User-perceived Quality of Service in Wireless Data Networks

Anthony J. Saliba, Michael A. Beresford, Milosh Ivanovich, Paul Fitzpatrick, User-perceived quality of service in wireless data networks, Personal and Ubiquitous Computing, Volume 9, Issue 6, Dec 2005, Pages 413 – 422 is a paper that brings the computer science and a rather quantitative perspective to my research. I would classify it in the Human-Network Interaction category, because the machine is rather irrelevant in their study that illustrates the advantages of defining and assessing user-perceived quality of service (QoS) when dimensioning critical network parameters for optimized network performance. The authors are from the Telstra Research Labs.

In the past the term of QoS was a way to dimension wireless networks to run in the most efficient way possible. That is assuming that optimization of performance at the lower network layers will translate directly into an improved user experience. Nowadays, there is a trend reversal, looking at the user perceptions of the network performance to decide where dimensioning can have the greatest impact. Effective network performance becomes less about network characterisitcs and efficiency and more about satisfactory of having the user experience drive the notion of QoS. However, QoS is very application and context-specific because various applications require different levels of network performance to satisfy users. The growth in the wireless data sector rests upon two critical dependencies: network capability and a positive overall user experience. Thus, by understanding the relationship between the user experience and perceived quality of the service and the associated network performance, conclusion can be drawn as the required network performance to provide the necessary service quality to users on an ongoing basis.

QoS is often defined in terms of network performance or network characteristics while aspected of the user experience also need to be captured and assessed accurately using tools and metrics. A user’s behavior is most accurately predicted by analysing their mental constructs, context of interaction, motivations and the tasks they perform while using the network services [Distributed Multimedia and QOS: A Survey].

Studies on network QoS usually identify four main factors:

• Reliability: how important and useful it is to know in advance the level of network performance
• Efficiency: a measure of how quickly the system responds to requests
• Predictability: the degree to which the user experience followed the expectations of the users
• Satisfaction: what degree the user was satisfied with each experience.

The authors propose a wireless QoS hierarchy fitting into the open systems interconnection (OSI) layers to obtain an in-depth understanding of each layer’s performance and the relationship between layers in the hierarchy.