In Search of Lost QoS

TL;DR

This paper reviews the history and challenges of QoS in networks, highlighting its limited practical adoption and proposing incentive-based schemes to improve its relevance and implementation.

Contribution

It provides a comprehensive 30-year review of QoS research, analyzes its phases, and introduces a novel incentive-based QoS scheme with practical recommendations.

Findings

QoS research has passed through multiple phases with changing focus.

Practical deployment of QoS remains limited despite extensive research.

An incentive-based QoS scheme can address some of the challenges identified.

Abstract

The area of quality of service (QoS) in communications networks has been the target of research for already several decades with tens of thousands of published journal and conference papers. However, the practical introduction of QoS systems in commercial networks has been limited (with a preference for simple overprovisioning). Despite this dissonance, most influential QoS papers do not discuss this lack of penetration or challenge any of the common assumptions used to argue for QoS systems. So far, the few critical QoS papers have had only a minor effect on QoS research and standardization. Therefore, there is a serious risk that QoS will remain an academic research topic without significant practical relevance. To help elucidate these issues, in this work, we first perform a comprehensive review of QoS including a general overview and an analysis of both influential and critical work from the past 30 years. We examine properties such as citations, keywords, and author traits to show that QoS has passed through several distinct phases with different topics while maintaining the overall attitude towards the role and objective of QoS systems. We then discuss QoS as a social phenomenon and in the context of current networking standards. Finally, we propose a QoS scheme based on incentives that avoids some of the problems identified in critical work, and we provide simple recommendations for network operators. Overall, we hope to spark the community to take a fresh look at QoS.