Statistical characterization for per-hop QoS

Authors:
Mohamed El-Gendy;Abhijit Bose;Haining Wang;Kang G. Shin
Affiliations:
Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI;Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI;Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI;Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI
Venue:
IWQoS'03 Proceedings of the 11th international conference on Quality of service
Year:
2003

Citing 3
Cited 2

End-to-end performance analysis with traffic aggregation

Computer Networks: The International Journal of Computer and Telecommunications Networking - Pioneering tomorrow's Internet Selected papers from the TERENA Networking Conference 2000 22–25 May 2000, Lisbon, Portugal
On the impact of policing and rate guarantees in DiffServ networks: a video streaming application perspective

Proceedings of the 2001 conference on Applications, technologies, architectures, and protocols for computer communications
Preferential Treatment of Acknowledgment Packets in a Differentiated Services Network

IWQoS '01 Proceedings of the 9th International Workshop on Quality of Service

IP Easy-pass: a light-weight network-edge resource access control

IEEE/ACM Transactions on Networking (TON)
Sapphire: Statistical Characterization and Model-Based Adaptation of Networked Applications

IEEE Transactions on Parallel and Distributed Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

In Differentiated Services (DiffServ) architecture, we define the per-hop QoS as the throughput, delay, jitter, and loss rate experienced by traffic crossing a Per-Hop Behavior (PHB). In this paper, we use a statistical approach that is based on experiments on a real network testbed to characterize the per-hop QoS of a given PHB. Specifically, we employ a full factorial statistical design of experiments to study the effects of different PHB configurations and input traffic scenarios on perhop QoS. We use Analysis of Variance (ANOVA) to identify the input and PHB configuration parameters that have the most significant influence on per-hop QoS. Then, multiple regression analysis is applied to construct models for the per-hop QoS with respect to these parameters. The overall approach is shown to be effective and capable of characterizing any given PHB, within the ranges of the experiments, and for construction of functional relationships for the PHB output parameters. The approach in this paper forms a "fundamental" step towards achieving predictable end-to-end QoS when applying statistical QoS control at intermediate nodes.