Peer-to-Peer systems form logical overlay networks on top of the Internet. Essentially, peers randomly
choose logical neighbours without any knowledge about underlying physical topology. This may cause
inefficient communications among peers. This topology mismatch problem may result in poor
performance and scalability for Peer-to-Peer systems. A possible way to improve the performance of
Peer-to-Peer systems is the overlay network construction based on the knowledge of the physical network
topology. In this paper, we will propose the use of the “Record Route” and “Timestamp” options
supported in the IP protocol to explore the paths between peers. By the topology-aware peer selection,
our approach outperforms traditional P2P systems using random peer selection. Our approach only
incurs a low overhead and can be deployed easily in various P2P systems.
The International Journal of Engineering and Science (IJES)theijes
The International Journal of Engineering & Science is aimed at providing a platform for researchers, engineers, scientists, or educators to publish their original research results, to exchange new ideas, to disseminate information in innovative designs, engineering experiences and technological skills. It is also the Journal's objective to promote engineering and technology education. All papers submitted to the Journal will be blind peer-reviewed. Only original articles will be published.
ON THE OPTIMIZATION OF BITTORRENT-LIKE PROTOCOLS FOR INTERACTIVE ON-DEMAND ST...IJCNCJournal
This paper proposes two novel optimized BitTorrent-like protocols for interactive multimedia streaming: the Simple Interactive Streaming Protocol (SISP) and the Exclusive Interactive Streaming Protocol (EISP). The former chiefly seeks a trade-off between playback continuity and data diversity, while the latter is mostly focused on playback continuity. To assure a thorough and up-to-date approach, related work is carefully examined and important open issues, concerning the design of BitTorrent-like algorithms, are analyzed as well. Through simulations, in a variety of near-real file replication scenarios, the novel protocols are evaluated using distinct performance metrics. Among the major findings, the final results show that the two novel proposals are efficient and, besides, focusing on playback continuity ends up being the best design concept to achieve high quality of service. Lastly, avenues for further research are included at the end of this paper as well
Textual based retrieval system with bloom in unstructured Peer-to-Peer networksUvaraj Shan
This document summarizes a research article about a textual retrieval system using Bloom filters in unstructured peer-to-peer networks. It discusses how Bloom Cast replicates document content across the network using Bloom filters to encode documents. This allows for efficient full-text searches with guaranteed recall rates while reducing communication costs compared to replicating raw documents. The system samples nodes randomly using a lightweight distributed hash table to support searches in an unstructured P2P network where the network size is unknown.
Analytical Modelling of Localized P2P Streaming Systems under NAT ConsiderationIJCNCJournal
This document summarizes an analytical model for localized peer-to-peer (P2P) streaming systems that considers the impact of network address translation (NAT). It introduces theoretical boundaries for the number of peers that may be expelled from the system due to NAT incompatibility. It also presents a mathematical model for startup delay in P2P streaming that accounts for peers' NAT types. The document proposes a new neighbor selection algorithm that considers both autonomous system numbers and NAT types to improve connectivity while reducing transit traffic and startup delays.
A Distributed Approach to Solving Overlay Mismatching ProblemZhenyun Zhuang
This document proposes an algorithm called Adaptive Connection Establishment (ACE) to address the topology mismatch problem between the logical overlay network and physical underlying network in unstructured peer-to-peer systems. ACE builds a minimum spanning tree among each source node and its neighbors within a certain diameter, optimizes connections not on the tree to reduce redundant traffic, while retaining search scope. It evaluates tradeoffs between topology optimization and information exchange overhead by changing the diameter. Simulation results show ACE can significantly reduce unnecessary P2P traffic by efficiently matching the overlay and physical network topologies.
Hybrid Periodical Flooding in Unstructured Peer-to-Peer NetworksZhenyun Zhuang
This document proposes a new search mechanism called Hybrid Periodical Flooding (HPF) for unstructured peer-to-peer networks. HPF aims to reduce unnecessary traffic like blind flooding while also addressing the "partial coverage problem" of some statistics-based search mechanisms. It introduces the concept of Periodical Flooding (PF), which controls the number of neighbors a query is forwarded to based on the time-to-live value. This allows the forwarding behavior to change periodically over the query's lifetime. HPF then combines PF with weighted selection of neighbors based on multiple metrics to guide queries towards potentially relevant results while exploring more of the network.
The International Journal of Engineering and Science (IJES)theijes
The International Journal of Engineering & Science is aimed at providing a platform for researchers, engineers, scientists, or educators to publish their original research results, to exchange new ideas, to disseminate information in innovative designs, engineering experiences and technological skills. It is also the Journal's objective to promote engineering and technology education. All papers submitted to the Journal will be blind peer-reviewed. Only original articles will be published.
ON THE OPTIMIZATION OF BITTORRENT-LIKE PROTOCOLS FOR INTERACTIVE ON-DEMAND ST...IJCNCJournal
This paper proposes two novel optimized BitTorrent-like protocols for interactive multimedia streaming: the Simple Interactive Streaming Protocol (SISP) and the Exclusive Interactive Streaming Protocol (EISP). The former chiefly seeks a trade-off between playback continuity and data diversity, while the latter is mostly focused on playback continuity. To assure a thorough and up-to-date approach, related work is carefully examined and important open issues, concerning the design of BitTorrent-like algorithms, are analyzed as well. Through simulations, in a variety of near-real file replication scenarios, the novel protocols are evaluated using distinct performance metrics. Among the major findings, the final results show that the two novel proposals are efficient and, besides, focusing on playback continuity ends up being the best design concept to achieve high quality of service. Lastly, avenues for further research are included at the end of this paper as well
Textual based retrieval system with bloom in unstructured Peer-to-Peer networksUvaraj Shan
This document summarizes a research article about a textual retrieval system using Bloom filters in unstructured peer-to-peer networks. It discusses how Bloom Cast replicates document content across the network using Bloom filters to encode documents. This allows for efficient full-text searches with guaranteed recall rates while reducing communication costs compared to replicating raw documents. The system samples nodes randomly using a lightweight distributed hash table to support searches in an unstructured P2P network where the network size is unknown.
Analytical Modelling of Localized P2P Streaming Systems under NAT ConsiderationIJCNCJournal
This document summarizes an analytical model for localized peer-to-peer (P2P) streaming systems that considers the impact of network address translation (NAT). It introduces theoretical boundaries for the number of peers that may be expelled from the system due to NAT incompatibility. It also presents a mathematical model for startup delay in P2P streaming that accounts for peers' NAT types. The document proposes a new neighbor selection algorithm that considers both autonomous system numbers and NAT types to improve connectivity while reducing transit traffic and startup delays.
A Distributed Approach to Solving Overlay Mismatching ProblemZhenyun Zhuang
This document proposes an algorithm called Adaptive Connection Establishment (ACE) to address the topology mismatch problem between the logical overlay network and physical underlying network in unstructured peer-to-peer systems. ACE builds a minimum spanning tree among each source node and its neighbors within a certain diameter, optimizes connections not on the tree to reduce redundant traffic, while retaining search scope. It evaluates tradeoffs between topology optimization and information exchange overhead by changing the diameter. Simulation results show ACE can significantly reduce unnecessary P2P traffic by efficiently matching the overlay and physical network topologies.
Hybrid Periodical Flooding in Unstructured Peer-to-Peer NetworksZhenyun Zhuang
This document proposes a new search mechanism called Hybrid Periodical Flooding (HPF) for unstructured peer-to-peer networks. HPF aims to reduce unnecessary traffic like blind flooding while also addressing the "partial coverage problem" of some statistics-based search mechanisms. It introduces the concept of Periodical Flooding (PF), which controls the number of neighbors a query is forwarded to based on the time-to-live value. This allows the forwarding behavior to change periodically over the query's lifetime. HPF then combines PF with weighted selection of neighbors based on multiple metrics to guide queries towards potentially relevant results while exploring more of the network.
On client’s interactive behaviour to design peer selection policies for bitto...IJCNCJournal
Peer-to-peer swarming protocols have been proven to be very efficient for content replication over Internet.
This fact has certainly motivated proposals to adapt these protocols to meet the requirements of on-demand
streaming system. The vast majority of these proposals focus on modifying the piece and peer selection
policies, respectively, of the original protocols. Nonetheless, it is true that more attention has often been
given to the piece selection policy rather than to the peer selection policy. Within this context, this article
proposes a simple algorithm to be used as basis for peer selection policies of BitTorrent-like protocols,
considering interactive scenarios. To this end, we analyze the client’s interactive behaviour when accessing
real multimedia systems. This analysis consists of looking into workloads of real content providers and
assessing three important metrics, namely temporal dispersion, spatial dispersion and object position
popularity. These metrics are then used as the main guidelines for writing the algorithm. To the best of our
knowledge, this is the first time that the client’s interactive behaviour is specially considered to derive an
algorithm for peer selection policies. Finally, the conclusion of this article is drawn with key challenges
and possible future work in this research field.
COMPARATIVE STUDY OF CAN, PASTRY, KADEMLIA AND CHORD DHTS ijp2p
Peer-to-Peer (P2P) systems allow decentralization, sharing of all the resources of a network with direct
communication and collaboration between nodes. There are three main families of P2P networks: the
centralized architecture, the decentralized architecture that can be structured or unstructured and the
hybrid architecture. Today, there are several implementations for structured decentralized architectures.
This implies that the insertion and search algorithms are different. Among them we have; Chord, Pastry,
Kademlia, CAN(Content Addressable Network) . The choice of these DHTs (Distributed Hash Table) for an
application is made on the basis of their performances. Studies of each of these DHTs mentioned have been
done, proving their performance. But a comparative study of the four DHTs Chord, Pastry, CAN, Kademlia
has not been clearly addressed by previous works. In this paper, we have conducted a comparative
theoretical study of the DHTs Chord, Pastry, CAN, Kademlia. Then, by simulation, we have evaluated the
performances in terms of latency, number of hops and number of transmitted messages. Our study clearly
shows the differences between mathematically established performance and actual performance in an
environment with less restriction. This analysis was made from the data obtained by using the simple
network layer of the PeerfactSim simulator. This simulator abstracts the different network layers, which
gives the advantage of testing the performances with reasonable accuracy. The use of the single network
layer can be considered an ideal case because the node searches are done locally
AN EFFECTIVE PREVENTION OF ATTACKS USING GI TIME FREQUENCY ALGORITHM UNDER DDOSIJNSA Journal
This document summarizes an algorithm called the GI (Group Intruders) Time Frequency Algorithm that is proposed to identify hackers attempting distributed denial of service (DDoS) attacks on websites. The algorithm works by maintaining a history of all user access to the site that includes their IP address and time/date of each access. It identifies users that access the site repeatedly from the same IP address on a single date by calculating the average time between accesses. If the time frequency of accesses exceeds a predefined threshold, the user is added to an intruders list to deny future access. This aims to improve server performance by preventing hackers from overloading the server with requests.
P2P DOMAIN CLASSIFICATION USING DECISION TREE ijp2p
The increasing interest in Peer-to-Peer systems (such as Gnutella) has inspired many research activities
in this area. Although many demonstrations have been performed that show that the performance of a
Peer-to-Peer system is highly dependent on the underlying network characteristics, much of the
evaluation of Peer-to-Peer proposals has used simplified models that fail to include a detailed model of
the underlying network. This can be largely attributed to the complexity in experimenting with a scalable
Peer-to-Peer system simulator built on top of a scalable network simulator. A major problem of
unstructured P2P systems is their heavy network traffic. In Peer-to-Peer context, a challenging problem
is how to find the appropriate peer to deal with a given query without overly consuming bandwidth?
Different methods proposed routing strategies of queries taking into account the P2P network at hand.
This paper considers an unstructured P2P system based on an organization of peers around Super-Peers
that are connected to Super-Super-Peer according to their semantic domains; in addition to integrating
Decision Trees in P2P architectures to produce Query-Suitable Super-Peers, representing a community
of peers where one among them is able to answer the given query. By analyzing the queries log file, a
predictive model that avoids flooding queries in the P2P network is constructed after predicting the
appropriate Super-Peer, and hence the peer to answer the query. A challenging problem in a schemabased Peer-to-Peer (P2P) system is how to locate peers that are relevant to a given query. In this paper,
architecture, based on (Super-)Peers is proposed, focusing on query routing. The approach to be
implemented, groups together (Super-)Peers that have similar interests for an efficient query routing
method. In such groups, called Super-Super-Peers (SSP), Super-Peers submit queries that are often
processed by members of this group. A SSP is a specific Super-Peer which contains knowledge about: 1.
its Super-Peers and 2. The other SSP. Knowledge is extracted by using data mining techniques (e.g.
Decision Tree algorithms) starting from queries of peers that transit on the network. The advantage of
this distributed knowledge is that, it avoids making semantic mapping between heterogeneous data
sources owned by (Super-)Peers, each time the system decides to route query to other (Super-) Peers.
