-the-shelf network devices. Therefore, this paper addresses these challenges by conducting a systematic review that focuses on the literature that has attempted to study the adoption of intelligence in end-to-end 5G network slicing. Since this study area is multidisciplinary, overlapping the fields of artificial intelligence, machine-learning, mobile networks, etc., we take the approach of formulating five research questions that contribute to the goal of this systematic review. The 5 questions are centered around the themes of 1) data collection procedures, toolkits, and strategies, 2) actual 5G open-source datasets with a description of associated features, 3) real-world implementations of end-to-end 5G network slices on physical hardware with low-level implementation details, 4) strategies of embedding intelligence using machine-learning techniques in such networks, and 5) possibilities of designing a network slice template before deployment using unsupervised machine-learning.
This graphical abstract gives an overview of the systematic review approach taken in this paper to understand the state of the art in literature on the deployment and man...
Published in: IEEE Access ( Volume: 12)