SLA Decomposition for Network Slicing: A Deep Neural Network Approach

TL;DR

This paper presents a neural network-based method for decomposing end-to-end SLAs in multi-domain network slicing, leveraging historical data to ensure SLA compliance across domains.

Contribution

It introduces NN-based risk models with monotonicity constraints for SLA decomposition, effective even with limited data, and demonstrates their efficiency on synthetic datasets.

Findings

Neural network models effectively decompose SLAs across domains.

Monotonicity constraints improve model performance with small datasets.

Empirical results show the approach's efficiency on synthetic data.

Abstract

For a network slice that spans multiple technology and/or administrative domains, these domains must ensure that the slice's End-to-End (E2E) Service Level Agreement (SLA) is met. Thus, the E2E SLA should be decomposed to partial SLAs, assigned to each of these domains. Assuming a two level management architecture consisting of an E2E service orchestrator and local domain controllers, we consider that the former is only aware of historical data of the local controllers' responses to previous slice requests, and captures this knowledge in a risk model per domain. In this study, we propose the use of Neural Network (NN) based risk models, using such historical data, to decompose the E2E SLA. Specifically, we introduce models that incorporate monotonicity, applicable even in cases involving small datasets. An empirical study on a synthetic multidomain dataset demonstrates the efficiency of our approach.