The set of SSP improves the robustness in queries routing mechanism, and the scalability in P2P
Network. Compared with a baseline approach,the proposal architecture shows the effect of the data
mining with better performance in respect to response time and precision.
International Journal of Engineering and Science Invention (IJESI) is an international journal intended for professionals and researchers in all fields of computer science and electronics. IJESI publishes research articles and reviews within the whole field Engineering Science and Technology, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online.
The document discusses optimizing bloom filters for peer-to-peer (P2P) networks to improve multi-keyword search efficiency. It proposes automatically updating bloom filters by polling peers for current service availability and using expiry dates. This reduces stale data and false positives. Evaluation shows the optimized bloom filter approach reduces search time for intersection and union queries by around 185 times compared to regular bloom filters.
IMPROVING BITTORRENT’S PEER SELECTION FOR MULTIMEDIA CONTENT ON-DEMAND DELIVERYIJCNCJournal
The great efficiency achieved by the BitTorrent protocol for the distribution of large amounts of data inspired its adoption to provide multimedia content on-demand delivery over the Internet. As it is not designed for this purpose, some adjustments have been proposed in order to meet the related QoS requirements like low startup delay and smooth playback continuity. Accordingly, this paper introduces a BitTorrent-like proposal named as Quota-Based Peer Selection (QBPS). This proposal is mainly based on the adaptation of the original peer-selection policy of the BitTorrent protocol. Its validation is achieved by means of simulations and competitive analysis. The final results show that QBPS outperforms other recent proposals of the literature. For instance, it achieves a throughput optimization of up to 48.0% in low-provision capacity scenarios where users are very interactive.
Content Distribution for Peer-To-Peer Overlays on Mobile Adhoc Networks to Fu...Editor IJCATR
Peer-to-peer (P2P) networks existing on a MANET are a natural evolution since both are decentralized and have dynamic
topologies. As MANETs grow in use due to the increasing popularity of wireless mesh and 4G networks, it is expected that P2P
applications will remain as a popular means of obtaining files. Network coding has been shown as an efficient means of sharing large
Files in a P2P network. With network coding, all file blocks have the same relative importance. This paper presents an efficient
content distribution scheme that uses network coding to share large files in a P2P overlay running on a MANET. Peers request file
blocks from multiple server nodes and servers multicast blocks to multiple receivers, providing efficient multipoint-to-multipoint
communication.
Simulation results show that compared to other common download techniques, the proposed scheme performs very well, having lower
download time and energy consumption. Also, more peers participate in uploading the file, resulting in greater fairness.
AOTO: Adaptive overlay topology optimization in unstructured P2P systemsZhenyun Zhuang
IEEE GLOBECOM 2003
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 mismatch-
ing 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
the performance gain from various search or routing tech-
niques. We propose the Adaptive Overlay Topology Optimiza-
tion (AOTO) technique, an algorithm of building an overlay
multicast tree among each source node and its direct logical
neighbors so as to alleviate the mismatching problem by choos-
ing closer nodes as logical neighbors, while providing a larger
query coverage range. AOTO is scalable and completely dis-
tributed in the sense that it does not require global knowledge
of the whole overlay network when each node is optimizing the
organization of its logical neighbors. The simulation shows that
AOTO can effectively solve the mismatching problem and re-
duce more than 55% of the traffic generated by the P2P system itself.
Adaptive Sliding Piece Selection Window for BitTorrent SystemsWaqas Tariq
Peer to peer BitTorrent (P2P BT) systems are used for video-on-Demand (VoD) services. Scalability problem could face this system and would cause media servers not to be able to respond to the users’ requests on time. Current sliding window methods face problems like waiting for the window pieces to be totally downloaded before sliding to the next pieces and determining the window size that affects the video streaming performance. In this paper, a modification is developed for BT systems to select video files based on sliding window method. Developed system proposes using two sliding windows, High and Low, running simultaneously. Each window collects video pieces based on the user available bandwidth, video bit rate and a parameter that determines media player buffered seconds. System performance is measured and evaluated against other piece selection sliding window methods. Results show that our method outperforms the benchmarked sliding window methods
ipoque has analyzed the Internet traffic in five regions of the world between August and September 2007. Comprehensive statistics about user behavior provide a unique overview of the Internet’s current state. While the emphasis of last year’s ipoque sur- vey was on peer-to-peer (P2P) file sharing, this study also includes data about Internet telephony (VoIP), Skype, video streaming, instant messaging (IM), file hosting and encrypted P2P protocols. Three petabytes of anonymous data representing over one million users in Australia, Eastern Europe, Germany, the Middle East and Southern Europe have been analyzed. The results for these different regions vary considerably. P2P produces, on average, between 49 and 83 percent of all Internet traffic with nighttime peaks of over 95 percent. About 20 percent of P2P traffic is already encryp- ted. Skype is by far the most popular Internet telephony service.
The document provides an overview of the BitTorrent protocol. It describes BitTorrent's history and operation, including how it uses trackers and a peer-to-peer network to efficiently distribute file pieces among users. Key terms are defined, such as torrents, peers, seeds, and leechers. The BitTorrent protocol specifications are also outlined, including how it uses bencoding for metadata, trackers for peer discovery, and a tit-for-tat incentive mechanism to encourage sharing.
This paper provides a comparison of the current Internet architecture based on TCP/IP and the proposed future Internet architecture of Named Data Networking (NDN). It discusses key differences in their approaches, components, packet formats, and security implementations. The TCP/IP model uses IP addresses and has a client-server request model, while NDN is information-centric, names content directly, and uses an interest-initiated model. NDN aims to more efficiently distribute popular content, optimize bandwidth usage, and reduce congestion compared to the TCP/IP architecture.
An approximation delay between consecutive requests for congestion control in...IJECEIAES
This research presents a way to avoid network congestion during unicast CoAP-based group communication using increased delays between consecutive requests (DCR) in LoWPAN border routers to limit request send rates. It also provides a way to determine DCR values that are suitable for various network group sizes with differing node counts. The optimal DCR is obtained using the least squares approximation method and the relative minimum. Results from experimentation shows a positive relation, that is, an increase in group size necessitates an increase in DCR value. Experiments in various group sizes show favorable network performance and support the proposed congestion control method using DCR.
This document discusses two approaches to peer-to-peer data mining: local algorithms and the Newscast model of computation. Local algorithms perform computations using only local communications between neighbors. The majority voting problem is presented as an example of an exact local algorithm. An approximate local algorithm for K-means clustering over a P2P network is also described. The Newscast model is then introduced as an alternative approach based on a gossip protocol that continuously rewires network connections, allowing data mining primitives to be computed in a decentralized manner even as the network dynamically changes.
The adaptation of the BitTorrent protocol to multimedia on-demand streaming systems essentially lies on
the modification of its two core algorithms, namely the piece and the peer selection policies, respectively.
Much more attention has though been given to the piece selection policy. Within this context, this article
proposes three novel peer selection policies for the design of BitTorrent-like protocols targeted at that type
of systems: Select Balanced Neighbour Policy (SBNP), Select Regular Neighbour Policy (SRNP), and
Select Optimistic Neighbour Policy (SONP). These proposals are validated through a competitive analysis
based on simulations which encompass a variety of multimedia scenarios, defined in function of important
characterization parameters such as content type, content size, and client´s interactivity profile. Service
time, number of clients served and efficiency retrieving coefficient are the performance metrics assessed in
the analysis. The final results mainly show that the novel proposals constitute scalable solutions that may
be considered for real project designs. Lastly, future work is included in the conclusion of this paper.
Flexible bloom for searching textual contentUvaraj Shan
This document describes BloomCast, a system that uses Bloom filters to encode document content and replicate it across peers in an unstructured peer-to-peer (P2P) network. BloomCast aims to achieve high recall for search queries at low communication cost. It hybridizes a lightweight distributed hash table (DHT) with the P2P overlay to enable random node sampling and network size estimation. The DHT helps meet two constraints: query and document replicas are randomly distributed across the network, and peers know the network size. By uniformly replicating content across the network, BloomCast can guarantee search recall. It utilizes Bloom filters to compress document replicas and reduce replication costs.
Flexible bloom for searching textual contentUvaraj Shan
This document presents the BloomCast scheme for efficient full-text retrieval in unstructured peer-to-peer networks. BloomCast replicates document content across the network in the form of Bloom filters to reduce communication costs while still guaranteeing recall. It hybridizes a lightweight distributed hash table with the unstructured overlay to support random node sampling and network size estimation. Queries are evaluated based on Bloom filter membership verification to support full-text searches with the potential for false positives but no false negatives. The system aims to provide efficient and effective full-text retrieval in unstructured P2P networks.
A Brief Note On Peer And Peer ( P2P ) Applications Have No...Brenda Thomas
The document discusses peer-to-peer (P2P) networks and server-based client/server networks. In a P2P network, all computers have equal privileges to share and access information directly without restrictions. P2P networks are easier to set up but provide less security. In a client/server network, file storage and management is centralized on a server. This provides better security but is more complex to set up and manage. The document explores the advantages and disadvantages of each type of network for different usage contexts.
SECURITY CONSIDERATION IN PEER-TO-PEER NETWORKS WITH A CASE STUDY APPLICATIONIJNSA Journal
Peer-to-Peer (P2P) overlay networks wide adoption has also created vast dangers due to the millions of users who are not conversant with the potential security risks. Lack of centralized control creates great risks to the P2P systems. This is mainly due to the inability to implement proper authentication approaches for threat management. The best possible solutions, however, include encryption, utilization of administration, implementing cryptographic protocols, avoiding personal file sharing, and unauthorized downloads. Recently a new non-DHT based structured P2P system is very suitable for designing secured communication protocols. This approach is based on Linear Diophantine Equation (LDE) [1]. The P2P architectures based on this protocol offer simplified methods to integrate symmetric and asymmetric cryptographies’ solutions into the P2P architecture with no need of utilizing Transport Layer Security (TLS), and its predecessor, Secure Sockets Layer (SSL) protocols.
A study of index poisoning in peer topeerIJCI JOURNAL
P2P file sharing systems are the most popular forms of file sharing to date. Its client-server architecture
attains faster file transfers, however with its peer anonymity and lack of authentication it has become a
gold mine for malicious attacks. One of the leading sources of disruptions in the P2P file sharing systems is
the index poisoning attacks. This attack seeks to corrupt the indexes used to reference files available for
download in P2P systems with false data. In order to protect the users from these attacks it is important to
find solutions to eliminate or mitigate the effects of index poisoning attacks. This paper will analyze index
poisoning attacks, their uses and solutions proposed to defend against them.
On client’s interactive behaviour to design peer selection policies for bitto...IJCNCJournal
Peer-to-peer swarming protocols have been proven to be very efficient for content replication over Internet.
This fact has certainly motivated proposals to adapt these protocols to meet the requirements of on-demand
streaming system. The vast majority of these proposals focus on modifying the piece and peer selection
policies, respectively, of the original protocols. Nonetheless, it is true that more attention has often been
given to the piece selection policy rather than to the peer selection policy. Within this context, this article
proposes a simple algorithm to be used as basis for peer selection policies of BitTorrent-like protocols,
considering interactive scenarios. To this end, we analyze the client’s interactive behaviour when accessing
real multimedia systems. This analysis consists of looking into workloads of real content providers and
assessing three important metrics, namely temporal dispersion, spatial dispersion and object position
popularity. These metrics are then used as the main guidelines for writing the algorithm. To the best of our
knowledge, this is the first time that the client’s interactive behaviour is specially considered to derive an
algorithm for peer selection policies. Finally, the conclusion of this article is drawn with key challenges
and possible future work in this research field.
COMPARATIVE STUDY OF CAN, PASTRY, KADEMLIA AND CHORD DHTS ijp2p
Peer-to-Peer (P2P) systems allow decentralization, sharing of all the resources of a network with direct
communication and collaboration between nodes. There are three main families of P2P networks: the
centralized architecture, the decentralized architecture that can be structured or unstructured and the
hybrid architecture. Today, there are several implementations for structured decentralized architectures.
This implies that the insertion and search algorithms are different. Among them we have; Chord, Pastry,
Kademlia, CAN(Content Addressable Network) . The choice of these DHTs (Distributed Hash Table) for an
application is made on the basis of their performances. Studies of each of these DHTs mentioned have been
done, proving their performance. But a comparative study of the four DHTs Chord, Pastry, CAN, Kademlia
has not been clearly addressed by previous works. In this paper, we have conducted a comparative
theoretical study of the DHTs Chord, Pastry, CAN, Kademlia. Then, by simulation, we have evaluated the
performances in terms of latency, number of hops and number of transmitted messages. Our study clearly
shows the differences between mathematically established performance and actual performance in an
environment with less restriction. This analysis was made from the data obtained by using the simple
network layer of the PeerfactSim simulator. This simulator abstracts the different network layers, which
gives the advantage of testing the performances with reasonable accuracy. The use of the single network
layer can be considered an ideal case because the node searches are done locally
AN EFFECTIVE PREVENTION OF ATTACKS USING GI TIME FREQUENCY ALGORITHM UNDER DDOSIJNSA Journal
This document summarizes an algorithm called the GI (Group Intruders) Time Frequency Algorithm that is proposed to identify hackers attempting distributed denial of service (DDoS) attacks on websites. The algorithm works by maintaining a history of all user access to the site that includes their IP address and time/date of each access. It identifies users that access the site repeatedly from the same IP address on a single date by calculating the average time between accesses. If the time frequency of accesses exceeds a predefined threshold, the user is added to an intruders list to deny future access. This aims to improve server performance by preventing hackers from overloading the server with requests.
