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.
Java Abs Peer To Peer Design & Implementation Of A Tuple Spacencct
Final Year Projects, IEEE Projects, Final Year Projects in Chennai, Final Year IEEE Projects, final year projects, college projects, student projects, java projects, asp.net projects, software projects, software ieee projects, ieee 2009 projects, 2009 ieee projects, embedded projects, final year software projects, final year embedded projects, ieee embedded projects, matlab projects, microcontroller projects, vlsi projects, dsp projects, free projects, project review, project report, project presentation, free source code, free project report, Final Year Projects, IEEE Projects, Final Year Projects in Chennai, Final Year IEEE Projects, final year projects, college projects, student projects, java projects, asp.net projects, software projects, software ieee projects, ieee 2009 projects, 2009 ieee projects, embedded projects, final year software projects, final year embedded projects, ieee embedded projects, matlab projects, final year java projects, final year asp.net projects, final year vb.net projects, vb.net projects, c# projects, final year c# projects, electrical projects, power electronics projects, motors and drives projects, robotics projects, ieee electrical projects, ieee power electronics projects, ieee robotics projects, power system projects, power system ieee projects, engineering projects, ieee engineering projects, engineering students projects, be projects, mca projects, mtech projects, btech projects, me projects, mtech projects, college projects, polytechnic projects, real time projects, ieee projects, non ieee projects, project presentation, project ppt, project pdf, project source code, project review, final year project, final year projects
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.
Improving the Proactive Routing Protocol using Depth First Iterative Deepenin...Yayah Zakaria
Owing to the wireless and mobility nature, nodes in a mobile ad hoc network are not within the transmission range. It needs to transfer data through the multi-intermediate nodes. Opportunistic data forwarding is an assuring solution to make use of the broadcast environment of wireless communication links. Due to absence of source routing capability with efficient proactive routing protocol, it is not widely used. To rectify the
problem, we proposed memory and routing efficient proactive routing protocol using Depth-First Iterative-Deepening and hello messaging scheme. This protocol can conserve the topology information in every node in the network. In experimental analysis and discussion, we implemented the proposed work using NS2 simulator tool and proved that the proposed technique is performed well in terms of average delay, buffer and throughput.
Analyse the performance of mobile peer to Peer network using ant colony optim...IJCI JOURNAL
The document describes analyzing the performance of a mobile peer-to-peer network using ant colony optimization. It proposes using a distributed spanning tree (DST) structure to improve efficiency by reducing the large number of messages. The DST is optimized using ant colony optimization to give an optimal solution. Simulation results show the approach reduces the number of messages, average delay, and increases packet delivery ratio in the 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
AN EFFICIENT ROUTING PROTOCOL FOR DELAY TOLERANT NETWORKS (DTNs)cscpconf
Delay-Tolerant Networks are those which lacks continuous communications among mobile
nodes . Distributed clustering scheme and cluster-based routing protocol are used for DelayTolerant
Mobile Networks (DTMNs). The basic idea is to distributive group mobile nodes with
similar mobility pattern into a cluster, which can then interchangeably share their resources for
overhead reduction and load balancing, aiming to achieve efficient and scalable routing in DTMN. Load balancing is carried out in two ways, Intra cluster load balancing and Inter cluster load balancing. The Convergence and stability become major challenges in distributed clustering in DTMN. An efficient routing protocol will be provided for the delay tolerant networks through which the stability of the network is maintained .Based on nodal contact probabilities, a set of functions including Sync(), Leave(), and Join() are devised for cluster formation and gateway selection. Finally, the gateway nodes exchange network information and perform routing
ANALYSE THE PERFORMANCE OF MOBILE PEER TO PEER NETWORK USING ANT COLONY OPTIM...ijcsity
A mobile peer-to-peer computer network is the one in which each computer in the network can act as a
client or server for the other computers in the network. The communication process among the nodes in the
mobile peer to peer network requires more no of messages. Due to this large number of messages passing,
propose an interconnection structure called distributed Spanning Tree (DST) and it improves the efficiency
of the mobile peer to peer network. The proposed method improves the data availability and consistency
across the entire network and also reduces the data latency and the required number of message passes for
any specific application in the network. Further to enhance the effectiveness of the proposed system, the
DST network is optimized with the Ant Colony Optimization method. It gives the optimal solution of the
DST method and increased availability, enhanced consistency and scalability of the network. The
simulation results shows that reduces the number of message sent for any specific application and average
delay and increases the packet delivery ratio in the network.
