Figure 1 - uploaded by Zhenyun Zhuang
Content may be subject to copyright.
Source publication
Peer-to-peer (P2P) systems are self-organized and decentralized. However, the mechanism of a peer randomly joining and leaving a P2P network causes topology mismatching between the P2P logical overlay network and the physical underlying network. The topology mismatching problem brings great stress on the Internet infrastructure and seriously limits...
Contexts in source publication
Context 1
... mechanism of a peer joining a P2P network and the fact of a peer randomly joining and leaving causes an interesting matching problem between a P2P overlay net- work topology and the underlying physical network topol- ogy. Figure 1 shows two examples of P2P overlay topology (A, B, and D are three participating peers) and physical topology (nodes A, B, C, and D) mappings, where solid lines denote physical connections and dashes lines denote overlay (logi- cal) connections. Consider the case of a message delivery from peer A to peer B. In the left figure, A and B are both P2P neighbors and physical neighbors. ...
Context 2
... Consider the case of a message delivery from peer A to peer B. In the left figure, A and B are both P2P neighbors and physical neighbors. Thus, only one communication is involved. In the right figure, since A and B are not P2P neighbors, A has to send the message to D before forwarding to B. This will involve 5 communications as indicated in Fig. 1. Clearly, such a mapping creates much unnecessary traffic and lengthens the query response time. We refer to this phenomenon as topology mismatching prob- lem. [3] show that only 2 to 5 percent of Gnutella connections link peers within a single autonomous system (AS). But more than 40 percent of all Gnutella peers are located within ...
Similar publications
The convergence of the Grid and Peer-to-Peer (P2P) worlds has led to many solutions that try to efficiently solve the problem of resource discovery on Grids. Some of these solutions are extensions of P2P DHT-based networks. We believe that these systems are not flexible enough in the case the data present in the network is very dynamic, i.e. the da...
In this paper we present an analytic study of the impact of churn in cluster-based overlay networks. Cluster-based overlays keep the best of unstructured and structured overlays in terms of scalability, fault-tolerance and stability. Most of join and leave events have no impact on the overall overlay topology making these overlays highly robust to...
As the substrate of structured peer-to-peer systems, distributed hash table (DHT) plays a key role in P2P routing infrastructures. Traditional DHT does not consider the location of the nodes for the assignment of identifiers, which will result in high end-to-end latency on DHT-based overlay networks. In this paper, we propose a design of locality-a...
Recently, many kinds of content are being circulated within a great many service-specific overlay networks. When the content is not extremely delay-sensitive, content circulation between wireless terminals can be realized without additional resources by using off-peak periods in wireless access links. In such content circulation, peer-to-peer conte...
This paper introduces a novel unstructured P2P system able to adapt its overlay network topology to the load conditions. The adaptation is performed by means of a mechanism which is run by the nodes in the network in an autonomous manner using only local information, so no global coordinator is needed. The aim of this adaptation is to build an effi...
Citations
... Theoretical results are pessimistic, showing that local routing becomes difficult if no localization is available [Bose et al. 2009;Braverman 2008]. Fig. 4: Illustration of the topology mismatching problem (redrawn from [Liu et al. 2003]). Solid edges and dashed edges give the underlay and overlay, respectively. ...
... Other Algorithms. AOTO [Liu et al. 2003] is another algorithm for reducing topology mismatching in Gnutella. This algorithm continuously performs two consecutive tasks. ...
Limiting the knowledge of individual nodes is a major concern for the design of distributed algorithms. With the LOCAL model, theoretical research already established a common model of locality that has gained little practical relevance. As a result, practical research de facto lacks any common locality model. The only common denominator among practitioners is that a local algorithm is distributed with a restricted scope of interaction. This article closes the gap by introducing four practically motivated classes of locality that successively weaken the strict requirements of the LOCAL model. These classes are applied to categorize and survey 36 local algorithms from 12 different application domains. A detailed comparison shows the practicality of the classification and provides interesting insights. For example, the majority of algorithms limit the scope of interaction to at most two hops, independent of their locality class. Moreover, the application domain of algorithms tends to influence their degree of locality.
... Numerous solutions [4] [5] [6] have been proposed to address the topology mismatch problem considering the properties of the underlying physical network like the location of the peer or delay and RTT between the peers. ...
... The solution in AOTO [6] use the formation of multicast spanning tree approaches which forms an overlay multicast tree from the source peer to all its neighbor nodes. Each peer maintains a cost table containing cost to reach all its neighbors. ...