P2P DOMAIN CLASSIFICATION USING DECISION TREE ijp2p
The increasing interest in Peer-to-Peer systems (such as Gnutella) has inspired many research activities
in this area. Although many demonstrations have been performed that show that the performance of a
Peer-to-Peer system is highly dependent on the underlying network characteristics, much of the
evaluation of Peer-to-Peer proposals has used simplified models that fail to include a detailed model of
the underlying network. This can be largely attributed to the complexity in experimenting with a scalable
Peer-to-Peer system simulator built on top of a scalable network simulator. A major problem of
unstructured P2P systems is their heavy network traffic. In Peer-to-Peer context, a challenging problem
is how to find the appropriate peer to deal with a given query without overly consuming bandwidth?
Different methods proposed routing strategies of queries taking into account the P2P network at hand.
This paper considers an unstructured P2P system based on an organization of peers around Super-Peers
that are connected to Super-Super-Peer according to their semantic domains; in addition to integrating
Decision Trees in P2P architectures to produce Query-Suitable Super-Peers, representing a community
of peers where one among them is able to answer the given query. By analyzing the queries log file, a
predictive model that avoids flooding queries in the P2P network is constructed after predicting the
appropriate Super-Peer, and hence the peer to answer the query. A challenging problem in a schemabased Peer-to-Peer (P2P) system is how to locate peers that are relevant to a given query. In this paper,
architecture, based on (Super-)Peers is proposed, focusing on query routing. The approach to be
implemented, groups together (Super-)Peers that have similar interests for an efficient query routing
method. In such groups, called Super-Super-Peers (SSP), Super-Peers submit queries that are often
processed by members of this group. A SSP is a specific Super-Peer which contains knowledge about: 1.
its Super-Peers and 2. The other SSP. Knowledge is extracted by using data mining techniques (e.g.
Decision Tree algorithms) starting from queries of peers that transit on the network. The advantage of
this distributed knowledge is that, it avoids making semantic mapping between heterogeneous data
sources owned by (Super-)Peers, each time the system decides to route query to other (Super-) Peers.
The set of SSP improves the robustness in queries routing mechanism, and the scalability in P2P
Network. Compared with a baseline approach,the proposal architecture shows the effect of the data
mining with better performance in respect to response time and precision.
International Journal of Engineering and Science Invention (IJESI) is an international journal intended for professionals and researchers in all fields of computer science and electronics. IJESI publishes research articles and reviews within the whole field Engineering Science and Technology, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online.
The document discusses optimizing bloom filters for peer-to-peer (P2P) networks to improve multi-keyword search efficiency. It proposes automatically updating bloom filters by polling peers for current service availability and using expiry dates. This reduces stale data and false positives. Evaluation shows the optimized bloom filter approach reduces search time for intersection and union queries by around 185 times compared to regular bloom filters.
IMPROVING BITTORRENT’S PEER SELECTION FOR MULTIMEDIA CONTENT ON-DEMAND DELIVERYIJCNCJournal
The great efficiency achieved by the BitTorrent protocol for the distribution of large amounts of data inspired its adoption to provide multimedia content on-demand delivery over the Internet. As it is not designed for this purpose, some adjustments have been proposed in order to meet the related QoS requirements like low startup delay and smooth playback continuity. Accordingly, this paper introduces a BitTorrent-like proposal named as Quota-Based Peer Selection (QBPS). This proposal is mainly based on the adaptation of the original peer-selection policy of the BitTorrent protocol. Its validation is achieved by means of simulations and competitive analysis. The final results show that QBPS outperforms other recent proposals of the literature. For instance, it achieves a throughput optimization of up to 48.0% in low-provision capacity scenarios where users are very interactive.
Content Distribution for Peer-To-Peer Overlays on Mobile Adhoc Networks to Fu...Editor IJCATR
Peer-to-peer (P2P) networks existing on a MANET are a natural evolution since both are decentralized and have dynamic
topologies. As MANETs grow in use due to the increasing popularity of wireless mesh and 4G networks, it is expected that P2P
applications will remain as a popular means of obtaining files. Network coding has been shown as an efficient means of sharing large
Files in a P2P network. With network coding, all file blocks have the same relative importance. This paper presents an efficient
content distribution scheme that uses network coding to share large files in a P2P overlay running on a MANET. Peers request file
blocks from multiple server nodes and servers multicast blocks to multiple receivers, providing efficient multipoint-to-multipoint
communication.
Simulation results show that compared to other common download techniques, the proposed scheme performs very well, having lower
download time and energy consumption. Also, more peers participate in uploading the file, resulting in greater fairness.
AOTO: Adaptive overlay topology optimization in unstructured P2P systemsZhenyun Zhuang
IEEE GLOBECOM 2003
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 mismatch-
ing 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
the performance gain from various search or routing tech-
niques. We propose the Adaptive Overlay Topology Optimiza-
tion (AOTO) technique, an algorithm of building an overlay
multicast tree among each source node and its direct logical
neighbors so as to alleviate the mismatching problem by choos-
ing closer nodes as logical neighbors, while providing a larger
query coverage range. AOTO is scalable and completely dis-
tributed in the sense that it does not require global knowledge
of the whole overlay network when each node is optimizing the
organization of its logical neighbors. The simulation shows that
AOTO can effectively solve the mismatching problem and re-
duce more than 55% of the traffic generated by the P2P system itself.
Adaptive Sliding Piece Selection Window for BitTorrent SystemsWaqas Tariq
Peer to peer BitTorrent (P2P BT) systems are used for video-on-Demand (VoD) services. Scalability problem could face this system and would cause media servers not to be able to respond to the users’ requests on time. Current sliding window methods face problems like waiting for the window pieces to be totally downloaded before sliding to the next pieces and determining the window size that affects the video streaming performance. In this paper, a modification is developed for BT systems to select video files based on sliding window method. Developed system proposes using two sliding windows, High and Low, running simultaneously. Each window collects video pieces based on the user available bandwidth, video bit rate and a parameter that determines media player buffered seconds. System performance is measured and evaluated against other piece selection sliding window methods. Results show that our method outperforms the benchmarked sliding window methods
ipoque has analyzed the Internet traffic in five regions of the world between August and September 2007. Comprehensive statistics about user behavior provide a unique overview of the Internet’s current state. While the emphasis of last year’s ipoque sur- vey was on peer-to-peer (P2P) file sharing, this study also includes data about Internet telephony (VoIP), Skype, video streaming, instant messaging (IM), file hosting and encrypted P2P protocols. Three petabytes of anonymous data representing over one million users in Australia, Eastern Europe, Germany, the Middle East and Southern Europe have been analyzed. The results for these different regions vary considerably. P2P produces, on average, between 49 and 83 percent of all Internet traffic with nighttime peaks of over 95 percent. About 20 percent of P2P traffic is already encryp- ted. Skype is by far the most popular Internet telephony service.
The document provides an overview of the BitTorrent protocol. It describes BitTorrent's history and operation, including how it uses trackers and a peer-to-peer network to efficiently distribute file pieces among users. Key terms are defined, such as torrents, peers, seeds, and leechers. The BitTorrent protocol specifications are also outlined, including how it uses bencoding for metadata, trackers for peer discovery, and a tit-for-tat incentive mechanism to encourage sharing.
This paper provides a comparison of the current Internet architecture based on TCP/IP and the proposed future Internet architecture of Named Data Networking (NDN). It discusses key differences in their approaches, components, packet formats, and security implementations. The TCP/IP model uses IP addresses and has a client-server request model, while NDN is information-centric, names content directly, and uses an interest-initiated model. NDN aims to more efficiently distribute popular content, optimize bandwidth usage, and reduce congestion compared to the TCP/IP architecture.
An approximation delay between consecutive requests for congestion control in...IJECEIAES
This research presents a way to avoid network congestion during unicast CoAP-based group communication using increased delays between consecutive requests (DCR) in LoWPAN border routers to limit request send rates. It also provides a way to determine DCR values that are suitable for various network group sizes with differing node counts. The optimal DCR is obtained using the least squares approximation method and the relative minimum. Results from experimentation shows a positive relation, that is, an increase in group size necessitates an increase in DCR value. Experiments in various group sizes show favorable network performance and support the proposed congestion control method using DCR.
This document discusses two approaches to peer-to-peer data mining: local algorithms and the Newscast model of computation. Local algorithms perform computations using only local communications between neighbors. The majority voting problem is presented as an example of an exact local algorithm. An approximate local algorithm for K-means clustering over a P2P network is also described. The Newscast model is then introduced as an alternative approach based on a gossip protocol that continuously rewires network connections, allowing data mining primitives to be computed in a decentralized manner even as the network dynamically changes.
The adaptation of the BitTorrent protocol to multimedia on-demand streaming systems essentially lies on
the modification of its two core algorithms, namely the piece and the peer selection policies, respectively.
Much more attention has though been given to the piece selection policy. Within this context, this article
proposes three novel peer selection policies for the design of BitTorrent-like protocols targeted at that type
of systems: Select Balanced Neighbour Policy (SBNP), Select Regular Neighbour Policy (SRNP), and
Select Optimistic Neighbour Policy (SONP). These proposals are validated through a competitive analysis
based on simulations which encompass a variety of multimedia scenarios, defined in function of important
characterization parameters such as content type, content size, and client´s interactivity profile. Service
time, number of clients served and efficiency retrieving coefficient are the performance metrics assessed in
the analysis. The final results mainly show that the novel proposals constitute scalable solutions that may
be considered for real project designs. Lastly, future work is included in the conclusion of this paper.
Flexible bloom for searching textual contentUvaraj Shan
This document describes BloomCast, a system that uses Bloom filters to encode document content and replicate it across peers in an unstructured peer-to-peer (P2P) network. BloomCast aims to achieve high recall for search queries at low communication cost. It hybridizes a lightweight distributed hash table (DHT) with the P2P overlay to enable random node sampling and network size estimation. The DHT helps meet two constraints: query and document replicas are randomly distributed across the network, and peers know the network size. By uniformly replicating content across the network, BloomCast can guarantee search recall. It utilizes Bloom filters to compress document replicas and reduce replication costs.
Flexible bloom for searching textual contentUvaraj Shan
This document presents the BloomCast scheme for efficient full-text retrieval in unstructured peer-to-peer networks. BloomCast replicates document content across the network in the form of Bloom filters to reduce communication costs while still guaranteeing recall. It hybridizes a lightweight distributed hash table with the unstructured overlay to support random node sampling and network size estimation. Queries are evaluated based on Bloom filter membership verification to support full-text searches with the potential for false positives but no false negatives. The system aims to provide efficient and effective full-text retrieval in unstructured P2P networks.
A Brief Note On Peer And Peer ( P2P ) Applications Have No...Brenda Thomas
The document discusses peer-to-peer (P2P) networks and server-based client/server networks. In a P2P network, all computers have equal privileges to share and access information directly without restrictions. P2P networks are easier to set up but provide less security. In a client/server network, file storage and management is centralized on a server. This provides better security but is more complex to set up and manage. The document explores the advantages and disadvantages of each type of network for different usage contexts.
SECURITY CONSIDERATION IN PEER-TO-PEER NETWORKS WITH A CASE STUDY APPLICATIONIJNSA Journal
Peer-to-Peer (P2P) overlay networks wide adoption has also created vast dangers due to the millions of users who are not conversant with the potential security risks. Lack of centralized control creates great risks to the P2P systems. This is mainly due to the inability to implement proper authentication approaches for threat management. The best possible solutions, however, include encryption, utilization of administration, implementing cryptographic protocols, avoiding personal file sharing, and unauthorized downloads. Recently a new non-DHT based structured P2P system is very suitable for designing secured communication protocols. This approach is based on Linear Diophantine Equation (LDE) [1]. The P2P architectures based on this protocol offer simplified methods to integrate symmetric and asymmetric cryptographies’ solutions into the P2P architecture with no need of utilizing Transport Layer Security (TLS), and its predecessor, Secure Sockets Layer (SSL) protocols.