The document discusses the Internet Protocol (IP) and its role in facilitating communication between hosts over the Internet. It describes the structure of IP addresses, including the use of four bytes to represent each address. It also discusses how classes and subnets divide IP addresses to allow for more devices to connect as the number of Internet users grows. Routers are used to create and manage subnets and allow communication between internal and external networks. Virtual IPs are also discussed as a way to extend the number of connectable devices since not all can be assigned a unique IP address.
Java Abs Peer To Peer Design & Implementation Of A Tuple Spacencct
Final Year Projects, IEEE Projects, Final Year Projects in Chennai, Final Year IEEE Projects, final year projects, college projects, student projects, java projects, asp.net projects, software projects, software ieee projects, ieee 2009 projects, 2009 ieee projects, embedded projects, final year software projects, final year embedded projects, ieee embedded projects, matlab projects, microcontroller projects, vlsi projects, dsp projects, free projects, project review, project report, project presentation, free source code, free project report, Final Year Projects, IEEE Projects, Final Year Projects in Chennai, Final Year IEEE Projects, final year projects, college projects, student projects, java projects, asp.net projects, software projects, software ieee projects, ieee 2009 projects, 2009 ieee projects, embedded projects, final year software projects, final year embedded projects, ieee embedded projects, matlab projects, final year java projects, final year asp.net projects, final year vb.net projects, vb.net projects, c# projects, final year c# projects, electrical projects, power electronics projects, motors and drives projects, robotics projects, ieee electrical projects, ieee power electronics projects, ieee robotics projects, power system projects, power system ieee projects, engineering projects, ieee engineering projects, engineering students projects, be projects, mca projects, mtech projects, btech projects, me projects, mtech projects, college projects, polytechnic projects, real time projects, ieee projects, non ieee projects, project presentation, project ppt, project pdf, project source code, project review, final year project, final year projects
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.
Improving the Proactive Routing Protocol using Depth First Iterative Deepenin...Yayah Zakaria
Owing to the wireless and mobility nature, nodes in a mobile ad hoc network are not within the transmission range. It needs to transfer data through the multi-intermediate nodes. Opportunistic data forwarding is an assuring solution to make use of the broadcast environment of wireless communication links. Due to absence of source routing capability with efficient proactive routing protocol, it is not widely used. To rectify the
problem, we proposed memory and routing efficient proactive routing protocol using Depth-First Iterative-Deepening and hello messaging scheme. This protocol can conserve the topology information in every node in the network. In experimental analysis and discussion, we implemented the proposed work using NS2 simulator tool and proved that the proposed technique is performed well in terms of average delay, buffer and throughput.
