one6G Work Item 207 contributors present a paper at 6GNet2024 in Paris
On October 23, 2024, one6G representatives presented a paper titled “In-Network Computing and Split-AI in 6G” at the 3rd International Conference on 6G Networking (6GNet2024) in Châtillon, near Paris, France.
The paper, authored by academic and industrial experts from Docomo (Germany), University of Calabria and CNIT (Italy), University of Aveiro (Portugal), University of Thessaly (Greece), Huawei (Germany), and NTNU (Norway) results from joint activities carried out within the one6G Work Item 207 -Intelligent User Plane and In-Network Computing.
The paper explores the integration of In-Network Computing to enhance the efficiency of AI-based applications within 6G networks. Theauthors demonstrate how INC-assisted Split-AI, which distributes neural network execution across multiple network nodes, can reduce latency, lower mobile device energy consumption, and optimize network resource usage. The paper also introduces key architectural enhancements in the Control Plane and User Plane of 6G networks, necessary to natively support these capabilities. These proposals are validated through a proof-of-concept testbed through which the study highlights the advantages of INC-assisted Split-AI technologies.
Paper abstract
Split-AI proposes a paradigm shift in which Neural Networks (NN)-relevant tasks are distributed across multiple networking entities. The capabilities envisaged for 6G mobile networking such as In-Network Computing (INC) can facilitate a smooth deployment of this technique. 6G’s Access Nodes (ANs) and User Plane Functions (UPFs) will be able to execute parts of an NN while performing their traditional functions (e.g. transmitting packets). Recent studies show that INC-assisted SplitAI has the potential to enhance Key Performance Indicators (KPIs) like inference time, network traffic load, and UE energy consumption. The incorporation of Split-AI in 6G networks needs to fulfill significant design and performance requirements such as optimal partitioning strategies for distributing NN parts across network elements. Additionally, the 6G User Plane must enable the dynamic and flexible allocation and deployment of these split NN parts. In this work we: i) propose possible enhancements for the Control Plane (CP) and UP, to make possible the integration of INC-assisted Split-AI in 6G networks; ii) present the design of a Proof-Of-Concept (PoC) simulation model aimed at demonstrating the feasibility of the technique with a comprehensive study on the impact that different NN partitioning options have on the network utilization, inference time, and the resources of both UE and network entities involved.
About 6GNet
6GNet is an annual conference, established in 2022, that covers advances in the area of 6G network design and management, encompassing enabling technologies, architectures and services. The conference provides a forum for researchers, students, and professionals to exchange ideas, share their experiences, and discuss advances and challenges for the advent of 6G networks. The 2024 edition took place in Châtillon, near Paris, France, on October 21-24, 2024.