A study of index poisoning in peer topeerIJCI JOURNAL
P2P file sharing systems are the most popular forms of file sharing to date. Its client-server architecture
attains faster file transfers, however with its peer anonymity and lack of authentication it has become a
gold mine for malicious attacks. One of the leading sources of disruptions in the P2P file sharing systems is
the index poisoning attacks. This attack seeks to corrupt the indexes used to reference files available for
download in P2P systems with false data. In order to protect the users from these attacks it is important to
find solutions to eliminate or mitigate the effects of index poisoning attacks. This paper will analyze index
poisoning attacks, their uses and solutions proposed to defend against them.
The document discusses peer-to-peer (P2P) networks. P2P allows for direct data exchange between users without centralized servers. While P2P is commonly associated with file sharing, many companies and social networks also use P2P technologies. There are different architectures for P2P networks, including centralized networks with indexing servers, decentralized distributed networks without servers, and hybrid networks that combine aspects of both. P2P software uses techniques like distributed hash tables and multi-source transfers to efficiently locate and download content across the network.
Analysis of threats and security issues evaluation in mobile P2P networks IJECEIAES
Technically, mobile P2P network system architecture can consider as a distributed architecture system (like a community), where the nodes or users can share all or some of their own software and hardware resources such as (applications store, processing time, storage, network bandwidth) with the other nodes (users) through Internet, and these resources can be accessible directly by the nodes in that system without the need of a central coordination node. The main structure of our proposed network architecture is that all the nodes are symmetric in their functions. In this work, the security issues of mobile P2P network system architecture such as (web threats, attacks and encryption) will be discussed deeply and then we propose different approaches and we analysis and evaluation of these mobile P2P network security issues and submit some proposal solutions to resolve the related problems with threats and other different attacks since these threats and attacks will be serious issue as networks are growing up especially with mobility attribute in current P2P networks.
This document discusses a study of BitTorrent characteristics through packet analysis. It provides background on BitTorrent, including how it works, key terminology, and the basic structure of torrent files. The study captured packet traces of BitTorrent files from different scenarios to analyze characteristics like peer connections and bandwidth usage over time.
FUTURE OF PEER-TO-PEER TECHNOLOGY WITH THE RISE OF CLOUD COMPUTINGijp2p
Peer-to-Peer (P2P) networking emerged as a disruptive business model displacing the server based
networks within a point in time.P2P technologies are on the edge of becoming all-purpose in developing
several applications for social networking. In the past seventeen years, research on P2P computing and
systems has received enormous amount of attention in the areas of academia and the industry. P2P rose to
triumphant profit-making systems in the internet. It represents the best incarnation of the end to end
argument, the frequently disputed design philosophies that guided the design of the internet. The doubting
factor then is why is research on P2P computing now fading from the spotlight and suffering a nose dive
fall as dramatic as its rise to its popularity. This paper is going to capture a quick look at past results in
peer-to-peer computing with focus on understanding what led to its rise, what contributed to its commercial
success and what has led to its lack of interest. The insight of this paper introduces cloud computing as a
paradigm to peer-to-peer computing.
FUTURE OF PEER-TO-PEER TECHNOLOGY WITH THE RISE OF CLOUD COMPUTINGijp2p
Peer-to-Peer (P2P) networking emerged as a disruptive business model displacing the server based
networks within a point in time.P2P technologies are on the edge of becoming all-purpose in developing
several applications for social networking. In the past seventeen years, research on P2P computing and
systems has received enormous amount of attention in the areas of academia and the industry. P2P rose to
triumphant profit-making systems in the internet. It represents the best incarnation of the end to end
argument, the frequently disputed design philosophies that guided the design of the internet. The doubting
factor then is why is research on P2P computing now fading from the spotlight and suffering a nose dive
fall as dramatic as its rise to its popularity. This paper is going to capture a quick look at past results in
peer-to-peer computing with focus on understanding what led to its rise, what contributed to its commercial
success and what has led to its lack of interest. The insight of this paper introduces cloud computing as a
paradigm to peer-to-peer computing.
We discuss the operation of bit-torrent, the framework behind its working nature and have listed the proposed modifications to deal with performance issues, research challenges and factors to be considered while implementing bit-torrent in a P2P social network. Bit torrent
has been slowly migrating from its conventional public file
sharing role to a social network. In todays world, recognition of having a more distributed and decentralized social network has grown largely among people. We have presented a paradigm for the design of bit-torrent style sharing feature in P2P social network.
International Journal of Computational Engineering Research(IJCER)ijceronline
International Journal of Computational Engineering Research(IJCER) is an intentional online Journal in English monthly publishing journal. This Journal publish original research work that contributes significantly to further the scientific knowledge in engineering and Technology
Bit Torrent is a peer-to-peer file sharing protocol that allows users to download files from each other in a swarm, reducing bandwidth usage and download times compared to single-source downloads. A tracker coordinates the file transfers between peers in a swarm by tracking which file fragments each client has and assisting in efficient data sharing. There are private trackers that require user accounts and enforce rules like minimum upload ratios, and public trackers that are open for any user but are less organized.
COMPARATIVE STUDY OF CAN, PASTRY, KADEMLIA AND CHORD DHTSijp2p
Peer-to-Peer (P2P) systems allow decentralization, sharing of all the resources of a network with direct
communication and collaboration between nodes. There are three main families of P2P networks: the
centralized architecture, the decentralized architecture that can be structured or unstructured and the
hybrid architecture. Today, there are several implementations for structured decentralized architectures.
This implies that the insertion and search algorithms are different. Among them we have; Chord, Pastry,
Kademlia, CAN(Content Addressable Network) . The choice of these DHTs (Distributed Hash Table) for an
application is made on the basis of their performances. Studies of each of these DHTs mentioned have been
done, proving their performance. But a comparative study of the four DHTs Chord, Pastry, CAN, Kademlia
has not been clearly addressed by previous works. In this paper, we have conducted a comparative
theoretical study of the DHTs Chord, Pastry, CAN, Kademlia. Then, by simulation, we have evaluated the
performances in terms of latency, number of hops and number of transmitted messages. Our study clearly
shows the differences between mathematically established performance and actual performance in an
environment with less restriction. This analysis was made from the data obtained by using the simple
network layer of the PeerfactSim simulator. This simulator abstracts the different network layers, which
gives the advantage of testing the performances with reasonable accuracy. The use of the single network
layer can be considered an ideal case because the node searches are done locally.
COMPARATIVE STUDY OF CAN, PASTRY, KADEMLIA AND CHORD DHTSijp2p
Peer-to-Peer (P2P) systems allow decentralization, sharing of all the resources of a network with direct
communication and collaboration between nodes. There are three main families of P2P networks: the
centralized architecture, the decentralized architecture that can be structured or unstructured and the
hybrid architecture. Today, there are several implementations for structured decentralized architectures.
This implies that the insertion and search algorithms are different. Among them we have; Chord, Pastry,
Kademlia, CAN(Content Addressable Network) . The choice of these DHTs (Distributed Hash Table) for an
application is made on the basis of their performances. Studies of each of these DHTs mentioned have been
done, proving their performance. But a comparative study of the four DHTs Chord, Pastry, CAN, Kademlia
has not been clearly addressed by previous works. In this paper, we have conducted a comparative
theoretical study of the DHTs Chord, Pastry, CAN, Kademlia. Then, by simulation, we have evaluated the
performances in terms of latency, number of hops and number of transmitted messages. Our study clearly
shows the differences between mathematically established performance and actual performance in an
environment with less restriction. This analysis was made from the data obtained by using the simple
network layer of the PeerfactSim simulator. This simulator abstracts the different network layers, which
gives the advantage of testing the performances with reasonable accuracy. The use of the single network
layer can be considered an ideal case because the node searches are done locally.
An efficient hybrid peer to-peersystemfordistributeddatasharingambitlick
The document proposes a hybrid peer-to-peer system that combines the advantages of structured and unstructured networks. It consists of two parts: 1) a structured core network that forms the backbone and provides efficient data lookup; 2) multiple unstructured networks attached to each core node, allowing flexible peer joining/leaving. This two-tier design decouples efficiency and flexibility. Simulation results show the hybrid system balances these properties better than single-approach networks.
Routing performance of structured overlay in Distributed Hash Tables (DHT) fo...journalBEEI
This paper presents a routing performance analysis of structured P2P overlay network. Due to the rapid development and hectic life, sharing data wirelessly is essential. P2P allows participating peers move freely by joining and leaving the network at any convenience time. Therefore, it exists constraint when one measuring the network performance. Moreover, the design of structured overlay networks is fragmented and with various design. P2P networks need to have a reliable routing protocol. In order to analyse the routing performance, this work simulates three structured overlay protocols-Chord, Pastry and Kademlia using OMNeT++ with INET and OverSim module. The result shows that Pastry is the best among others with 100% routing efficiency. However, Kademlia leads with 12.76% and 18.78% better than Chord and Pastry in lookup hop count and lookup success latency respectively. Hence, Pastry and Kamelia architectures will have a better choice for implementing structured overlay P2P network.
SECURITY PROPERTIES IN AN OPEN PEER-TO-PEER NETWORKIJNSA Journal
This paper proposes to address new requirements of confidentiality, integrity and availability properties fitting to peer-to-peer domains of resources. The enforcement of security properties in an open peer-topeer network remains an open problem as the literature have mainly proposed contribution on availability of resources and anonymity of users. That paper proposes a novel architecture that eases the administration of a peer-to-peer network. It considers a network of safe peer-to-peer clients in the sense that it is a commune client software that is shared by all the participants to cope with the sharing of various resources associated with different security requirements. However, our proposal deals with possible malicious peers that attempt to compromise the requested security properties. Despite the safety of an open peer-to-peer network cannot be formally guaranteed, since a end user has privileges on the target host, our solution provides several advanced security enforcement. First, it enables to formally define the requested security properties of the various shared resources. Second, it evaluates the trust and the reputation of the requesting peer by sending challenges that test the fairness of its peer-to-peer security policy. Moreover, it proposes an advanced Mandatory Access Control that enforces the required peer-to-peer security properties through an automatic projection of the requested properties onto SELinux policies. Thus, the SELinux system of the requesting peer is automatically configured with respect to the required peer-to-peer security properties. That solution prevents from a malicious peer that could use ordinary applications such as a video reader to access confidential files such as a video requesting fee paying. Since the malicious peer could try to abuse the system, SELinux challenges and traces are also used to evaluate the fairness of the requester. That paper ends with different research perspectives such as a dedicated MAC system for the peer-to-peer client and honeypots for testing the security of the proposed peer-to-peer infrastructure.
BitTorrent is a peer-to-peer file sharing protocol that allows users to distribute large files over the internet. It works by breaking files into pieces that can be downloaded from multiple peers simultaneously, reducing costs and load on any single source. A BitTorrent client connects to a tracker to find peers sharing pieces of a file, downloads different pieces from many peers in random order, and begins sharing pieces with others once enough of the file is obtained. BitTorrent usage accounts for significant internet traffic and is commonly used to distribute large files like Linux distributions, movies, music and other content.
Similar to EFFECTIVE TOPOLOGY-AWARE PEER SELECTION IN UNSTRUCTURED PEER-TO-PEER SYSTEMS (20)
International Journal of Peer to Peer Networks .docxijp2p
The International Journal of peer-to-peer networking is a quarterly open access peer-reviewed journal that publishes articles that contribute new results in all areas of P2P Networks. The journal provides a platform to disseminate new ideas and new research, advance theories, and propagate best practices in the area of P2P networking. This will include works that relate to peer-to-peer systems, peer-to-peer applications, grid systems, large-scale distributed systems, and overlay networks. The journal offers a forum in which academics, consultants, and practitioners in a variety of fields can exchange ideas to further research and improve practices in all areas of P2P.
Authors are solicited to contribute to the journal by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the areas of P2P networks.
International Journal of peer-to-peer networks (IJP2P)ijp2p
The International Journal of peer-to-peer networking is a quarterly open access peer-reviewed journal that publishes articles that contribute new results in all areas of P2P Networks. The journal provides a platform to disseminate new ideas and new research, advance theories, and propagate best practices in the area of P2P networking.
International Journal of peer-to-peer networks (IJP2P)ijp2p
The International Journal of peer-to-peer networking is a quarterly open access peer-reviewed journal that publishes articles that contribute new results
in all areas of P2P Networks. The journal provides a platform to disseminate new ideas and new research, advance theories, and propagate best
practices in the area of P2P networking. This will include works that relate to peer-to-peer systems, peer-to-peer applications, grid systems,
large-scale distributed systems, and overlay networks. The journal offers a forum in which academics, consultants, and practitioners in a variety
of fields can exchange ideas to further research and improve practices in all areas of P2P.
2nd International Conference on Big Data, IoT and Machine Learning (BIOM 2022)ijp2p
2nd International Conference on Big Data, IoT and Machine Learning (BIOM 2022) will act as a major forum for the presentation of innovative ideas, approaches, developments, and research projects in the areas of Big Data, Internet of Things (IoT) and Machine Learning. It will also serve to facilitate the exchange of information between researchers and industry professionals to discuss the latest issues and advancement in the area of Big Data, IoT and Machine Learning.