Analyse the performance of mobile peer to Peer network using ant colony optim...IJCI JOURNAL
The document describes analyzing the performance of a mobile peer-to-peer network using ant colony optimization. It proposes using a distributed spanning tree (DST) structure to improve efficiency by reducing the large number of messages. The DST is optimized using ant colony optimization to give an optimal solution. Simulation results show the approach reduces the number of messages, average delay, and increases packet delivery ratio in the 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
AN EFFICIENT ROUTING PROTOCOL FOR DELAY TOLERANT NETWORKS (DTNs)cscpconf
Delay-Tolerant Networks are those which lacks continuous communications among mobile
nodes . Distributed clustering scheme and cluster-based routing protocol are used for DelayTolerant
Mobile Networks (DTMNs). The basic idea is to distributive group mobile nodes with
similar mobility pattern into a cluster, which can then interchangeably share their resources for
overhead reduction and load balancing, aiming to achieve efficient and scalable routing in DTMN. Load balancing is carried out in two ways, Intra cluster load balancing and Inter cluster load balancing. The Convergence and stability become major challenges in distributed clustering in DTMN. An efficient routing protocol will be provided for the delay tolerant networks through which the stability of the network is maintained .Based on nodal contact probabilities, a set of functions including Sync(), Leave(), and Join() are devised for cluster formation and gateway selection. Finally, the gateway nodes exchange network information and perform routing
ANALYSE THE PERFORMANCE OF MOBILE PEER TO PEER NETWORK USING ANT COLONY OPTIM...ijcsity
A mobile peer-to-peer computer network is the one in which each computer in the network can act as a
client or server for the other computers in the network. The communication process among the nodes in the
mobile peer to peer network requires more no of messages. Due to this large number of messages passing,
propose an interconnection structure called distributed Spanning Tree (DST) and it improves the efficiency
of the mobile peer to peer network. The proposed method improves the data availability and consistency
across the entire network and also reduces the data latency and the required number of message passes for
any specific application in the network. Further to enhance the effectiveness of the proposed system, the
DST network is optimized with the Ant Colony Optimization method. It gives the optimal solution of the
DST method and increased availability, enhanced consistency and scalability of the network. The
simulation results shows that reduces the number of message sent for any specific application and average
delay and increases the packet delivery ratio in the network.
The document discusses the Internet Protocol (IP) and its role in facilitating communication between hosts over the Internet. It describes the structure of IP addresses, including the use of four bytes to represent each address. It also discusses how classes and subnets divide IP addresses to allow for more devices to connect as the number of Internet users grows. Routers are used to create and manage subnets and allow communication between internal and external networks. Virtual IPs are also discussed as a way to extend the number of connectable devices since not all can be assigned a unique IP address.
Presentation outline:
P2P Basics
Architecture
Lookup in P2P
Related work in P2P Lookup Protocols
Chord Protocol
Cluster based and Routing Balanced P2P Lookup Protocol
PathFinder
LiChord
Proposed P2P Lookup Model based on RCC8 and Scalable Bloom Filter
Future work for proposed P2P lookup model
The document discusses replica allocation techniques in mobile ad hoc networks to improve data accessibility in the presence of selfish nodes. It describes existing static and dynamic replica allocation methods, including Static Access Frequency, Dynamic Connectivity Based Grouping, and Dynamic Access Frequency and Neighborhood. It also proposes a Dynamic Connectivity Based Grouping method with selfish node detection to properly allocate replicas while accounting for selfish nodes. The goal is to maximize data accessibility, communication cost, and minimize query delay.
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...IJERD Editor
This document summarizes a research paper on secure cooperative caching in wireless peer-to-peer (P2P) networks. The paper proposes using encryption and digital signatures to authenticate data and ensure its integrity as it is cached and shared between wireless nodes. It describes greedy cache placement and mutual cache-based data access schemes to efficiently manage caching in wireless P2P networks. Experimental results show that a hybrid caching approach, which dynamically selects between caching data or caching paths to data, has better performance than solely caching data or paths. Future work areas include mechanisms for data owners to control caching scope while maintaining flexibility, and defending against security threats to data consistency.
Content Sharing over Smartphone-Based Delay-Tolerant NetworksIJERA Editor
With the growing number of smartphone end users, peer-to-peer ad hoc content giving is likely to occur often. Thus, new articles sharing mechanisms must be developed since traditional information delivery schemes will not be efficient with regard to content sharing due to the sporadic connectivity between smartphones on the market. To obtain data delivery such challenging environments, researchers include proposed the employment of store-carry-forward methodologies, in which a node stores a communication and holds it until a forwarding prospect arises through an encounter together with other nodes. Most past works in this field have dedicated to the conjecture of whether two nodes could encounter the other, without thinking about the place and also time from the encounter. In this particular paper, we propose to her discover-predict-deliver as a possible efficient articles sharing scheme for delay-tolerant touch screen phone networks. In this proposed scheme, contents are usually shared while using the mobility information of people. Specifically, our strategy employs the mobility understanding algorithm to spot places inside your own home and outdoor.
