A Slicing Model for Transport Networks with Traffic Burst Control and QoS Compliance for Traffic Flows

TL;DR

This paper introduces a new resource control mechanism for transport network slices that manages traffic bursts and ensures QoS compliance, improving bandwidth utilization and slice performance.

Contribution

The paper presents a novel fine-grained control model for transport slices that handles traffic bursts and QoS, aligned with standards and tested experimentally.

Findings

Effective traffic burst control within slices

Maximized bandwidth utilization

Maintained QoS compliance

Abstract

Network slicing has emerged as a key network technology, providing network operators with the means to offer virtual networks to vertical users over a single physical network infrastructure. Recent research has resulted mainly in techniques for managing and deploying network slices, but the implementation of network slices on a real physical transport network infrastructure has received much less attention. Standardization bodies, such as the Internet Engineering Task Force (IETF), have provided some implementation recommendations. Still, there is a lack of mechanisms to implement network slices capable of handling traffic bursts while simultaneously meeting the Quality of Service (QoS) requirements of the traffic flows associated with the slices. In this paper, we propose a novel fine-grained resource control mechanism to implement transport network slices that meet traffic QoS requirements while both accepting limited traffic bursts, and enabling efficient bandwidth sharing within and across slices. The mechanism is executed at the edge of the transport network. The proposed model aligns with current standards on network slicing and has been tested on an experimental platform. Using this platform, we have conducted an extensive experimental campaign that demonstrates that our proposal can effectively control traffic bursts generated within the network slices while maximizing bandwidth utilization across the network.