7th International Conference on Networks, Communications, Wireless and Mobile...ijp2p
7th International Conference on Networks, Communications, Wireless and Mobile Computing (NCWMC 2022) looks for significant contributions to the Computer Networks, Communications, wireless and mobile computing for wired and wireless networks in theoretical and practical aspects. Original papers are invited on computer Networks, network protocols and wireless networks, Data communication Technologies, network security and mobile computing. The goal of this Conference is to bring together researchers and practitioners from academia and industry to focus on advanced networking concepts and establishing new collaborations in these areas.
4th International Conference on Internet of Things (CIoT 2022)ijp2p
4th International Conference on Internet of Things (CIoT 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of IoT.
11th International conference on Parallel, Distributed Computing and Applicat...ijp2p
11th International conference on Parallel, Distributed Computing and Applications (IPDCA 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Parallel, Distributed Computing. Original papers are invited on Algorithms and Applications, computer Networks, Cyber trust and security, Wireless networks and mobile Computing and Bioinformatics. The aim of the conference is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
3rd International Conference on Machine learning and Cloud Computing (MLCL 2022)ijp2p
3rd International Conference on Machine learning and Cloud Computing (MLCL 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of on Machine Learning & Cloud computing. The aim of the conference is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
4th International Conference on Internet of Things (CIoT 2022) ijp2p
4th International Conference on Internet of Things (CIoT 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of IoT.
4th International Conference on Internet of Things (CIoT 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of IoT.
International Journal of peer-to-peer networks (IJP2P)ijp2p
The International Journal of peer-to-peer networking is a quarterly open access peer-reviewed journal that publishes articles that contribute new results in all areas of P2P Networks. The journal provides a platform to disseminate new ideas and new research, advance theories, and propagate best practices in the area of P2P networking. This will include works that relate to peer-to-peer systems, peer-to-peer applications, grid systems, large-scale distributed systems, and overlay networks. The journal offers a forum in which academics, consultants, and practitioners in a variety of fields can exchange ideas to further research and improve practices in all areas of P2P.
3rd International Conference on Networks, Blockchain and Internet of Things (...ijp2p
3rd International Conference on Networks, Blockchain and Internet of Things (NBIoT 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Networks, Blockchain and Internet of Things. The Conference looks for significant contributions to all major fields of the Networks, Blockchain and Internet of Things in theoretical and practical aspects.
Authors are solicited to contribute to the conference by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the following areas but are not limited to:
3rd International Conference on NLP & Information Retrieval (NLPI 2022)ijp2p
3rd International Conference on NLP & Information Retrieval (NLPI 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Natural Language Computing and Information Retrieval.
Authors are solicited to contribute to the conference by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the following areas, but are not limited to.
CALL FOR PAPERS - 14th International Conference on Wireless & Mobile Networks...ijp2p
14th International Conference on Wireless & Mobile Networks (WiMoNe 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Wireless & Mobile computing Environment. Current information age is witnessing a dramatic use of digital and electronic devices in the workplace and beyond. Wireless, Mobile Networks & its applications had received a significant and sustained research interest in terms of designing and deploying large scale and high performance computational applications in real life. The aim of the conference is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
PUBLISH YOUR PAPER - INTERNATIONAL JOURNAL OF PEER-TO-PEER NETWORKS (IJP2P)ijp2p
Authors are solicited to contribute to the journal by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the areas of P2P networks.
International Journal of peer-to-peer networks (IJP2P)ijp2p
The International Journal of peer-to-peer networking is a quarterly open access peer-reviewed journal that publishes articles that contribute new results in all areas of P2P Networks. The journal provides a platform to disseminate new ideas and new research, advance theories, and propagate best practices in the area of P2P networking. This will include works that relate to peer-to-peer systems, peer-to-peer applications, grid systems, large-scale distributed systems, and overlay networks. The journal offers a forum in which academics, consultants, and practitioners in a variety of fields can exchange ideas to further research and improve practices in all areas of P2P.
3rd International Conference on Blockchain and Internet of Things (BIoT 2022)ijp2p
3rd International Conference on Blockchain and Internet of Things (BIoT 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and
applications of Blockchain and Internet of Things. The Conference looks for significant contributions to all major fields of the Blockchain and Internet of Things in theoretical and
practical aspects. Authors are solicited to contribute to the conference by submitting articles that illustrate research
results, projects, surveying works and industrial experiences that describe significant advances in the areas of Blockchain and Internet of Things.
International Journal of peer-to-peer networks (IJP2P)ijp2p
The International Journal of peer-to-peer networking is a quarterly open access peer-reviewed journal that publishes articles that contribute new results in all areas of P2P Networks. The journal provides a platform to disseminate new ideas and new research, advance theories, and propagate best practices in the area of P2P networking. This will include works that relate to peer-to-peer systems, peer-to-peer applications, grid systems, large-scale distributed systems, and overlay networks. The journal offers a forum in which academics, consultants, and practitioners in a variety of fields can exchange ideas to further research and improve practices in all areas of P2P.
CALL FOR PAPERS - 4th International Conference on Internet of Things (CIoT 2022)ijp2p
4th International Conference on Internet of Things (CIoT 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of IoT.
Authors are solicited to contribute to the conference by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the areas of Internet of Things.
CALL FOR PAPER - 13th International Conference on Ad hoc, Sensor & Ubiquitous...ijp2p
13th International Conference on Ad hoc, Sensor & Ubiquitous Computing (ASUC 2022) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of Ad Hoc & Ubiquitous computing. Current information age is witnessing a dramatic use of digital and electronic devices in the workplace and beyond. Ubiquitous Computing presents a rather arduous requirement of robustness, reliability and availability to the end user. Ad hoc, Sensor & Ubiquitous computing has received a significant and sustained research interest in terms of designing and deploying large scale and high performance computational applications in real life.
Understanding User Behavior with Google Analytics.pdfSEO Article Boost
Unlocking the full potential of Google Analytics is crucial for understanding and optimizing your website’s performance. This guide dives deep into the essential aspects of Google Analytics, from analyzing traffic sources to understanding user demographics and tracking user engagement.
Traffic Sources Analysis:
Discover where your website traffic originates. By examining the Acquisition section, you can identify whether visitors come from organic search, paid campaigns, direct visits, social media, or referral links. This knowledge helps in refining marketing strategies and optimizing resource allocation.
User Demographics Insights:
Gain a comprehensive view of your audience by exploring demographic data in the Audience section. Understand age, gender, and interests to tailor your marketing strategies effectively. Leverage this information to create personalized content and improve user engagement and conversion rates.
Tracking User Engagement:
Learn how to measure user interaction with your site through key metrics like bounce rate, average session duration, and pages per session. Enhance user experience by analyzing engagement metrics and implementing strategies to keep visitors engaged.
Conversion Rate Optimization:
Understand the importance of conversion rates and how to track them using Google Analytics. Set up Goals, analyze conversion funnels, segment your audience, and employ A/B testing to optimize your website for higher conversions. Utilize ecommerce tracking and multi-channel funnels for a detailed view of your sales performance and marketing channel contributions.
Custom Reports and Dashboards:
Create custom reports and dashboards to visualize and interpret data relevant to your business goals. Use advanced filters, segments, and visualization options to gain deeper insights. Incorporate custom dimensions and metrics for tailored data analysis. Integrate external data sources to enrich your analytics and make well-informed decisions.
This guide is designed to help you harness the power of Google Analytics for making data-driven decisions that enhance website performance and achieve your digital marketing objectives. Whether you are looking to improve SEO, refine your social media strategy, or boost conversion rates, understanding and utilizing Google Analytics is essential for your success.
Meet up Milano 14 _ Axpo Italia_ Migration from Mule3 (On-prem) to.pdfFlorence Consulting
Quattordicesimo Meetup di Milano, tenutosi a Milano il 23 Maggio 2024 dalle ore 17:00 alle ore 18:30 in presenza e da remoto.
Abbiamo parlato di come Axpo Italia S.p.A. ha ridotto il technical debt migrando le proprie APIs da Mule 3.9 a Mule 4.4 passando anche da on-premises a CloudHub 1.0.
Instagram has become one of the most popular social media platforms, allowing people to share photos, videos, and stories with their followers. Sometimes, though, you might want to view someone's story without them knowing.
Discover the benefits of outsourcing SEO to Indiadavidjhones387
"Discover the benefits of outsourcing SEO to India! From cost-effective services and expert professionals to round-the-clock work advantages, learn how your business can achieve digital success with Indian SEO solutions.
EFFECTIVE TOPOLOGY-AWARE PEER SELECTION IN UNSTRUCTURED PEER-TO-PEER SYSTEMS
1. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
DOI : 10.5121/ijp2p.2011.2101 1
EFFECTIVE TOPOLOGY-AWARE PEER SELECTION
IN UNSTRUCTURED PEER-TO-PEER SYSTEMS
Yen-Cheng Chen and Chen-Chih Liao
Department of Information Management, National Chi Nan University, Nantou, Taiwan
ycchen@ncnu.edu.tw, catsky@yahoo.com.tw
ABSTRACT
Peer-to-Peer systems form logical overlay networks on top of the Internet. Essentially, peers randomly
choose logical neighbours without any knowledge about underlying physical topology. This may cause
inefficient communications among peers. This topology mismatch problem may result in poor
performance and scalability for Peer-to-Peer systems. A possible way to improve the performance of
Peer-to-Peer systems is the overlay network construction based on the knowledge of the physical network
topology. In this paper, we will propose the use of the “Record Route” and “Timestamp” options
supported in the IP protocol to explore the paths between peers. By the topology-aware peer selection,
our approach outperforms traditional P2P systems using random peer selection. Our approach only
incurs a low overhead and can be deployed easily in various P2P systems.
KEYWORDS
Peer to Peer, Topology Awareness, Peer Selection, IP, Route Record, Timestamp
1. INTRODUCTION
Peer-to-Peer (P2P) applications have become very popular in the Internet. Some traffic
measurements show that P2P traffic is starting to dominate the bandwidth in certain segments of
the Internet [1]. In a typical P2P system, a peer realizes its neighbouring peers at the application
level. Accordingly, peers only know the existence of other peers somewhere in the Internet, but
are not aware of how far the peers are physically. The network logically constructed by peers is
called an overlay network. Each logical link of the overlay network corresponds to a path in the
underlying Internet. Since there exists a topology mismatch between the overlay network and
the Internet, communications through an overlay network may be inefficient and may consume
much overall traffic. Therefore, such a P2P file sharing mechanism will incur a large amount of
unnecessary traffic and may spend much more time in downloading files. This mismatch
problem has been described in [2]. Since peer nodes of an overlay network may be distributed
over the Internet, P2P traffic flows may spread over a number of ASs (Autonomous Systems).
Consequently, P2P applications may consume more overall Internet traffic than client-server
applications. It is shown in [3] and [4] that P2P traffic contributes the largest portion of the
Internet traffic, based on measurements on some popular P2P systems, including KaZaA [5],
Gnutella [6], and DirectConnect [7]. For better understanding of bandwidth use in P2P
communications, a mathematical model for P2P bandwidth is proposed in [8]. Measurements in
[9] also show that the flooding-based routing generates 330 TB/month in a Gnutella network
with only 5000 nodes even if 95% of any two nodes are less than 7 hops distant. This heavy use
of Internet traffic gets worse as the number of peer nodes increases. This leads to poor
scalability in unstructured P2P systems [10]. This also raises a demand for a better use of
Internet bandwidth for P2P communications.
There are three major categories of P2P systems: centralized, decentralized structured and
decentralized unstructured. In the centralized model, centralized index servers are employed to
maintain a directory of shared files stored in peers so that a peer can easily find the location of a
desired content from an index server. But, this model suffers from the risk of a single point of
2. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
2
failure. In decentralized P2P systems, without the need of any centralized index server, a peer
finds desired contents and downloads them directly from other peers. In the decentralized
structured model, such as Chord [11], Pastry [12], Tapestry [13], and CAN [14], the shared data
placement and topology characteristics of the network are strictly controlled based on
distributed hash functions. On the other hand, the decentralized unstructured architecture
doesn’t have any hierarchical structure among peers. In other words, all peers are of the same
role [15] and files are distributed in peers in a random manner, independent of the actual
network topology of peers [16]. Decentralized unstructured P2P systems are most commonly
seen in today’s Internet. The main theme of this study is the peer selection issue in decentralized
unstructured P2P systems.