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
Migrating packet dropping in mobile ad hoc network based on modified ack-base...Alexander Decker
This document discusses migrating packet dropping in mobile ad-hoc networks using a modified ACK-based authentication scheme. It begins by introducing mobile ad-hoc networks and some of the security challenges they face, such as packet dropping attacks. It then discusses existing ACK-based authentication schemes and their limitations, such as generating a large overhead and having ambiguity when nodes refuse acknowledgments. The document proposes a modified ACK-based scheme that uses a secure channel to overcome these limitations and minimize packet dropping. It evaluates the proposed scheme through simulation experiments and finds it improves performance metrics like packet delivery ratio and throughput compared to not using a secure channel. In conclusion, the modified scheme is presented as a way to enhance node authentication and reduce packet dropping in mobile
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.
A Comprehensive Study on Vehicular Ad-Hoc Delay Tolerant Networking for Infra...inventionjournals
Generally, traditional networks presume the presence of some path between endpoints. Today, however, new applications, environments and types of devices are challenging these assumptions. In Delay Tolerant Networks (DTNs), an end-to-end path from source to destination may not exist. Nodes may connect and exchange their information in an opportunistic way. This book represents a broad overview of DTNs, particularly focusing on Vehicular Ad-hoc DTNs, their main characteristics, challenges and our research on this field. In the near future, cars are expected to be equipped with devices that will allow them to communicate wirelessly i.e. Wi-Fi. However, there will be strict restrictions to the duration of their connections with other vehicles, whereas the conditions of their links will greatly vary; DTNs as well as Ad-hoc DTNs present an attractive solution. Therefore, Vehicular Ad-hoc DTNs constitute an attractive research field. For practical implementation, we have used two Android devices for a little range of Wi-Fi. So, by this we are trying to give us better accuracy to go further. Thorough out this document, we have mentioned those techniques we came across and also those techniques and algorithms that we used in our proposed method.
Scale-Free Networks to Search in Unstructured Peer-To-Peer NetworksIOSR Journals
This document discusses using scale-free networks to improve search efficiency in unstructured peer-to-peer networks. It proposes the EQUATOR architecture, which creates an overlay network topology based on the scale-free Barabasi-Albert model. Simulation results show that EQUATOR achieves good lookup performance comparable to the ideal Barabasi-Albert network, with low message overhead even under node churn. The scale-free topology allows random walks to efficiently locate resources by directing searches to high-degree "hub" nodes with greater knowledge of the network.
The document proposes a distributed database system for delay tolerant networks (DTNs). It discusses two key contributions: 1) practical heuristics for query scheduling that optimize for lower delays by moving relevant data to sites with the largest amounts of data or highest likelihood of contact, and 2) a prereplication scheme that reduces latency for on-demand retrieval by actively pre-caching popular data between nodes based on bandwidth availability. The paper evaluates the proposed system through simulations on artificial and real-world DTN traces.
In the last decade Peer to Peer technology has been thoroughly explored, becauseit overcomes many limitations compared to the traditional client server paradigm. Despite its advantages over a traditional approach, the ubiquitous availability of high speed, high bandwidth and low latency networks has supported the traditional client-server paradigm. Recently, however, the surge of streaming services has spawned renewed interest in Peer to Peer technologies. In addition, services like geolocation databases and browser technologies like Web-RTC make a hybrid approach attractive.