... In these algorithms each peer also uses a local operator for changing the connections among the peers with the aim of decreasing the number of mismatched paths of the overlay. This type of algorithm is used in both structured [13,14,15] and unstructured peer-to-peer networks [6,7,16]. These algorithms reconfigure a given overlay graph ...
... In the review of the relevant literature, we specifically focus on studies addressing topology mismatch problems in unstructured peer-to-peer networks. Reported algorithms can be classified into three classes of algorithms ( Fig. 4): algorithms in which each peer in the network uses some services such as GPS 1 or Oracle to gather information about the location of peers [23][24][25], algorithms in which each peer in the network uses information about landmark peers [26][27][28][29][30][31][32][33], and algorithms in which each peer in the network uses local information about its neighbors [6,7,9,11,12,16,17,[34][35][36][37][38][39]. In the rest of this section, we briefly review these classes of algorithms. ...
... Deterministic approaches can be further classified into two classes: tree based algorithms and heuristic algorithms. In tree based algorithms such as those reported in [16,38,39], some algorithms for finding minimum spanning trees and multicast trees are used to reconfigure the links of peers. A main drawback of tree based algorithms is that they not only need to tune their parameters with an adaptive mechanism, but they also need another algorithm to create multicast trees or minimum spanning trees. ...
Peer-to-peer networks are overlay networks that are constructed over underlay networks. These networks can be structured or unstructured. In these networks, peers choose their neighbors without considering underlay positions, and therefore, the resultant overlay network may have a large number of mismatched paths. In a mismatched path, a message may meet an underlay position several times, which causes redundant network traffic and end-to-end delay. In some of the topology matching algorithms called the heuristic algorithms, each peer uses a local search operator for gathering information about the neighbors of that peer located in its neighborhood radius. In these algorithms, each peer also uses a local operator for changing the connections among the peers. These matching algorithms suffer from two problems; neither the neighborhood radius nor the local operator can adapt themselves to the dynamicity of the network. In this paper, a topology matching algorithm that uses learning automata to adapt the neighborhood radius and an adaptation mechanism inspired from the Schelling segregation model to manage the execution of the local operator is proposed. To evaluate the proposed algorithm, computer simulations were conducted and then the results were compared with the results obtained for other existing algorithms. Simulation results have shown that the proposed algorithm outperforms the existing algorithms with respect to end-to-end delay and number of mismatched paths.
... This topology does not require the global information of all the p2p networks when each peer in the network attempts to organize the peers in its neighborhood. The simulation shows that the AOTO topology has 55% better performance in reducing traffic compared with the regular tree peer-to-peer topology [35]. ...
A Review on Peer-to-Peer Live Video Streaming Topology
... The solutions AOTO [6] and ACE [7] use the formation of multicast spanning tree approaches which forms an overlay multicast tree from the source peer to all its neighbor nodes. Each peer maintains a cost table containing cost to reach all its neighbors. ...
The formation of unstructured overlays without considering the underlying physical network properties introduces a substantial amount of redundant traffic into the P2P network. Also the frequent changes to the overlay topology and message overhead involved in maintaining the P2P overlays affects the performance of the system. By optimizing the routes traversed by the peers during the communication and implementing a simple but scalable and robust P2P overlay, the proposed system tries to reduce the redundant traffic in the network. The system is assessed by proper simulation and analysis with the findings providing improvements in the total system performance. General Terms Distributed Networks, Network Traffic, Simulation, Overlays
... This topology does not require the global information of all the p2p networks when each peer in the network attempts to organize the peers in its neighborhood. The simulation shows that the AOTO topology has 55% better performance in reducing traffic compared with the regular tree peer-to-peer topology [35]. ...
... In 2000, a company called " Clip2 " developed a Gnutella crawler and published their results on the web. Although not validated by peer-review, their analysis and topology captures have been widely used in simulation studies of improvements for Gnutella-like networks [1, 19,293031. In [8], Clip2 presents analysis of snapshots they captured between June and August of 2000. ...
The combination of large scale and geographically distributed nature of
P2P system has led to their significant impact on the Internet. It is essential to characterize
deployed P2P system for at least three reasons: (1) Accurately assessing
their impact on the Internet, (2) identifying any performance bottleneck as well as
any opportunity for performance improvement, (3) understanding user-driven dynamics
in P2P systems.
To characterize a P2P system, one needs to accurately capture snapshots of the
resulting P2P overlay. This is challenging because the overlay is often large and dynamic.