Among unstructured P2P systems, BitTorrent [17][18] is one of the most popular P2P
applications nowadays [19]. We choose BitTorrent as the target P2P system whose performance
is evaluated and improved by our proposed scheme. In fact, BitTorrent is a P2P communication
protocol, which is supported by a lot of BitTorrent client programs. In this paper, the BitTorrent
protocol is first reviewed and an improved scheme is then proposed to be compliant with the
current BitTorrent protocol. Without any change of the BitTorrent protocol, the proposed
approach is developed based on the option support in IP headers, which are available in all
BitTorrent messages. By enabling the “Record Route” (RR) option of the IP header, we can
infer the length of the path between two peers during their initial BitTorrent conversation. By
enabling the “Timestamp” option of the IP header, we can further infer the latency of message
transmissions between two peers. Based on the path and latency information, the proposed
solution constructs an effective overlay network with a logical topology partially resembling the
physical network topology. It will be shown that the traffic and latency incurred by BitTorrent
communication can be reduced by selecting near peers in terms of path length and transmission
latency.
2. PRELIMINARIES
In this section, we will review the BitTorrent protocol and give a brief overview of the Record
Route and Timestamp options in the IP header.
2.1. Overview of BitTorrent
BitTorrent is a P2P application designed to facilitate fast downloading of popular files. The
basic idea of BitTorrent is to divide a single file into pieces with size at most 256 KB. A
BitTorrent client, who attempts to download a file, will first get a seed-file for the file. The
seed-file describes the location of a Tracker server and the hash values of each piece of the file.
A Tracker server is responsible for maintaining a list of peers who are participating in the
sharing of a certain file. Figure 1 illustrates the entities involved in BitTorrent and the initial
procedure for downloading files. After obtaining the seed-file for a file to download, a client
will get connected to the corresponding Tracker server. The Tracker server will give the client a
list of peer IDs which are participating in the file sharing. The default number of peer IDs
included in a list is 50. The number can be changed if required. If there are fewer peers in the
current torrent of file sharing, a shorter list will be returned. If more than 50 peers are
participating in file sharing, the Tracker will randomly pick up 50 peers to include their ID in
the list. After receiving a list of peer IDs, the client then send Handshake messages to all the
peers to gather further information for effective downloading. A Handshake message consists of
the peer’s ID and some hash values which are transmitted in tracker requests. If a peer receives
a Handshake with the hash values that it is not currently serving or the peer’s ID does not match
the expected peer’s IDs, then the peer must drop the connection. After handshaking is
completed, each of the requested peers will immediately send a bit-field message to the client.
Each bit of the bit-field message, corresponding to a specific piece of the wanted file, indicates
3. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
3
whether a specific piece is owned by the requested peer. Thus, from the received bit-field
messages, the client could determine the most effective way to download all the pieces of a file
from a set of peers. To be allowed to download a file piece, the client will send an “interested”
message to the peers who have the specific piece. After completing the download of a piece, the
client verifies the downloaded piece by comparing its hashed value with the hashed value
declared in the seed-file. The client then sends a HAVE message to other peers to announce that
the client has a copy of the piece.
Figure 1. The initial procedure of BitTorrent downloading
In order to facilitate the effective distribution of file pieces, a Rarest First algorithm [17] is used
to determine which piece should be downloaded first. By keeping the received bit-field
messages and updating them whenever receiving any HAVE message, a client knows which
piece is indicated least frequently in those bit-field messages. The piece will be downloaded
first. In addition, each peer allows at most four clients to download pieces from it at the same
time. If there are more than four clients attempting to download pieces from the same peer
simultaneously, these clients will be scheduled to download pieces in turns with an interval of
30 seconds. For the purpose, a peer allows or disallows downloading requests by sending back
“Unchoking” or “Choking” messages respectively. The reciprocation of Choking and
Unchoking is maintained in four flags for each client (called remote peer).
am_choking: the local peer is choking the remote peer
am_interested: the local peer is interested in the remote peer
peer_choking: the remote peer is choking the local peer
peer_interested: the remote peer is interested in the local peer
Basically, a peer allows four remote peers for downloading at the same time and changes the
allowed peers in periods of 30 seconds. This rule can be overwritten by the Optimistic
Unchoking algorithm [17], which first selects the interested peers with best upload rates. To
ensure fairness, each peer uses a tit-for-tat algorithm to get a consistent download rate. In
addition, there are several criteria that a good choking algorithm should meet. First, the number
of simultaneous uploads should be limited for good TCP performance. Choking and unchoking
quickly, known as “fibrillation”, should be avoided. All the peers should have opportunity to
download pieces. Finally, choked connections could be unchoked if they are anticipated to have
better throughput than the current unchoked ones. According to BitTorrent’s specification,
whenever there a request from a client, the Tracker server randomly picks up 50 peers among
the current peers participating the sharing of a file. Therefore, it is very possible that the selected
neighboring peers are far from the client in the physical network topology. This will increase
both transmission latency and traffic overhead for file sharing.
4. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
4
2.2. IP Header’s Options: Record Route and Timestamp
In this section, a brief introduction of the options in the IP header is given. As defined in RFC
791 [20], the last part of the IP header is a list of optional fields. Several types of IP header
options are defined. Each type of options is identified by the option-type octet, which appears at
the first octet of an option. The option-type octet consists of three fields: copied flag (one bit),
option class (two bits) and option number (five bits). The copied flag indicates if the option
should be copied to each fragment of an IP packet. Option class and number are used to classify
options. Table 1 lists the option class, option number, length, and description of each option. In
this paper, we will adopt the Record Route and Timestamp options for path cost evaluation.
Table 1. Options in the IP header.
Class Number Length Description
0 0 -
End of Option list. This option occupies only 1 octet; it has no
length octet.
0 1 -
No Operation. This option occupies only 1 octet; it has no length
octet.
0 2 11
Security. Used to carry Security, Compartmentation, User
Group (TCC), and Handling Restriction Codes compatible with
DOD requirements.
0 3 var.
Loose Source Routing. Used to route the internet datagram
based on information supplied by the source.
0 9 var.
Strict Source Routing. Used to route the internet datagram based
on information supplied by the source.
0 7 var.
Record Route. Used to trace the route an internet datagram
takes.
0 8 4 Stream ID. Used to carry the stream identifier.
2 4 var. Internet Timestamp.
The Record Route option allows an IP packet to record its route in the IP header. If this option is
present in an IP packet, each router which forwards this packet will append its IP address to the
option. As depicted in Figure 2, the Record Route option begins with an option type of value 7.
Following the option type field is the length field to specify the total length of the option,
including option fields and appending route data. The pointer field indicates the position in the
route data where the next route address should be placed. A recorded route is composed of a
sequence of IP addresses. If the value of the pointer is greater than the length, the recorded route
area becomes full. To accommodate all the IP addresses of the routers in a route, the originating
host of the IP packet should specify the length field with a value large enough to place IP
addresses in the route data area.
00000111 length pointer Route data
Figure 2. The Record Route option
When forwarding a datagram, a router checks to see if the record route option is present. If
present, the router inserts its own IP address into the record route option beginning at the octet
indicated by the pointer and increments the pointer by four. If the route data area is already full,
the datagram is forwarded without inserting the IP address into the record route.
Similar to Record Route, the Timestamp option is used to record the timestamps of the routers
within the route of an IP packet. Whenever receiving an IP datagram with the Timestamp option
in its header, a router will inserts the current timestamp into the routed datagram. As illustrated
5. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
5
in Figure 3, the option-type code of Timestamp is 131. The length field is to indicate the number
of octets of the option (maximum length 40). The pointer field indicate the position where the
next timestamp will be appended. The timestamp area is full when the value of the pointer is
greater than the length. The Overflow (oflw) field (4 bits) indicates the number of IP modules
(i.e. routers) that cannot append timestamps into the option due to lack of space. The Flag (flg)
is used to specify the way of appending timestamps:
0: timestamp only, stored in consecutive 32-bit words.
1: Each timestamp is preceded with the IP address of the registering entry.
2: The IP address fields are pre-specified. An IP module only registers its timestamp if it
matches its own address with the next specified IP address
01000100 length pointer oflw flg
Internet address
Timestamp
...
Figure 3. The Timestamp option
A timestamp is a right-justified, 32-bit one in milliseconds since midnight UT (Universal Time).
If the time is not available in milliseconds or can’t be provided with respect to midnight UT,
then any time may be inserted as a timestamp provided the high order bit of the timestamp field
is set to one to indicate the use of a non-standard value. If the timestamp data area of the
forwarded datagram is full, the datagram is forwarded without inserting the timestamp, but the
overflow count is increased by one.
3. RELATED WORK
There have been a number of approaches to improve the performance of unstructured P2P
overlay networks [2][21][22][23]. Most of these approaches use proposed algorithms or
particular protocols to avoid the waste of unnecessary traffic in long-distance message travelling
over the Internet. In these approaches, physical network distances are estimated and near peers
are selected for downloading files more quickly. Essentially, a peer will pick up the closest peer
from its neighbours if network topology information is available. Indeed, we can manually use
the ping or traceroute tool to estimate the network distance between source and destined nodes.
However, these network tools were designed primarily for network diagnostics and are too
heavy-weighted to be used by large-scale applications. For example, the use of BGP routing
table dumps [24] can provide comprehensive topology information, but incurs the same
problem. Furthermore, such topology information is obtained or learned from the underlying
Internet layer and therefore cannot be directly available to end-user applications.
In [23], the authors proposed many approaches, including off-line and on-line ones, to perform
accurate topology-aware operations for server selection using only static information. Based on
the four metrics, namely geographic distance (DIST), number of hops (HOPS), number of
Autonomous Systems (AS), and DEPTH, each peer could estimate the network distance
between peers by regression analysis, which is also evaluated in [23]. Another approach
proposed in [25] is the use of an extended DNS system providing location information of hosts,
networks, and subnets. The geographic information of hosts and networks are added to the DNS
system using the extension to the protocol as defined in RFC 1876 [26] (also see [26] for a
similar implementation). That is, each network element will be associated with static location
6. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
6
information, such as latitude and longitude. The major idea is that two geographically near peers
are most probably two ones with shorter network distance in the physical network. Many
Internet users get connected to the Internet via dialup, DSL, ADSL or cable modem. DHCP
(Dynamic Host Configuration Protocol) protocol is typically used for dynamically assigning IP
address to these hosts. For use in the extended DNS system, all IP addresses available to a
DHCP server can be associated with a single geographic location. We can query the DNS server
to obtain the geographic information of particular peers. Therefore, one can select the closest
server using regression analysis, as proposed in [23], to calculate the network distance metrics
(i.e. round trip time) approximately between any pair of peers. It was shown that 90% accuracy
for an allowed deviation of 30ms can be achieved to utilize geographic information in the
evaluation of transmission latency based on the best off-line model proposed in [23]. This
approach significantly outperforms the RANDOM approach, where the server is chosen
randomly. The off-line approach can achieve the same performance as the best on-line model,
Nevertheless, this off-line approach has an obvious disadvantage: a geographic information
protocol, i.e., RFC 1876, should be supported in each DNS server. In practice, it is very difficult
to deploy this approach in the Internet.
In order to estimate the network distance accurately for a better construction of an overlay
network, SL algorithm and BinSL algorithm were proposed in [21]. In SL algorithm, a node
picks its k neighbours by picking the k/2 nodes in the system closest to itself (these connections
are called short links) and then picks another k/2 nodes at random (these connections are called
long links). It was shown in [27] that if a greedy algorithm is used to merely pick up closer peer
nodes, the constructed overlay network will become disconnected. Therefore, the SL algorithm
picks up k/2 random links to keep the graph connected. This algorithm assumes that the latency
between any peer pairs is available in an n×n IP-latency matrix. Obviously, SL algorithm is
more efficient than the RAN algorithm, which chooses all nodes at random. The major
drawback of the SL algorithm is the difficulty of maintaining the n×n IP-latency matrix, which
also results in poor scalability [28].
Figure 4. An example of the BinSL algorithm in [22]
Ratnasamy et al. [21] propose the distributed BinSL algorithm, which was based on the
evaluated end-to-end delays using a system of k landmarks spread across the Internet. Each
node measures the round trip times (RTTs) from itself to k well-known landmarks on the
Internet. The ordering of the landmarks by the latencies represents the “bin” that the node
belongs to. That is, nodes in the same bin are those with the same ordering of landmarks. The
major idea of the BinSL algorithm is that each node will pick up k/2 nodes in the bin it belongs
to and then picks up other k/2 nodes at random. In the algorithm, latencies are classified into
levels by ranges of latency values. For example, we might divide possible latency values into 3
levels: level 0 for latencies in the range [0,100] ms, level 1 for latencies between (100,200] ms,
and level 2 for latencies greater than 200ms. The landmark ordering for a node is then
augmented with a level vector. Each level value of the vector corresponds to a landmark in the
ordering. For example, the landmark ordering for the peer shown in Figure 4 is l2l1l3. If three
levels as defined above are used, the corresponding level vector will be “0 1 1”. Thus, the bin of
7. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
7
the peer is denoted by “l2l1l3:011”. We can infer that nodes belonging to the same bin are
usually topologically close to each other. It is possible that some nodes do belong to the same
“bin” but are not topologically close to each other. In [22], it is seen that fewer landmarks will
degrade the performance of the binning scheme and may increase the probability of improper
binning. The performance of the BinSL algorithm is a little worse than SL Algorithm while
constructing the overlay network. Each landmark should sustain the load of being pinged by all
nodes to determine landmark orderings. In fact, the overhead is substantial. For example,
assume there are a million nodes and each node needs 10 Pings sent to a landmark to obtain a
good sample of RTT. If nodes refresh their bin information once per hour, it would impose a
load of 2700 Pings per second on each landmark [21]. Although this drawback could be
possibly overcome by the use of multiple collocated nodes acting as a single logical landmark,
the overall communication cost is still high and the deployment of numerous landmark nodes
complicates the implementation of the BinSL algorithm.
