This paper contains a detailed description of the congestion control algorithm of TRAM, a tree-based reliable multicast protocol. This algorithm takes advantage of regular acknowledgements from the receivers that propagate back to the sender via the repair tree. This scalable feedback mechanism is used to collect receiver credits. Complementing the windowing mechanism, packet transmission is smoothed by using a data rate commensurate with the window size. Additional details, such as how to prune slow receivers, and how to implement the rate scheduler on non-real-time systems are also discussed. The performance of the congestion control algorithm is then evaluated in extended LANs, and wide area networks. The fairness of bandwidth-sharing with other (TCP) traffic is also evaluated.

Current reliable multicast protocols do not have scalable congestion control mechanisms and this deficiency leads to concerns that multicast deployment may endanger stability of the network. In this paper, we present a sender-based approach for multicast congestion control targeted towards reliable bulk data transfer. We assume that there are a few bottleneck links in a large scale multicast group at any time period and these bottlenecks persist long enough to be identified and adapted to. Our work focus on dynamically identifying the worst congested path in the multicast tree and obtaining TCP-friendly throughput on this selected path. We devise novel selection (amongst receivers) and aggregation (over time) methods to achieve our goal. The response time of our protocol is then compatible to TCP once the worst path is identified. Only when switching between worst paths, the protocol response time is relaxed to multiple RTTs (less than 10) for the reasons of scalability and stability. We use the network simulator (NS2) to validate and evaluate our congestion control algorithm with both drop-tail and RED gateways.

This paperpresen ts a case studyin implemen tin a moderately complex, useful service on an activen work platform. The active application is reliable multicast withcon'O con trol; the platform comprises the Bowman Node Operatin Systeman the Composable Active Network Elemen ts (CANEs)Execution En vironio t. The importan of the work stems from thelesson it provides about thedesign an implemen tation of active platformsin genorm an BowmanOR in particular. For example, our experien shows that timer-driven activen ode processin is as importan t as packet-arrival-driven processin' Thus, execution en vironon tscan focus exclusivelyon forwardin' but must also providee#cien t timers an allow timerhanrO the same capabilities as packet-driven computationR Other areasin which the implemen tation provides inesO tin service decomposition approaches for active application an inica-O' sharin amon service compon ts. I. Introducti Active networfl pr vide aplatfor for networ ser vices that can be builtor customized by injecting code or other infor]B(/K into the nodes of the networ; For the pur oses of this pap er the salient char;;K(/ istic of "active networ[[flS is the placement ofuser contr(][S;( (for some definition of"user3 computing capabilities inshar[ infr[O(/KB[W3 of the communication networS wher they can be utilized by applications that need those capabilities. Thispar(K3K o#er a number of potential advantages, including the ability to develop and deploy new networ prS; cols andser]3]B quickly, and the ability to customizeser vices to meet the needs ofdi#erW t classes ofuserfl As an example of a desirK(/ networ serSWO( consider thepr(;OS of multicast applications that need red(SBfl]fl y anddesir toshar bandwidth cooperK tively withother applications. Thepr(KK t IP multicast...

We propose an end-to-end single-rate source-based multicast congestion control scheme (LE-SBCC) for reliable or unreliable multicast transport protocols. It addresses key issues such as drop-to-zero issues [27], TCP friendliness [1] and RTT estimation. The scheme consists of a cascaded set of filters and a rate adaptation policy module (AIMD or TFRC [9]) which transform the multicast tree to appear like a unicast path for the purposes of congestion control. The scheme is not self-clocked but acts upon a stream of loss indications (LIs), which are filtered to get a stream of loss events (LEs) [9] (at most one per RTT per receiver). This LE stream is further filtered to extract the maximum LEs from any one receiver, which results in at most one rate-reduction per RTT. A range of results (simulation and experimental) is presented and compared against the mathematical model of the scheme components.

A key issue in the design of source-based multicast congestion control schemes is how to aggregate loss indications from multiple receivers into a single rate control decision at the source. Such aggregation entails filtering out a portion of the loss indications received by the source, and then using the remaining for rate adjustments. In this paper, we first propose a set of goals guiding the design of loss indication filters. We then present a novel loss indication filtering approach, the Linear Proportional Response (LPR) approach. Analysis and simulation is used to compare LPR to two well-known approaches -- the Random Listening Algorithm (RLA) ([1]), and the Worst Estimate-Based Tracking (WET) [2] approach. Our results indicate that LPR achieves a desirable tradeoff between stability and response, thereby making it more suitable than WET and RLA for deployment in an Internet-like environment.

The danger of congestion collapse, and the role of congestion control in the Internet. Change and heterogeneity as conditions of the Internet. Speculations on the future evolution of end-to-end congestion control in the Internet. 2

This paper presents a simple, generic source-based end-to-end multicast congestion control (GSC) algorithm for reliable multicast transport (RMT) protocols. The algorithm is completely implemented at the source and leverages the reverse control information ow in RMT protocols like PGM or RMTP [13,44]. Speci cally, itdoes not introduce any new control tra c or new elds in RMT protocol headers. It addresses the drop-to-zero problem [43] by introducing a robust, adaptive timelter based upon RTT estimates collected by observing NAK tra c. This solution allows it to scale for large multicast groups while being very adaptive to congestion situation changes in any part of the tree. The algorithm is friendly to TCP in terms of competition for bandwidth shares. The scheme has minimal control tra c requirements and weak RTT estimation requirements which allows a large deployment space including multi-sender multicast and combination with receiver-based schemes. Key words: reliable multicast, congestion control, transport protocol 1

Abstract — A significant impediment to deployment of multicast services is the daunting technical complexity of developing, testing and validating congestion control protocols fit for wide-area deployment. Protocols such as pgmcc and TFMCC have recently made considerable progress on the single rate case, i.e. where one dynamic reception rate is maintained for all receivers in the session. However, these protocols have limited applicability, since scaling to session sizes beyond tens of participants necessitates the use of multiple rate protocols. Unfortunately, while existing multiple rate protocols exhibit better scalability, they are both less mature than single rate protocols and suffer from high complexity. We propose a new approach to multiple rate congestion control that leverages proven single rate congestion control methods by orchestrating an ensemble of independently controlled single rate sessions. We describe a new multiple rate congestion control algorithm for layered multicast sessions that employs a single rate multicast congestion control as the primary underlying control mechanism for each layer. Our new scheme combines the benefits of single rate congestion control with the scalability and flexibility of multiple rates to provide a sound multiple rate multicast congestion control policy. I.

This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet- Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at

This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026 [1]. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or become obsolete by other documents at any time. It is inappropriate to use Internet- Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at