Multi-domain Network Slice Partitioning: A Graph Neural Network Algorithm

TL;DR

This study tackles partitioning multi-domain network slices by distributing a VNF-FG across domains to minimize intra- and inter-domain costs while balancing load. It introduces a four-term objective incorporating DC, DL, IC, and a KL-divergence-based load-balancing term, with standardization to handle scale differences. The authors propose two heuristic ILP-based approaches and a Graph Neural Network partitioner (GNNP) to accelerate plan generation, and they evaluate performance on 200 VNF-FGs across four domains, comparing total cost and load-balancing alignment. Results show that GNNP achieves fast inferences with good alignment to target distributions, while Approximate ILP attains lower cost at the expense of KL adherence and Branch-and-Bound is slower; the work provides a scalable, flexible framework for rapid partitioning in multi-domain network slicing.

Abstract

In the context of multi-domain network slices, multiple domains need to work together to provide a service. The problem of determining which part of the service fits within which domain is referred to as slice partitioning. The partitioning of multi-domain network slices poses a challenging problem, particularly when striving to strike the right balance between inter-domain and intra-domain costs, as well as ensuring optimal load distribution within each domain. To approach the optimal partition solution while maintaining load balance between domains, a framework has been proposed. This framework not only generates partition plans with various characteristics but also employs a Graph Neural Network solver, which significantly reduces the plan generation time. The proposed approach is promising in generating partition plans for multi-domain network slices and is expected to improve the overall performance of the network.

Figures (3)

• Figure 1: GNN Partitioner
• Figure 2: Example of VNF-FG
• Figure 5: KL divergency value distribution