Zeinalipour-Yazti et al. [22] propose the use of domain names to select near peers. The
proposed DDNO (Distributed Domain Name Order) method is a distributed technique to make
unstructured overlay networks topologically aware. Each node participating in a DDNO
topology has a domain name (dn), which is a string that conforms to the syntax rules of RFC
1035 [29]. In order to determine whether two peers, saying peer A and peer B with domain
names dnA and dnB respectively, belong to the same domain, this approach defines two
functions, called Split-hash and dnMatch. The Split-hash function is to split a domain name dn
into k hashes, where k is the number of sub-domain strings in dn. After splitting the domain
name, the procedure will use the dnMatch(dnA, dnB) comparison function to compare the
individual hashes of sub-domains in dnA and dnB. For example, if dnA = “a.aol.com” and dnB =
“b.aol.com”, then dnMatch(dnA, dnB) = true. If dnA = “a.yahoo.de” and dnB = “a.yahoo.com”,
then dnMatch(dnA, dnB) = false. In addition to the use of Split-hash and dnMatch, a node may
send a lookupDN message to discover other sibling nodes under the same sub-domain. The
lookupDN message is flooding over the overlay network. The flooding may traverse a number
of randomly selected neighbours before the node finds its siblings. By DDNO, a node tries to
connect to k/2 nodes among the sibling nodes found by lookupDN and picks up other k/2 nodes
at random. Similar to the SL algorithm and BinSL algorithm, the DDNO algorithm makes well-
connected clusters of nodes that are topologically close to each other without jeopardizing
network connectivity. Moreover, DDNO only uses the DNS system in selecting peers.
Therefore, neither the n×n IP-latency matrix nor additional landmarks are required in DDNO.
This leads to good scalability but incurs a little performance degradation compared to SL and
BinSL algorithms. Another disadvantage of the DDNO algorithm is the additional traffic
introduced by the flooding of lookupDN messages, although the authors declared that the
lookups in a completely decentralized fashion will not actually impose large communication
overhead even in larger topologies.
In [2], the authors propose the location-aware topology matching (LTM) algorithm based on the
use of timestamps. A message called TTL2-detector is used in LTM to record timestamps of
nodes for measuring latencies between nodes. These measured latencies will help the
construction of a better overlay network, which could topologically match the physical
underlying network. It is assumed that the clocks in all nodes are synchronized by NTP
(Network Time Protocol) [30]. When a TTL2-detector message traverses a node, the node will
record its timestamp in the message such that the latencies among these traversed nodes can be
calculated. Thus, the topology of the overlay network is constructed by pruning the connections
with a large cost in terms of latency and bandwidth. Indeed, it is confirmed that LTM achieves
the same performance improvement as Gnutella-like overlay networks. However, it is not
applicable to utilize LTM in some P2P overlay networks, like BT-like applications, which don’t
use flooding query messages. In addition, connection pruning used by LTM may jeopardize
network connectivity and probably result in a shattered topology.
8. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
8
4. THE PROPOSED SCHEME
P2P protocols on the Internet are built over the IP layer. That is, P2P protocol messages are
carried by IP packets. The proposed scheme is developed based on the inherent supports in the
IP protocol to gather topological information of the P2P overlay network. The proposed scheme
enables the Timestamp option of the IP header to record the current time of each router within
the path from a peer to another one. Thus, we can evaluate the latency of data transmission from
those timestamp records. Another approach of the proposed scheme is the use of the Record
Route option of the IP header to record the routers which a packet traverses, which can be used
to evaluate the network distance between peers. The proposed scheme is described in three
aspects.
Selection of IP Options
Timestamp (TS): IP supports the Timestamp option to allow routers to record their current times
in IP packets. For a proper use of timestamps, the clocks in all routers should be synchronized
using time synchronization techniques [30] in an acceptable accuracy. If the timestamp option is
present in the IP header of a packet, any router forwarding the packet will put its IP address and
current timestamp in the IP header. For example, if there are n routers between two peers,
timestamp option of the IP header will be (Router1, Timestamp1, Router2, Timestamp2, …,
Routerr, Timestampr). Note that the length of the option in the IP header should not be more
than 40 bytes. If a packet traverse many routers, it is possible that there is no room to allow all
timestamps to be recorded. For saving spaces, recording only timestamps in the header can be
adopted. In this case, the timestamp option area will be (Timestamp1, Timestamp2, …,
Timestampt), which allows more timestamps included. Each timestamp is 32-bit timestamp
string in milliseconds. Thus, at most 10 timestamps can be contained in the IP header option.
The overflow flag field in the timestamp option is used to record the number of the routers
which fail to add timestamps into the option due to the lack of space.
Record Route (RR): IP supports the Record Route option to allow routers to record their IP
addresses in IP packets. If the Record Route option is present in the IP header of a packet, any
router forwarding the packet will put its IP address in the IP header. The Record Route option
will looks like (Router1, Router2, …, Routerr). The space limitation problem also exists in the
use of the Record Route option.
Selection of P2P Messages to Enable IP Options
Both TS and RR scheme can be realized in all messages of a P2P protocol. In fact, it is not
necessary to implement the schemes in each P2P message. To minimize communication and
processing overheads, the proposed approach enables the IP options only during the very initial
message exchange between peers before file downloading. In the BitTorrent protocol, before
selecting peers for downloading files, a peer has to handshake with other peers in the peer-list
given by a tracker. These two IP options can be enabled only in the handshaking packets
because the handshaking is the initial and mandatory message exchange between peers. Let p
denotes a peer which wants to download a file available in a BitTorrent P2P network N.
Suppose that tracker will provide p with a list of participating peers, L = {p1, p2,…, pn}, where n
is the maximum number of peers that a tracker will provide in the list. The default value of n is
50 by the BitTorrent specification. A peer may request a longer list of peers by specifying the
NumWant parameter in the request message to the tracker. If there are fewer peers participating
in the sharing of a specific file, then a shorter list will be provided. By default, the tracker will
randomly select 50 peers to fill into the peer-list and send the peer-list to peer p. Upon receiving
the peer-list, peer p then sends handshake messages to all the peers in list L to request
connections. Those recipients will send responses to peer p. By the propose scheme, the TS or
RR option is enabled in those recipients’ responses. Therefore, after the handshaking procedure,
peer p can acquire the latency or the number of hops between peer p and each peer in list L. For
9. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
9
Gnutella protocol, the TS or RR option will be enabled in Response messages from peers are.
The rationale of enabling IP options in the Response messages is that it is an initial message
before file downloading and is the only one message replied from other peers.
Selection of Peers
After the initial message exchanges between peers, the peer p will obtain a set of estimated
latency or path length data for all the other peers. Then, peer p can select a certain number of
peers with lower latency or shorter path length for downloading files concurrently. For a
widespread distribution of files and better network connectivity, it is not recommended that only
the near peers be selected for file sharing. The idea of our approach is similar to the SL
algorithm. Suppose the number of peers to be selected is k. Given a list of n peers with n ≥ k,
peer p will select the first k/2 peers with lower latencies or shorter path length and randomly
select k/2 peers from the other n - k/2 peers. Recall that more peers randomly picked in file
sharing will increase connectivity, but may lead to worse downloading performance. On the
other hand, more near peers involved in downloading file pieces may achieve better
performance. However, the connectivity of the overlay network will decrease. As shown in [22],
the selection of k/2 random peers in file sharing is a good trade-off between P2P communication
performance and network connectivity.
Figure 5. Procedures of the proposed schemes
Figure 5 depicts the message communications of the proposed scheme. In summary, the
proposed scheme is performed in four steps.
1. A client peer gets the torrent file for a specific file on the web..
2. The client peer sends a request to a tracker to ask for a peer list for downloading the file of
interest. By setting a bigger value of the “NumWant” parameter, the client gets a longer peer
list than the original BT system does.
3. The client peer sends a Handshake message to each peer specified in the given peer list.
Each peer responds to the handshaking by sending a response in an IP packet with Record
Route or Timestamp option enabled.
4. The client picks up the first k/2 peers with the shorter latencies or the smaller number of hops
measured by the TS or RR scheme. The client peer further picks up another k/2 peers
randomly from the remaining n - k/2 peers. Those k selected peers form a new peer list. The
client peer starts to download file pieces from the peers in the new peer list.
One may argue that the latency evaluated via only one handshaking message is not accurate as
the actual time spent in file downloading. The evaluated latency is still a valuable assessment
for peer selection. In fact, the construction of an optimal overlay network is known to be NP-
complete [30]. The proposed scheme can be regarded as a heuristic approach for better
10. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
10
performance than current BitTorrent-based P2P systems. In addition, if we keep measuring the
latency in other messages, e.g. keep-alive and HAVE messages, considerable traffic will be
introduced due to the extra 40 bytes needed in each IP packet. The traffic overhead is also
proportional to the number of participating peers.
In practice, it is difficult to get the latency metric via the Timestamp option of the IP header.
The prerequisite for proper use of timestamps is that the clocks of all the routers should be
synchronized. The requirement is hard to be satisfied in the public Internet, which is constructed
by a lot of networks under different administrative domains. Thus, we prefer the use of the
Record Route option of the IP header for the selection of peers. By counting the entries in the
Record Route data area, we can determine the number of hops for each route. A route with a
smaller hop number is considered as a better one for file sharing.
5. SIMULATION
5.1. Performance Metrics
A study in [19] presents the importance of scalability and efficiency in P2P file sharing systems.
These issues receive more attention recently because of the increased Internet bandwidth
consumption by P2P systems. In fact, BitTorrent provides a built-in incentive unchoking
algorithm to encourage users to share files more efficiently. Each peer prefers uploading to the
peers with a higher downloading rate, since it usually implies more available bandwidth and
better downloading performance. However, it doesn’t imply a better usage of the overall
Internet traffic. To evaluate the effectiveness of the Internet traffic usage, we introduce a new
performance metric Overall Traffic based on the aggregated link costs of downloading file
pieces. The overall traffic consumption of P2P communications becomes one of the important
parameters concerned by network administrators. Network administrators usually use
bandwidth shaping tools to limit the overall bandwidth usage of P2P systems. This still allows
proper operations of P2P services but limits their scalability [10]. Therefore, a better traffic
usage is crucial for the development of contemporary P2P systems. The overall traffic metric for
use in our experiments is defined as follows.
0 0
( [ ][ ]) 256KB (1)
[ ][ ]
p k
i j
Overall Traffic PieceLink i j
PieceLink i j
= =
= ×
≠ ∞
∑∑
where PieceLink is a p*k traffic matrix, assuming that there are p file pieces transferred within a
network with k links. PieceLink[i][j] indicates the consumed traffic in link j for the transmission
of file piece i. In other words, the Overall Traffic metric represents the total traffic volumes for
downloading a file with p pieces within a network with k links. We recommend that Overall
Traffic metric be minimized for better performance of BitTorrent-based P2P systems.
In practice, there are a number of downloading events occurring simultaneously, A piece will
become available in a peer after the peer finishes downloading of the piece. Therefore, a
downloader can soon become an uploader. As the downloading events proceed, a peer can
choose more nearer peers to download file pieces. Therefore, the time required to download a
file will be related to the number of downloading events. Therefore, we use the average
downloading time of all the downloading events to evaluate the downloading performance of
the P2P system. In a downloading event, a user may give up to download a file if the
downloading has taken a long time. Therefore, only successful downloading events are involved
in the following performance evaluation. Assumes the number of successful downloading
events is e, and the downloading time of the i-th event is dti, 0≤i<e. The average downloading
time DTavg is defined as follows.
DTavg = (dt0+dt1+…+dte-1)/e (2)
11. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
11
In addition, we take into account the average number of links required to download a file piece.
It is an important metric in the evaluation of the proposed Record Route scheme. The average
number of links, called Linkavg, is defined as follows.
n
jiLink
Link
n
i
m
j
avg
∑∑
−
=
−
=
=
1
0
1
0
]][[
(3)
where Link is an n m× matrix, assuming that there are n pieces transferred over a network with
m links. Links[i][j] = 1 if link j is within the route of transmitting the file piece i.
Table 2. Configuration of the simulation.