In this paper we present algorithms for the construction and the maintenance of a hybrid P2P overlay multicast tree based on topological distances. The essential idea of these algorithms is to build a multicast tree by choosing neighbours close to each other. The topological distances can be easily obtained by the browser using the geolocation API. Thus the implementation of algorithms can be done web-based in a distributed manner.
We present proofs of our algorithms as well as experimental results and evaluations.
Maximizing P2P File Access Availability in Mobile Ad Hoc Networks though Repl...1crore projects
IEEE PROJECTS 2015
1 crore projects is a leading Guide for ieee Projects and real time projects Works Provider.
It has been provided Lot of Guidance for Thousands of Students & made them more beneficial in all Technology Training.
Dot Net
DOTNET Project Domain list 2015
1. IEEE based on datamining and knowledge engineering
2. IEEE based on mobile computing
3. IEEE based on networking
4. IEEE based on Image processing
5. IEEE based on Multimedia
6. IEEE based on Network security
7. IEEE based on parallel and distributed systems
Java Project Domain list 2015
1. IEEE based on datamining and knowledge engineering
2. IEEE based on mobile computing
3. IEEE based on networking
4. IEEE based on Image processing
5. IEEE based on Multimedia
6. IEEE based on Network security
7. IEEE based on parallel and distributed systems
ECE IEEE Projects 2015
1. Matlab project
2. Ns2 project
3. Embedded project
4. Robotics project
Eligibility
Final Year students of
1. BSc (C.S)
2. BCA/B.E(C.S)
3. B.Tech IT
4. BE (C.S)
5. MSc (C.S)
6. MSc (IT)
7. MCA
8. MS (IT)
9. ME(ALL)
10. BE(ECE)(EEE)(E&I)
TECHNOLOGY USED AND FOR TRAINING IN
1. DOT NET
2. C sharp
3. ASP
4. VB
5. SQL SERVER
6. JAVA
7. J2EE
8. STRINGS
9. ORACLE
10. VB dotNET
11. EMBEDDED
12. MAT LAB
13. LAB VIEW
14. Multi Sim
CONTACT US
1 CRORE PROJECTS
Door No: 214/215,2nd Floor,
No. 172, Raahat Plaza, (Shopping Mall) ,Arcot Road, Vadapalani, Chennai,
Tamin Nadu, INDIA - 600 026
Email id: 1croreprojects@gmail.com
website:1croreprojects.com
Phone : +91 97518 00789 / +91 72999 51536
Iaetsd an enhancement for content sharing overIaetsd Iaetsd
This document proposes an efficient content sharing scheme called discover-predict-deliver (DPD) for smartphone-based delay tolerant networks (DTNs). It uses a hidden Markov model to predict individual mobility information to improve content discovery and delivery. The existing system results in a 2% CPU overhead and reduces smartphone battery life by 15%. To minimize energy consumption, the document proposes using sensor scheduling schemes in an opportunistic context. It aims to solve problems in content sharing and minimize energy use through sensor scheduling.
This document summarizes a research paper that proposes a replica placement scheme called MAXDISJOINT for tree-based routing distributed hash tables (DHTs). MAXDISJOINT aims to create route diversity between replicas to improve routing robustness against node failures or attacks. The paper proves that MAXDISJOINT creates disjoint routes and evaluates its performance using simulations of the Pastry DHT. Simulation results show that with MAXDISJOINT placement, lookups can still succeed with a high probability even when a significant portion of the network is compromised. The paper also explores using a technique called neighbor set routing to further increase route diversity.
A collaborative wireless sensor network routingambitlick
This document proposes a new routing scheme called node reliance for wireless sensor networks. Node reliance rates how much each node is relied upon for routing data from sources to sinks. Sources will route through nodes with low reliance ratings to avoid overusing critical nodes and maximize network lifetime. The scheme is evaluated using an example network and compared to other routing methods. Node reliance aims to reduce energy waste from disconnected sources by encouraging collaboration between sources in path selection.
