Truthful Application-Layer Multicast in Mesh-based Selfish Overlays

Recently, application-layer multicast (ALM) has emerged as a means to solve most of the problems of IP-based multicast. Topology-aware approach of ALM is more attractive because it exploits underlying network topology data to construct multicast overlay networks. In this paper, a novel overlay network construction mechanism called balanced multi-domain topology-aware grouping (BMTAG) is introduced. BMTAG adopts a special node (domain monitor) to be the parent of all other nodes in the same domain. The mechanism can not only save the time used to discover the topology and execute the path matching algorithm but also lower the depth of the multicast tree. Building a multicast tree with balanced traffic to avoid the bottleneck is another purpose of BMTAG. Simulation results show that BMTAG can get higher performance than TAG especially when the group size is large or the percentage of domain nodes is high.

Video streaming services over the Internet are growing rapidly owing to the broadbandization of various networks. However, in such systems where a large number of users simultaneously access the same contents server, the convergence of the network traffic and service requests presents a significant problem. From this point of view, application layer multicast (ALM) has recently attracted much attention due to its scalability and feasibility in the current network systems. In ALM, user nodes construct a multicast tree structure to efficiently deliver an identical video streaming content. Various techniques to construct ALM trees have been designed to achieve improved performance in terms of shorter delay and higher reliability. However, in most of the existing ALM schemes, the performance in channel switching events has not been studied sufficiently; channel switching can be considered as the combination of the procedures for departing from the current and joining the future ALM trees, respectively. In general, node joining and departure processes take a certain time, which can be increased significantly depending on the network conditions whereby users are left waiting for the start of a new program on the switched channel. Therefore, reducing the channel switching delay is a significant issue in multichannel video streaming in ALM. In this paper, we propose a multichannel streaming scheme to reduce channel switching delay in ALM. Our scheme is based on an advanced ALM method dubbed network-aware hierarchical arrangement graph, which constructs node-disjoint multicast trees by utilizing the arrangement graph theory to achieve high robustness on node departures. The performance of the proposed scheme is verified through extensive computer simulations, which demonstrate that our proposed scheme succeeds in reducing the delay in channel switching without system throughput degradation.

The geographically distributed system can be interconnected via an overlay multicast network. In this overlay network the multicast data are routed and replicated on the application layer along a multicast tree. This paper presents the techniques of the network reduction and the multicast tree construction. The multicast tree in the form of shortest path tree (SPT) can be build up upon the linear programming formulation. To control the load of each host, the additional constraints on the maximal number of directly outgoing connections and integer variables are added and subsequently form the degree-bounded shortest path tree problem (db-SPT). This theoretically based problem is formulated in integer linear programming framework

Despite the availability of high bandwidth Internet access for end-users, video broadcasting over the Internet is not widely spread. Multicast communication decreases the network load by eliminating redundancy of the data transfer. However IP multicast was never widely accepted by commercial Internet service providers (ISP). Existing solutions solving this problem, like MBONE tunneling, are not available for end-users accessing the Internet via xDSL or TV cable. Application layer multicast using peer-to-peer (P2P) overlay networks could solve the problem of sparse IP multicast support in the Internet. A limitation of this approach is the lack of standardized interfaces for existing IP multicast applications. We propose a solution, which bridges application layer multicast and IP multicast and uses a P2P (overlay) network to transport multicast data. Our solution - including a proof-of-concept prototype - enables video broadcasting over the Internet using existing IP multicast applications without requiring additional service deployment.

It is proved that network coding can achieve multicast capacity and therefore improves the throughput of a multicast network significantly. This chapter will focus on applying network coding to Application layer multicast(ALM). The benefits are two folds: first, ALM is built on peer-to-peer networks whose topology can be arbitrary so it is easy to tailor the topology to facilitate network coding; second, the nodes in ALM are end hosts which are powerful enough to perform complex encoding and decoding operations.

In view of the free content distribution in multicast on Internet, a mechanism for constructing the minimum cost multicast routing tree is proposed based on the theory of economic mechanism design. The minimum cost multicast tree can minimize the total transmission cost of the tree, minimize the links connecting the specified multicast group members and save the data transmitting cost and network resources. The model of multicast networks is provided and a routing model based on minimum cost multicast tree is formulate. According to the existing multicast protocols, the design of an algorithm for implementing the routing mechanism is described. The mechanism is evaluated by means of a numerical example. Some prospects on the application of mechanism design theory in multicast research fields are presented.

International audience

Collaborative virtual environments (CVEs) such as massive multiuser 3D games and military training environments can place strict requirements on networks when participating users share the 3D virtual environment through mobile devices in an ad-hoc network. In this paper, the authors show how a CVE application can benefit from the application layer multicast when deployed on the Gnutella peer-to-peer network over an ad-hoc network. The authors propose a protocol called GALM (Gnutella application layer multicast). GALM requires no infrastructure support such as a multicast router to maintain the group state. It has the following characteristics: a) it is adaptable to mobility and network group size by managing the mobile device resources by using gateway node, b) it is reliable; the CVE application can choose at a running time the adequate transport protocol for each data type - for example, using TCP for scene and object data and RTP to send video and audio data, c) it is independent from the lower layer; any link failure or mobility in the physical layer will not affect the application layer, eliminating the need to perform a multicast tree reconfiguration, d) it has link quality; therefore, a cross layer can be used between the network and the application layer in order to provide optimal paths in the multicast tree configuration process. In addition, the protocol handles tolerance to mobility and multicast tree recovery using a smart logical Gnutella network that is based on a novel discovery technique in which mobile nodes are found by means of both their state and position in the CVE.

