Correlation of Bottleneck Metrics and Quality of Experience

Describing service quality from the end-users' point of view is fulfilled by using Quality of Experience (QoE) metrics. In a networking environment QoE metrics are very close to Quality of Service (QoS), but – because of the subjective nature of QoE – they could not be derived directly from QoS. Collection and analysis of QoS and SLS (Service Level Specification) properties of networking services are daily tasks of the operators. User satisfaction, however, cannot be mapped directly to these measures related to a network path or aggregated access traffic. Our ultimate aim is to find methods and metrics that correlate properly with QoE, and could be measured passively on aggregated network links. In this paper we present our experimental results on correlating the severity of network overload and the experienced service quality by using our advanced metrics on bottleneck detection. Our measurements were carried out in a real network environment, in which we loaded the network with various kinds of service requests and made notes on the perceived quality. We captured packet level traffic, and then derived measures based on packet interarrival times, packet size information and packet loss information. Our results show that the proposed bottleneck-detection metrics correlate much more to QoE than any common QoS metrics, thus could be used successfully as indicators of user satisfaction. BME – TMIT (2) APNOMS 2006 Introduction Quality of Service (QoS) Objective – Derived at measured interfaces – Typically: • Throughput • Loss • Delay • Jitter, etc. Quality of Experience (QoE) Subjective – Reflecting end-user experience – Related to • service unavailability, • response time, • audio/video quality, etc. Our aim: -Estimate QoE with passive, single-point measurements Find appropriate measures Among QoS metrics Among Bottleneck metrics