Hash tables – which map “keys ” onto “values” – are an essential building block in modern software systems. We believe a similar functionality would be equally valuable to large distributed systems. In this paper, we introduce the concept of a Content-Addressable Network (CAN) as a distributed infrastructure that provides hash table-like functionality on Internet-like scales. The CAN is scalable, fault-tolerant and completely self-organizing, and we demonstrate its scalability, robustness and low-latency properties through simulation.

Most currently proposed solutions to application-level multicast organize the group members into an application-level mesh over which a DistanceVector routing protocol, or a similar algorithm, is used to construct source-rooted distribution trees. The use of a global routing protocol limits the scalability of these systems. Other proposed solutions that scale to larger numbers of receivers do so by restricting the multicast service model to be single-sourced. In this paper, we propose an application-level multicast scheme capable of scaling to large group sizes without restricting the service model to a single source. Our scheme builds on recent work on Content-Addressable Networks (CANs). Extending the CAN framework to support multicast comes at trivial additional cost and, because of the structured nature of CAN topologies, obviates the need for a multicast routing algorithm. Given the deployment of a distributed infrastructure such as a CAN, we believe our CAN-based multicast scheme offers the dual advantages of simplicity and scalability.

Although IP Multicast is an effective network primitive for best-effort, large-scale, multi-point communication, many multicast applications such as shared whiteboards, multi-player games and software distribution require reliable data delivery. Building services like reliable sequenced delivery on top of IP Multicast has proven to be a hard problem. The enormous extent of network and end-system heterogeneity in multipoint communication exacerbates the design of scalable end-to-end reliable multicast protocols. In this paper, we propose a radical departure from the traditional end-to-end model for reliable multicast and instead propose a hybrid approach that leverages the successes of unicast reliability protocols such as TCP while retaining the efficiency of IP multicast for multi-point data delivery. Our approach splits a large heterogeneous reliable multicast session into a number of multicast data groups of co-located homogeneous participants. A collection of application-aware agents--Reliable Multicast proxies (RMXs)--organizes these data groups into a spanning tree using an overlay network of TCP connections. Sources transmit data to their local group, and the RNLX in that group forwards the data towards the rest of the data groups. RMXs use detailed knowledge of application semantics to adapt to the effects of heterogeneity in the environment. To demonstrate the efficacy of our architecture, we have built a prototype implementation that can be customized for different kinds of applications.

The efficacy of end-to-end multicast transport protocols depends critically upon their ability to scale efficiently to a large number of receivers. Several research multicast protocols attempt to achieve this high scalability by identifying sets of co-located receivers in order to enhance loss recovery, congestion control and so forth. A number of these schemes could be enhanced and simplified by some level of explicit knowledge of the topology of the multicast distribution tree, the value of the bottleneck bandwidth along the path between the source and each individual receiver and the approximate location of the bottlenecks in the tree. In this paper, we explore the problem of inferring the internal structure of a multicast distribution tree using only observations made at the end hosts. By noting correlations of loss patterns across the receiver set and by measuring how the network perturbs the fine-grained timing structure of the packets sent from the source, we can determine both ...

The current Internet does not offer any quality of service guarantees or support to Internet multimedia applications such as Internet telephony and video-conferencing, due to the besteffort nature of the Internet. Their performance may be adversely affected by network congestion. Also, since these applications commonly employ the UDP transport protocol, which lacks congestion control mechanisms, they may severely overload the network and starve other applications. We present an overview of recent research efforts in developing adaptive delivery models for Internet multimedia applications, which dynamically adjust the transmission rate according to network conditions. We classify the approaches used to develop adaptive delivery models with brief descriptions of representative research work. We then evaluate the approaches based on important design issues and performance criteria, such as the scalability of the control mechanism, responsiveness in detecting and reacting to congestion, and ability to accommodate receiver heterogeniety. Some conclusions are developed regarding the suitability of particular design choices under various conditions.