This paper considers the problem of constructing minimum-delay multicast trees for message dissemination in overlay networks in which nodal delays for processing messages are taken into account in addition to communication delays. The objective is to find a multicast tree to deliver a message from a source to multiple destinations in minimum delay time. This paper introduces a new delay measure called reception-and-processing delay that takes nodal processing delays into account for constructing multicast trees. In the process of constructing a multicast tree, the branches of the multicast tree may intersect or overlap with each other. An efficient method is devised to avoid this undesirable phenomenon. Several heuristic algorithms are developed for constructing multicast trees. The performances of the heuristic algorithms are studied via simulation. Our simulation results show that constructing the multicast tree by iteratively attaching the destination node with the smallest reception-and-processing delay to the current multicast tree yields the lowest average multicast delay among the heuristic algorithms studied in this paper.

Overlay network monitoring is one of the most important issues in the design and operation of overlay networks. It faced with increasing demand for better throughput and response time performance. Given an overlay network with n end hosts, existing systems either require O (n2) measurements, and thus lack scalability, or can only estimate the latency but not failures or congestion. We propose a scalable deterministic network monitoring scheme that selectively monitors k linearly independent paths which can fully describe all the O (n2) paths. We use the loss rates and latency of the collected k path to deduce the loss rate and latency of the rest paths. Our method only assumes knowledge of the underlying IP topology, with links dynamically varying between normal and lossy. In this paper, we propose a novel Path Loss Inference Algorithm (in short, PLIA) which improves implements and extensively estimates such a monitoring method. We carry out our experiments in two different scenarios and the results from the experiments demonstrate that our approach increases accuracy in monitoring congested paths compared to other representative schemas'. Especially in a router-level topology, our approach can deliver almost a 12.5% increase.

Application layer multicast (ALM) is highly expected to replace IP multicasting as the new technological choice for content delivery. Depending on the streaming application, ALM nodes will construct a multicast tree and deliver the stream through this tree. However, if a node resides in the tree leaves, it cannot deliver the stream to its descendant nodes. In this case, quality of service (QoS) will be compromised dramatically. To overcome this problem, topology-aware hierarchical arrangement graph (THAG) was proposed. By employing multiple description coding (MDC), THAG first splits the stream into a number of descriptions, and then uses arrangement graph (AG) to construct node-disjoint multicast trees for each description. However, using a constant AG size in THAG creates difficulty in delivering descriptions appropriately across a heterogeneous network. In this paper, we propose a method, referred to as network-aware hierarchical arrangement graph (NHAG), to change the AG size dynamically to enhance THAG performance, even in heterogeneous networks. Finally, we evaluate the proposed scheme by experiments using the network simulator ns-2. By comparing our proposed method to THAG and SplitStream, we show that our method provides better performance in terms of throughput and QoS. The results indicate that our approach is more reliable than other methods in heterogeneous networks.

Various streaming services have been gaining much attention in the Internet. In order to reduce the investment cost many companies decide to offer the streaming by the use of the application-layer multicasting constructed on the top of an overlay network. In this paper we address the problem of overlay networks design. The main goal is to minimize the cost of access links selected for nodes participating in the application-layer multicasting. The second objective is to minimize the network delay, which is mainly caused in the Internet by inter-ISP (Internet Service Providers) transmissions. Therefore, in the optimization model the underlying network is composed of ISPs. The idea is to provide enough upload capacity within each ISP in order to reduce the number of links between ISPs. The optimization problem is formulated as an Integer Program. We use optimal methods (branch-and-cut algorithm included in CPLEX) as well as our own heuristic that yields results very close to optimum. Moreover, we use simulations to examine the tradeoff of the delay-aware approach in terms of the cost and the number of inter-ISP links.

This paper presents a new heuristic algorithm called DDBMB (dynamic delay bounded multicast algorithm) to construct a minimum cost multicast tree. The heuristic depends on: (A) bounded delay along paths from source node to each destination node; (B) minimum cost of the multicast tree where the algorithm starts with the minimum delay tree and monotonically decreases the cost iteratively to construct a delay bounded minimum cost tree; and (C) dynamic multicast tree status which is maintained by updating the existing multicast tree when nodes in the network request join or leave. The construction of such constrained multicast tree is known as an NP-complete problem. Compared with other existing relative algorithms, the heuristic demonstrates a better average case behavior and can be very close to having a minimum cost depending on the delay bounds imposed.

CoolStreaming is the first protocol which introduces the mesh structure into Application Layer Multicast(ALM) in media streaming application, i.e. each agent may have two or more parents instead of only one parent. However, since the agents in Mesh-based ALM(MALM) are strategy and selfish, the effect of cheating behavior may not be ignored. To solve this problem, we apply the VCG mechanism design into MALM network model, and devise a strategyproof mechanism to avoid the agent cheating. As a result, the goal to maximize the system outcome can be achieved. In addition, we design a distributed algorithm to realize our mechanism. The algorithm can dynamically adapts to form a better multicast mesh, though ALM network parameters and constraints change dynamically in reality. The correctness and performance of this distributed algorithm are verified by the following experimental results.

Service Overlay Network (SON) was proposed to alleviate the difficulties encountered in providing end-to-end Quality of Service (QoS) guarantees. SON is able to provide QoS guarantees by purchasing bandwidth from individual network domains and building a logical end-to-end data delivery infrastructure on top of the existing Internet. We focus on SON topology design problems under a generalized cost model. Earlier research in this topic considered two distinct cost models - fixed (leased) cost model and variable (usage-based) cost model. However in most applications, the costs of both nodes and links have a fixed component as well as a variable component that often depends on usage. Our generalized cost model takes this fact into account and our topology design algorithm uses this cost model to And the optimal topology. Since the SON topology design problem is NP-complete, we provide approximation algorithm with guaranteed performance bound. We validate the effectiveness of our algorithm through extensive simulation.