Content area
Full text
Abstract-Burst ratio is a parameter that quantifies packet loss patterns in transmission networks. It has been defined for an end-to-end scenario, therefore burst ratio can be determined only if the characteristics of the whole transmission path are known. In this paper, the burst ratio parameter applicability to cases when the transmission path consists of a series of transmission channels with known packet loss rate and burst ratio values is extended. The paper also presents the results of simulations performed with NS2 software, demonstrating the validity of the burst ratio analysis. Consequently, the research makes it possible to determine the value of the burst ratio parameter in concatenated packet networks, which in turn supports delivering higher quality VoIP services.
Keywords-bursty packet loss, E-model, quality of experience, voice over IP.
(ProQuest: ... denotes formulae omitted.)
1. Introduction
Voice over Internet Protocol (VoIP) applications play a crucial role in connecting people and businesses around the world. It is a huge business for hardware manufacturers, network operators and service providers. In order to assure end customer satisfaction, the transmission networks must be designed well, and the quality of the provided VoIP service must be constantly monitored and maintained. In order to achieve this, all factors that affect the application quality of experience (QoE) [1] must be recognized.
The quality of VoIP carried over packet networks is influenced by multiple factors [2]. They include user-dependent aspects (e.g. user expectations), terminal quality (e.g. microphone sensitivity) and application settings (e.g. audio codec). The quality is also affected by transmission network-dependent factors, which include throughput, round-trip time and packet loss. To some extent, they can be controlled by network design and maintenance.
One of the transmission network-dependent factors that influences the perceived quality of VoIP transmissions is the burst ratio parameter [3]. It quantifies the packet loss pattern by describing the extent to which the packets were lost in bursts. The burstiness of packet loss affects the perceived media quality. If the number of audio packets lost sequentially is low enough not to be noticed by the human cognitive system, or it can be concealed by the packet loss concealment (PLC) technique [4], then the event has no impact on the perceived quality. In contrast, long sequences of lost packets can be...