While the overlay is discovered by a crawler, it is changing which leads to a
distorted view of the system. Capturing unbiased view of the traffic in the overlay is
equally challenging because it is difficult to show that the captured behavior represent
the observed behavior by all peers.
In this chapter, we describe some of the fundamental problems in empirical characterization
of widely deployed P2P systems. We present several examples to illustrate
the effect of ad-hoc measurement/data collection on the resulting analysis/characterization.
We then present two sampling techniques as a powerful approach to
capture unbiased view of peer properties in a scalable fashion.
... We further contrast this design with three previously proposed approaches, AOTO [10], LTM [12], and SBO [14] in Figs. 13 and 14. The convergent speed of AOTO is the slowest, so its performance in dynamic environments is not as good as the other three approaches. ...
Abstract—The efficiency of Peer-to-Peer (P2P) systems,is largely dependent,on the overlay constructions. Due to the random selection of logical neighbors, there often exists serious topology mismatch problem between overlay and physical topologies in P2P systems. Such mismatching causes unnecessary query message duplications at both the overlay and IP level, as well as an increase in query response time. In this work, we define the optimal overlay problem and prove its NP-hardness. We then propose a distributed overlay optimization algorithm,to address,this issue and evaluate its effectiveness,through trace-driven simulations. The proposed design has four strengths. First, it does not need any global knowledge. Second, its optimization convergent speed is fast. Third, it is orthogonal to other types of advanced search approaches. Fourth, it reduces both the traffic cost and the search latency. Index Terms—Peer-to-Peer, topology mismatch, optimal overlay, NP-hard, distributed approach. Ç 1I NTRODUCTION
... Our scheme manages those types of connections so that, eventually , contributors become more close to each other in the network, and free riders become isolated. There exist some other studies which also focus on modifying P2P topology such as16171829,30]. However, these works aim to solve the topology mismatching problem and improve the search quality; they do not attack the free riding problem directly. ...
... However, these works aim to solve the topology mismatching problem and improve the search quality; they do not attack the free riding problem directly. In [16], Liu et al. proposed a solution called the Adaptive Overlay Topology Optimization (AOTO) to optimize inefficient overlay topologies for improving P2P search and routing efficiency. In another work [17], Crameret et al. also aimed to create a topology refinement by modifying the bootstrapping mechanism in the P2P network. ...
The existence of a high degree of free riding in Peer-to-Peer (P2P) networks is an important threat that should be addressed while designing P2P protocols. In this paper we propose a connection-based solution that will help to reduce the free riding effects on a P2P network and discourage free riding. Our solution includes a novel P2P connection type and an adaptive connection management protocol that dynamically establishes and adapts a P2P network topology considering the contributions of peers. The aim of the protocol is to bring contributing peers closer to each other on the adapted topology and to push the free riders away from the contributors. In this way contribution is promoted and free riding is discouraged. Unlike some other proposals against free riding, our solution does not require any permanent identification of peers or a security infrastructure for maintaining a global reputation system. It is shown through simulation experiments that there is a significant improvement in performance for contributing peers in a network that applies our protocol.
... Liu et al. propose AOTO (Adaptive Overlay Topology Optimization) [8] and a location-aware topology matching scheme [6] in which each peer issues a detector message in a small region so that other peers receiving the detector message can record relative delay information from the peer. Moreover, to improve the studies, the authors also discuss ACE (Adaptive Connection Establishment) [7] which builds an overlay multicast tree among a source peer and peers within a certain diameter from the source peer. ...
Peer-to-peer (P2P) systems are widely used in various types of applications. In this paper, we evaluate the superpeer-based two-layer (SBTL) P2P overlay network with the charge-based flooding (CBF) algorithm to detect target peers which have target files, proposed as our previous work. The SBTL P2P overlay network is composed of two layers, normal peer and superpeer layers which include normal peers and superpeers, respectively. Multiple normal peers with some common properties, e.g. peers which have replicas of a file, are interconnected with a superpeer. A collection of a superpeer and normal peers is referred to as a cluster. In a cluster, a normal peer tries to find a target peer which has a target file by itself without help of its superpeer. If no target peer is detected in the cluster, the normal peer asks the superpeer to find the target peer. Then, the superpeer forwards the request message to other superpeers by using a type of flooding algorithm named the CBF algorithm at the superpeer layer. We evaluate the SBTL P2P model in terms of the number of messages exchanged among peers and communication load compared with other models.
