An Adaptive Mechanism for End-to-End Multirate Multicast Congestion Control

Abstract

One of the most significant challenges that confront End-to-End Multicast Congestion Control is providing a scalable solution that addresses the heterogeneity of the receivers. There exist two categories of rate based multicast congestion control approaches. The first category is a single rate where the source adjusts the sending rate dynamically according to the slowest receiver within the session. The second category is the Multirate where the source has to disseminate the data in multiple-layers and each receiver could subscribe to the desired layer according to its own capacities. Multirate mechanisms exhibit better flexibility at the expense of a higher complexity. Furthermore, they are less mature than Single-Rate schemes. This paper proposes an improved version of a recently proposed scheme SMCC. The first improvement consists of using an equation to estimate the rates of the individual layers. The second improvement is with respect to the Join/Leave rules followed by the multicast receivers to migrate between the layers. Experimental results show remarkable performance improvement with respect to achieved throughput, without affecting the TCP friendliness property of the protocol.

Efficient Joining and Leaving for Receiver Driven Multicast Congestion Control

“…There are many schemes available in this category such as RLM [10], RLC [4], FLID-DL [6], TFMCC [8], SMCC [5], ESMCC [15] etc. The brief description of related propose approach are following Receiver driven Multicast layered (RLM) [10] is a transmission scheme which use layered approach for stream transmission.…”

“…The join attempt is ceased if a loss event occurs meanwhile. Extended Smooth Multirate Multicast Congestion Control (ESMCC) [15] is extension of SMCC which tackle the problem of SMCC using adaptive RTT. In this approach joining or leaving operation are based on RTT.…”

“…We have compare the performance to existing approach such RLM [10] SMCC [5], ESMCC [15] but we are taking receiver driven layered multicast (RLM) as an example to describe our approach. RLM is adaptive approach to carry out active experiment by addition a layer at fixed time.…”

“…The authors provide the cumulative, in which receiver join layer in increment order (base layer then enhance layers) whereas leave the layers in reverse order and in case of non-commutative receiver joins or leave the layers without order. There are may approach QIRLM [14], PLM [9], TFMCC [8], MILDA [13], RLM [10], RLC [4], FLID-DL [6], SMCC [5], ESMCC [15] etc. which are using single layer joining approach at a time and direct leave the layer which are reason for underutilization of resources and provide the less performance.…”