The document lists 23 networking and mobile computing projects implemented in NS2. It includes projects on topics like mobility in wireless networks, intrusion detection, neighbor discovery, energy renewal with wireless power transfer, load balancing, and spectrum access control. Contact information is provided for those interested in the project reports, presentations, source code, or implementing new projects.
Stichting van Jacques van Ypersele kreeg geld van rijke AmerikanenThierry Debels
Op 24 september 2013 richt Jacques van Ypersele de stichting Centres Ste-Thérèse op.
Uit een officieel document van de VS blijkt dat de stichting een geldsom kreeg van rijke Amerikaanse geldschieters.
Het gaat meer bepaald om The Stephen F & Camilla T Brauer Charitable Trust.
Hieruit blijkt dat Van Ypersele nog steeds erg goede relaties onderhoudt met onderdanen uit de Verenigde Staten.
Onder het mom van COREPER worden door de Europese Unie diverse snoepreisjes georganiseerd. OPvallend is dat de leden van COREPER hun partner kunnen meenemen op een dergelijke snoepreis.
Admission control and routing in multi hop wireless networksambitlick
This document summarizes a research paper about admission control and routing algorithms for multi-hop wireless networks that provide quality of service guarantees for flows requesting a pre-specified bandwidth. The paper develops an optimal admission control and routing algorithm that has performance close to an offline algorithm with complete future knowledge. The algorithm makes no assumptions about flow arrival patterns and can be implemented in a distributed manner. It also proves this algorithm is asymptotically optimal with respect to competitive ratio, a metric that measures performance compared to an offline algorithm. Finally, the paper discusses how the algorithm can be modified to allow use of standard shortest-path algorithms for distributed implementation.
Presentation outline:
P2P Basics
Architecture
Lookup in P2P
Related work in P2P Lookup Protocols
Chord Protocol
Cluster based and Routing Balanced P2P Lookup Protocol
PathFinder
LiChord
Proposed P2P Lookup Model based on RCC8 and Scalable Bloom Filter
Future work for proposed P2P lookup model
The document discusses replica allocation techniques in mobile ad hoc networks to improve data accessibility in the presence of selfish nodes. It describes existing static and dynamic replica allocation methods, including Static Access Frequency, Dynamic Connectivity Based Grouping, and Dynamic Access Frequency and Neighborhood. It also proposes a Dynamic Connectivity Based Grouping method with selfish node detection to properly allocate replicas while accounting for selfish nodes. The goal is to maximize data accessibility, communication cost, and minimize query delay.
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...IJERD Editor
This document summarizes a research paper on secure cooperative caching in wireless peer-to-peer (P2P) networks. The paper proposes using encryption and digital signatures to authenticate data and ensure its integrity as it is cached and shared between wireless nodes. It describes greedy cache placement and mutual cache-based data access schemes to efficiently manage caching in wireless P2P networks. Experimental results show that a hybrid caching approach, which dynamically selects between caching data or caching paths to data, has better performance than solely caching data or paths. Future work areas include mechanisms for data owners to control caching scope while maintaining flexibility, and defending against security threats to data consistency.
Content Sharing over Smartphone-Based Delay-Tolerant NetworksIJERA Editor
With the growing number of smartphone end users, peer-to-peer ad hoc content giving is likely to occur often. Thus, new articles sharing mechanisms must be developed since traditional information delivery schemes will not be efficient with regard to content sharing due to the sporadic connectivity between smartphones on the market. To obtain data delivery such challenging environments, researchers include proposed the employment of store-carry-forward methodologies, in which a node stores a communication and holds it until a forwarding prospect arises through an encounter together with other nodes. Most past works in this field have dedicated to the conjecture of whether two nodes could encounter the other, without thinking about the place and also time from the encounter. In this particular paper, we propose to her discover-predict-deliver as a possible efficient articles sharing scheme for delay-tolerant touch screen phone networks. In this proposed scheme, contents are usually shared while using the mobility information of people. Specifically, our strategy employs the mobility understanding algorithm to spot places inside your own home and outdoor.