Parameter Value
delay / bandwidth
transit-transit:
transit-stub:
within stub:
5ms / 1000Mbps
10ms / 100Mbps
30ms / 10Mbps
number of peer 300
number of agents per peer 10
DFN-Like topology:
number of nodes
number of edges
420
2100
number of nodes (overlay network) 300
Number of peers in a peer list 100 (RR, TS), 50 (BT)
Number of peers selected 50
5.2 Simulation Configuration
In the simulation, we use the Georgia Tech Internetwork Topology Model (GT-ITM) [32] to
generate network topologies. There are two topology models generated by GT-ITM: Waxman
model [33] and Transit-Stub (TS) model [34]. The Waxman model describes the Internet as a
graph in which the nodes in the network are placed at random points in a two-dimensional grid
and the links are added to the network by connecting any pair of nodes. This model has a
probability function used to decide whether a link should exist. This probability function
involves that how far apart the two nodes are and how many links are expected to exist in the
whole network [35]. The Transit-Stub (TS) model generates a network topology from the top
level of a hierarchy, i.e. Transit domain or WAN, and proceeds downward to the lowest level of
the hierarchy, i.e. Stub domain or LAN. Each node in the generated topology is labelled with a
string of identifiers: an identifier indicating whether it is a transit of Stub-node, a global
identifier of the domain to which it belongs, and a domain-local identifier. Each edge of the
generated topology is associated with a routing policy weight, which can be used to find routes
that follow the standard domain-based routing outlined earlier. In the simulation, we adopt the
TS model as our topology model because the Internet is constructed hierarchically. In addition,
the topology is generated based on the parameters specified in [36] to simulate a real ISP
network, DFN G-Win (German research network). As indicated in [36], the topology generated
by the parameters achieves a high degree of similarity: 0.966. In addition to the topology
generator, we conduct experiments using the General Peer-to-Peer Simulator (GPS), which is a
message-level and event driven P2P simulation framework to model P2P protocols, including
the BitTorrent protocol. In the simulation on GPS, we assume that there are some peers
randomly located in the nodes of the generated topology. A node is also randomly selected as
the tracker. Initially, a document of size 500,000 KB is stored in some peer. The performance of
12. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
12
the P2P system is evaluated regarding different number of downloading events for the
document. Because of the use of an unchoking algorithm in the simulation, the simulator needs
a way to calculate the available bandwidth, as defined follows.
(4)
MSS C
BW
RTT p
=
where BW is the available bandwidth, MSS is the maximum segment size of TCP, RTT is the
round trip time, C is a constant for TCP flavor, and p is the estimated packet loss rate. In our
simulation, we set MSS=536, RTT=1.7, C=1.22, and p=0.001.Three schemes are implemented in
the simulation: Route Record (RR), Timestamp (TS), and the original BitTorrent scheme (BT).
Table 2 lists the settings of the simulation.
5.3 Simulation Results
In the following, the simulation results of average downloading time, overall traffic, and
average number of links of three schemes are presented, with respect to four different numbers
of downloading events: 50, 85, 113, and 149. These four numbers are obtained by counting only
the successful events of all the downloading events in the simulation.
5.3.1. Downloading Time
Figure 6 shows the average downloading times of three schemes in four cases of downloading
events. It can be seen that the proposed TS and RR schemes achieve downloading times slightly
smaller than the BT scheme in the simulation of 50 downloading events. In fact, in the case of
50 downloading events, all the three schemes choose the same peers to download file pieces.
The better performance achieved is due to the different orders of peers involved in file sharing.
When the TS or RR scheme is used, the client will first choose closer peers to download file
pieces, although in all the three schemes downloading connections will be rotated among all
participating peers if the number of downloading events is below 50. Therefore, the proposed
schemes can achieve smaller downloading times slightly even if the number of all participating
peers is below 50. As the number of downloading events increases, more peers become
uploaders and the TS and RR schemes take effect in the selection of peers. By TS and RR
schemes, more nearer peers will be selected in file sharing than the BT scheme. Therefore,
compared to the BT scheme, both TS and RR schemes can achieve smaller downloading times.
In summary, the simulation results show that the average downloading times of the proposed TS
and RR schemes are smaller than the BT scheme in all cases of downloading events. The
improvements of downloading times over the BT scheme are about 3~21% and 2~21%
respectively.
Figure 6. Average downloading time
13. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
13
5.3.2. Overall Traffic
The overall traffic evaluation of our simulations is to investigate the total network traffic
consumed by the proposed schemes. The simulation results of overall traffic are shown in
Figure 7. The improvement of overall traffic achieved by our schemes is not significant in the
case of 50 downloading events. This is because all peers are selected in file sharing. When the
number of downloading events increases, the overall traffic consumed can be reduced by the
proposed schemes. It is shown that our schemes reduce 7%~18% of the overall traffic compared
to the BT scheme in the cases of 85, 113, and 149 downloading events. Furthermore, comparing
the two proposed schemes, we find that the RR scheme can reduce more overall traffic than the
TS scheme with respect to the same number of downloading events. As expected, the RR
scheme provides an accurate estimation of number of links used in file sharing. Therefore, the
client can exactly select the peers with less overall traffic usage.
Figure 7. Overall traffic
5.3.3. Average Number of Links
The last simulation result is the average number of links required to transmit a file piece. The
simulation results are shown in Figure 8. In average, the number of links by the proposed
schemes is about 7% ~ 18% shorter than the BT scheme for more than 50 downloading events.
In addition, the RR scheme slightly outperforms the TS scheme in this simulation. In fact, the
evaluation of average number of links is very similar to the overall traffic evaluation. This
measuring metric is to reveal the degree of topology mismatch between the P2P overlay
network and physical network. A smaller number of links between peers implies a better
topology consistency. Thus, this explains why the performance of P2P systems can be improved
by the proposed schemes.
Figure 8. Average number of links
14. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
14
6. CONCLUSIONS
This paper has presented a simple and scalable approach to inferring network proximity
information. We apply two schemes to the construction of topologically-aware overlay
networks to achieve better usage of overall traffic. Simulations have shown that the performance
improvement can be achieved by our schemes. The performance metrics of our research are
based on three parameters, overall traffic, average downloading time, and average number of
links. Our simulations indicated that our schemes are attractive implementations for
topologically-aware overlay construction as they reduce about 7% ~ 18% of overall traffic and
downloading time. This paper takes the BitTorrent protocol as an unstructured P2P example in
our schemes and simulations. Conceptually, our approaches can be applied in a variety of P2P
systems. In practice, there is a need for a more generalized scheme applicable for different P2P
frameworks, including centralized, decentralized and semi-centralized P2P systems. In the
future, we will further investigate the developement of the proposed schemes in other P2P
systems, such as Gnutella, Foxy, and eMule.
REFERENCES
[1] Top applications (bytes) for subinterface: SD-NAP traffic, 2006.
http://www.caida.org./analysis/workload/byapplication/sdnap.
[2] Yumhao Liu, Xiaomei Liu, Li Xiao, Lionel M. Ni, Xizodong Zhang, “Location-Aware Topology
Matching in P2P Systems,” in Proceedings of IEEE INFOCOM 2004, 2004.
[3] S. Sen and J. Wang, “Availability and locality measurements of peer-to-peer file systems,” in
Proceedings of ACM SIGCOMM Internet Measurement Workshop, 2002.
[4] S. Saroiu, K. P. Gummadi, R. J. Dunn, S. D. Gribble, and H. M. Levy, “An Analysis of Internet
Content Delivery Systems,” in Proceedings of the 5th
Symposium on Operating Systems Design and
Implementation, 2002.
[5] KaZaA, http://www.kazaa.com/.
[6] Gnutella, http://gnutella.wego.com/.
[7] DirectConnect, http://www.neo-modus.com/.
[8] Kiarash Mizanian, Mehdi Vasef and Morteza Analoui, “Bandwidth Modeling and Estimation in
Peer to Peer Networks,” International Journal of Computer Networks & Communications (IJCNC),
Volume 2, Number 3, May 2010, pp. 65-83.
[9] M. Ripeanu, A. Iamnitchi, and I. Foster, “Mapping the Gnutella Network,” IEEE Internet
Computing, Vol. 6, No.1, 2002, pp.50-64.
[10] Ritter, “Why Gnutella can’t scale. No, really”, http://www.darkridge.com/~jpr5/doc/gnutella.html.
[11] R. M. Stoica, D. Karger, M. F. Kaashoek, and H. Balakrishnan, “Chord: A scalable peer-to-peer
lookup service for Internet applications,” in Proceedings of ACM SIGCOMM’01, 2001.
[12] A. Rowstron and P. Druschel, “Pastry: Scalable, distributed object location and routing for large-
scale peer-to-peer systems,” in Proceedings of International Conference on Distributed Systems
Platforms, 2001.
[13] B. Y. Zhao, J. D. Kubiatowicz, and A. D. Joseph, “Tapestry: An infrastructure for fault-resilient
wide-area location and routing,” Technical Report UCB//CSD-01-1141, U.C.Berkely, 2001.
[14] R. Ratnasamy, P. Francis, M. Handley, R. Karp, and S. Shenker, “A scalable content-addressable
network,” in Proceedings of ACM SIGCOMM’01, 2001.
[15] L. Mekouar, Y. Iraqi, R. Boutaba, “Peer-to-peer’s most wanted: Malicious peers”, The
International Journal of Computer and Telecommunications Networking, Vol. 50 , No. 4 , March
2006.
15. International Journal of Peer to Peer Networks (IJP2P) Vol.2, No.1, January 2011
15
[16] B. Yang and H Gracia-Molina, “Efficient search in peer-to-peer networks,” in Proceedings of
International Conference on Distributed Computing Systems (ICDCS’02), 2002.
[17] BitTorrent official website, http://www.BitTorrent.com/, 2005.
[18] B. Cohen. Incentives to Build Robustness in BitTorrent, May 2003.
http://bitconjurer.org/BitTorrent/BitTorrentecon.pdf.
[19] D. Qiu and R. Srikant, “Modeling and Performance Analysis of BitTorrent-Like Peer-to-Peer
Networks,” in Proceedings of ACM SIGCOMM’04, 2004.
[20] J. Postel, “Internet Protocol,” RFC 791, Sep. 1981.
[21] S. Ratnasamy, M. Handley, R. Karp, S. Shenker, “Topologically-aware Overlay Construction and
Server Selection,” INFOCOM 2002. in Proceedings of the Twenty-First Annual Joint Conference of
the IEEE Computer and Communications Societies, Vol. 3, 2002.
[22] D. Zeinalipour-Yazti, V. Kalogeraki, “Structuring topologically aware overlay networks using
domain names,” The International Journal of Computer and Telecommunications Networking
archive, Vol. 50 , No. 16, Nov. 2006.
[23] P. Sinha, D. Raz, N. Choudhuri, “Estimation of network distances using off-line measurements,”
Computer Communications, Vol. 29, 2006, pp. 3295-3305.
[24] B. Krishnamurthy and J. Wang, “On network-aware clustering of web clients,” in Proceedings of
SIGCOMM ’00, Stockholm, Sweden, Aug. 2000.
[25] C. Davis, P. Vixie, T. Goodwin, I. Dickinson, “A Means for Expressing Location Information in the
Domain Name System,” RFC 1876, Jan. 1996.
[26] C. Davis, H. Rose, “DNS LOC: geo-enabling the domain name system,”
http://www.ckdhr.com/dns-loc/.
[27] V. Batagelj, A.Mrvar. “PAJEK-Program for large network analysis,” Connections 21, 1998.
[28] Y. Chawathe, S. Ratnasamy, L. Breslau, N. Lanham, S. Shenker, “Making Gnutella-like P2P
system scalable,” in Proceedings of the 2003 Conference on Applications, Technologies,
Architectures, and Protocols for Computer Communications, Karlsruhe, Germany, 2003, pp.407-
418
[29] P. Mockapetris. “Domain names – Implementation and specification,” RFC 1035, Network Working
Group, November 1987.
[30] David L. Mills, “Network Time Protocol (Version 3) Specification, Implementation and Analysis,”
RFC 1305, March 1992.
[31] M. R. Garey, D. S. Johnson, Computers and Intractability. “Guide to the Theory of NP-
Completeness,” W. H. Freeman San Francisco, CA, 1979.
[32] GT-ITM: Georgia Tech Internetwork Topology Models, http://www.cc.gatech.edu/projects/gtitm/.
[33] B. M. Waxman, “Routing of Multipoint Connections,” IEEE JSAC, col. No. 9, 1988, pp. 1617-22.
[34] D. Knuth, “The Stanford GraphBase: A Platform for Combinatorial Computing,” Reading, MA:
Addison-Wesley, 1994.
[35] Ken Calvert, Matt Doar and Ellen W. Zegura. ,”Modeling Internet Topology,” IEEE
Communications Magazine, June 1997.
[36] O. Heckmann, M. Piringer, J. Schmitt, and R. Steinmetz, “Generating Realistic ISP-Level Network
Topology”, IEEE Communications Letters, Vol. 7, No. 7, July 2003.