Abstract: We present a network model that allows processing of QoS (quality of service) information about media flows to enable applications to make adaptation decisions. Our model is based around a multi-dimensional spatial representation that allows QoS information describing the flow constructions and QoS parameters – flow-states – to interact with a representation of the network QoS. The model produces reports about the compatibility between the flow-states and the network QoS, indicating which flow-states the network can currently support. The simple nature of the reports allows the application to make decisions, dynamically, on which flow-state it should use. The model is relatively lightweight and scaleable. We demonstrate the use of the model by simulation of a dynamically adaptive audio tool. Our work is ongoing.

The IP service model retains its simplicity and robustness by deferring reliability and congestion control to higher layers through end-to-end transport protocols. While the IP unicast service has proven successful, extending end-to-end adaptation to multicast has been a difficult problem. Unlike the unicast case, multicast protocols must support large and heterogeneous receiver sets. While proposed approaches to multicast transports attempt to localize problems and/or organize receivers into a hierarchy through a divide-and-conquer approach, this approach succeeds only if the resulting hierarchy is congruent with the underlying routing tree topology. This implies the need for some level of topological information at the end systems which the IP multicast service deliberately hides. In this paper, we explore the problem of inferring the required topological information using only observations made at the end hosts. To this end, we present a Group Formation Protocol (GFP) whereby receivers dynamically organize themselves into a multi-level hierarchy of multicast groups that corresponds to the underlying routing tree. GFP can serve as a core component across a wide range of multicast applications and protocols such as local recovery for reliable multicast, self organized transcoding, self organizing web caches, the optimal and dynamic placement of proxies, repeaters, designated receivers, recorders and so forth. Our simulations indicate that GFP structures receivers in accordance with the underlying topology for a range of workloads and network topologies.

We present Dagster, an end-host multicast scheme for delivering multimedia streams. Unlike previous schemes, Dagster does not constrain the amount of bandwidth a node must donate. Instead, it relies on a novel incentive scheme to encourage nodes to contribute more bandwidth to improve the total capacity of the system. The key idea behind the incentive is that Dagster allows a node with more donated bandwidth to preempt another node in the system. Dagster also allows a host to receive from multiple parents at the same time, thus is more resilient to node failures. Our simulation results show that Dagster’s design leads to low rejection rate, high resilience to failure and smaller diameter. Furthermore, nodes that donate more bandwidth have lower rejection rate and are positioned fewer overlay hops away from the source, providing incentives for nodes to increase their contribution. 1.

We introduce a scalable application-level QoS adaptation service for roaming between wireless networks that are based on different technologies (‘inter-tech’ roaming). The service is part of a platform that supports the distribution of multimedia streams (e.g., a streamed TV channel) to mobile clients operating in a heterogeneous environment. Central to our approach is the notion of a service class, which is a domain-specific perceptual QoS level. Each domain in a wireless infrastructure must support a limited number of these service classes. Our adaptation service handles inter-tech roaming by handing a client off from one service class to another. In this paper, we focus on the design of the adaptation service’s client-side components. They combine the loss characteristics of the client’s network interfaces with configurable policies to decide when to initiate a handoff to a target service class and when to complete it. We conclude with some experimental resuults. 1.

Distributed Partial Mixing is an approach to creating a distributed audio service that supports optimisation of bandwidth utilization across multiple related audio streams (e.g. from concurrently active audio sources) while maintaining fairness to TCP traffic in best effort networks. Rate adaptation of streamed audio is difficult because of its rate sensitivity, the relatively limited range of encoding bandwidths available and the potential impact on the end user of rate-adaptation artefacts (such as changes of encoding). This paper describes and demonstrates how our design combines TCP-fairness with the stability that is desirable for streaming audio and other rate sensitive media. In particular, our design combines: a distributed multi-stream management/mixing architecture, loss event and round-trip time monitoring, rate limiting based on a TCP rate equation, tuned increase and decrease strategies and a loss-driven network probing mode. Experimental validation is performed over a wide range of network conditions including against various congesting levels, TCP and independent DPM traffic.