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
Migrating packet dropping in mobile ad hoc network based on modified ack-base...Alexander Decker
This document discusses migrating packet dropping in mobile ad-hoc networks using a modified ACK-based authentication scheme. It begins by introducing mobile ad-hoc networks and some of the security challenges they face, such as packet dropping attacks. It then discusses existing ACK-based authentication schemes and their limitations, such as generating a large overhead and having ambiguity when nodes refuse acknowledgments. The document proposes a modified ACK-based scheme that uses a secure channel to overcome these limitations and minimize packet dropping. It evaluates the proposed scheme through simulation experiments and finds it improves performance metrics like packet delivery ratio and throughput compared to not using a secure channel. In conclusion, the modified scheme is presented as a way to enhance node authentication and reduce packet dropping in mobile
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.
A Comprehensive Study on Vehicular Ad-Hoc Delay Tolerant Networking for Infra...inventionjournals
Generally, traditional networks presume the presence of some path between endpoints. Today, however, new applications, environments and types of devices are challenging these assumptions. In Delay Tolerant Networks (DTNs), an end-to-end path from source to destination may not exist. Nodes may connect and exchange their information in an opportunistic way. This book represents a broad overview of DTNs, particularly focusing on Vehicular Ad-hoc DTNs, their main characteristics, challenges and our research on this field. In the near future, cars are expected to be equipped with devices that will allow them to communicate wirelessly i.e. Wi-Fi. However, there will be strict restrictions to the duration of their connections with other vehicles, whereas the conditions of their links will greatly vary; DTNs as well as Ad-hoc DTNs present an attractive solution. Therefore, Vehicular Ad-hoc DTNs constitute an attractive research field. For practical implementation, we have used two Android devices for a little range of Wi-Fi. So, by this we are trying to give us better accuracy to go further. Thorough out this document, we have mentioned those techniques we came across and also those techniques and algorithms that we used in our proposed method.
Scale-Free Networks to Search in Unstructured Peer-To-Peer NetworksIOSR Journals
This document discusses using scale-free networks to improve search efficiency in unstructured peer-to-peer networks. It proposes the EQUATOR architecture, which creates an overlay network topology based on the scale-free Barabasi-Albert model. Simulation results show that EQUATOR achieves good lookup performance comparable to the ideal Barabasi-Albert network, with low message overhead even under node churn. The scale-free topology allows random walks to efficiently locate resources by directing searches to high-degree "hub" nodes with greater knowledge of the network.
The document proposes a distributed database system for delay tolerant networks (DTNs). It discusses two key contributions: 1) practical heuristics for query scheduling that optimize for lower delays by moving relevant data to sites with the largest amounts of data or highest likelihood of contact, and 2) a prereplication scheme that reduces latency for on-demand retrieval by actively pre-caching popular data between nodes based on bandwidth availability. The paper evaluates the proposed system through simulations on artificial and real-world DTN traces.
In the last decade Peer to Peer technology has been thoroughly explored, becauseit overcomes many limitations compared to the traditional client server paradigm. Despite its advantages over a traditional approach, the ubiquitous availability of high speed, high bandwidth and low latency networks has supported the traditional client-server paradigm. Recently, however, the surge of streaming services has spawned renewed interest in Peer to Peer technologies. In addition, services like geolocation databases and browser technologies like Web-RTC make a hybrid approach attractive.
In this paper we present algorithms for the construction and the maintenance of a hybrid P2P overlay multicast tree based on topological distances. The essential idea of these algorithms is to build a multicast tree by choosing neighbours close to each other. The topological distances can be easily obtained by the browser using the geolocation API. Thus the implementation of algorithms can be done web-based in a distributed manner.
We present proofs of our algorithms as well as experimental results and evaluations.
