Academic Commons


Multi-Layer Utilization Maximal Fairness for Multi-Rate Multimedia Sessions

Mendes, Paulo; Schulzrinne, Henning G.; Monteiro, Edmundo

We present a fairness model, called Multi-Layer Utilization Maximal Fairness (MLUM). The motivation of the MLUM model is to accomplish intra-session and inter-session fairness in the presence of multi-rate (multi-layer) multimedia sessions, distributing bandwidth between multicast trees of different sessions, considering the number of receivers in each session, and improving bandwidth efficiency. To accomplish this goal, the model is divided in three components: a fairness definition,a policy and a protocol. The MLUM definition guarantees fairness between receivers in a session and fairness between different sessions, considering the number of receivers in each session. The MLUM policy implements the MLUM definition in multicast aware border routers of Autonomous Systems (AS). It's functionality is divided in the control plane and the data plane. In the control plane, sessions fair rates are estimated based upon the MLUM fairness definition. In the data plane, a queueing discipline will fairly distribute bandwidth between sessions, considering their fair rates. The MLUM protocol allows the exchange of control information (sessions number of receivers and fair rates) between MLUM policy routers in order to accomplish a fair distribution of bandwidth between concurrent sessions. This paper aim to present the MLUM fairness definition, to describe the MLUM fairness policy control and data plane functionality and to evaluate how the policy data plane can fairly distribute bandwidth. We present some simulations that evaluate the performance of the MLUM policy data plane in different scenarios and that compare its performance to other queueing disciplines. We also briefly describe the MLUM protocol, due to its close relation with of the MLUM policy functionality.



More About This Work

Academic Units
Computer Science
Department of Computer Science, Columbia University
Columbia University Computer Science Technical Reports, CUCS-008-01
Published Here
April 22, 2011