A study of QoS performance for real time applications over a differentiated services network

Abstract

Abstract:The Differentiated Service Architecture (DiffServ) is currently a popular research topic as a lowcost method to provide Quality of Service to the different applications in the Internet. This paper addresses the definition and deployment of specific network services in a DiffServ environment. The proposed network services are described, while our main focus is the real-time delay-sensitive applications such as voice and video.The corresponding services for those applications are further analyzed and simulated. The analysis includes the influence of different algorithms that could implement the corresponding services for voice and video as well as the QoS performance under these algorithms. Sim ulation outcomes prove that the fundamental principles of the network services are fulfilled and provide an evaluation of the proposed algorithms.

Robust design of communication systems: The case of expedited forwarding of voice traffic in differentiated services networks

“…If the target packet loss rate is assumed to be 10 À2 ; then according to an admission control algorithm [40], the minimum SA is around 1.47. However, it turns out through the preliminary experiments that WFQ drops packets excessively when SA is set to 1.47, which suggests that larger SA values need to be investigated for the EF service in the present experiment.…”

“…Noise factor U: EF packet size. It is known that the packet size affects the performance characteristics of an IP network [40]. Since packets are available in various sizes depending on their application type and their sizes are uncontrollable to the network management, the EF packet size is considered as a noise factor in the present study.…”

“…Since packets are available in various sizes depending on their application type and their sizes are uncontrollable to the network management, the EF packet size is considered as a noise factor in the present study. In general, EF-based applications generate packets of small size ð5256 bytesÞ so as not to provoke long transmission delays [40]. Therefore, the two levels of the EF packet size are set to 40 and 160 bytes, respectively, as in Reference [40].…”

“…In general, EF-based applications generate packets of small size ð5256 bytesÞ so as not to provoke long transmission delays [40]. Therefore, the two levels of the EF packet size are set to 40 and 160 bytes, respectively, as in Reference [40].…”

A User-Driven Transmission Scheme for Voice Over IP Applications

“…A typical approach to address the delay issues in Voice over IP (VoIP) networks is packet scheduling. Presently, there have been several packet scheduling schemes proposed, including Weighted Fair Queuing (WFQ) [15][16], Priority Queuing (PQ) [11], Class-Based Queuing (CBQ) [12], Weighted Round Robin (WRR) Scheduling [14], or Priority Round Robin Scheduling (PRR) [5]. These schemes reduce the endto-end delay and improve the quality of voice over IP networks, however, performance improvement is based on aggregated voice flows.…”