Maximizing P2P File Access Availability in Mobile Ad Hoc Networks though Repl...1crore projects
IEEE PROJECTS 2015
1 crore projects is a leading Guide for ieee Projects and real time projects Works Provider.
It has been provided Lot of Guidance for Thousands of Students & made them more beneficial in all Technology Training.
Dot Net
DOTNET Project Domain list 2015
1. IEEE based on datamining and knowledge engineering
2. IEEE based on mobile computing
3. IEEE based on networking
4. IEEE based on Image processing
5. IEEE based on Multimedia
6. IEEE based on Network security
7. IEEE based on parallel and distributed systems
Java Project Domain list 2015
1. IEEE based on datamining and knowledge engineering
2. IEEE based on mobile computing
3. IEEE based on networking
4. IEEE based on Image processing
5. IEEE based on Multimedia
6. IEEE based on Network security
7. IEEE based on parallel and distributed systems
ECE IEEE Projects 2015
1. Matlab project
2. Ns2 project
3. Embedded project
4. Robotics project
Eligibility
Final Year students of
1. BSc (C.S)
2. BCA/B.E(C.S)
3. B.Tech IT
4. BE (C.S)
5. MSc (C.S)
6. MSc (IT)
7. MCA
8. MS (IT)
9. ME(ALL)
10. BE(ECE)(EEE)(E&I)
TECHNOLOGY USED AND FOR TRAINING IN
1. DOT NET
2. C sharp
3. ASP
4. VB
5. SQL SERVER
6. JAVA
7. J2EE
8. STRINGS
9. ORACLE
10. VB dotNET
11. EMBEDDED
12. MAT LAB
13. LAB VIEW
14. Multi Sim
CONTACT US
1 CRORE PROJECTS
Door No: 214/215,2nd Floor,
No. 172, Raahat Plaza, (Shopping Mall) ,Arcot Road, Vadapalani, Chennai,
Tamin Nadu, INDIA - 600 026
Email id: 1croreprojects@gmail.com
website:1croreprojects.com
Phone : +91 97518 00789 / +91 72999 51536
Iaetsd an enhancement for content sharing overIaetsd Iaetsd
This document proposes an efficient content sharing scheme called discover-predict-deliver (DPD) for smartphone-based delay tolerant networks (DTNs). It uses a hidden Markov model to predict individual mobility information to improve content discovery and delivery. The existing system results in a 2% CPU overhead and reduces smartphone battery life by 15%. To minimize energy consumption, the document proposes using sensor scheduling schemes in an opportunistic context. It aims to solve problems in content sharing and minimize energy use through sensor scheduling.
This document summarizes a research paper that proposes a replica placement scheme called MAXDISJOINT for tree-based routing distributed hash tables (DHTs). MAXDISJOINT aims to create route diversity between replicas to improve routing robustness against node failures or attacks. The paper proves that MAXDISJOINT creates disjoint routes and evaluates its performance using simulations of the Pastry DHT. Simulation results show that with MAXDISJOINT placement, lookups can still succeed with a high probability even when a significant portion of the network is compromised. The paper also explores using a technique called neighbor set routing to further increase route diversity.
A collaborative wireless sensor network routingambitlick
This document proposes a new routing scheme called node reliance for wireless sensor networks. Node reliance rates how much each node is relied upon for routing data from sources to sinks. Sources will route through nodes with low reliance ratings to avoid overusing critical nodes and maximize network lifetime. The scheme is evaluated using an example network and compared to other routing methods. Node reliance aims to reduce energy waste from disconnected sources by encouraging collaboration between sources in path selection.
The document lists 23 networking and mobile computing projects implemented in NS2. It includes projects on topics like mobility in wireless networks, intrusion detection, neighbor discovery, energy renewal with wireless power transfer, load balancing, and spectrum access control. Contact information is provided for those interested in the project reports, presentations, source code, or implementing new projects.
Stichting van Jacques van Ypersele kreeg geld van rijke AmerikanenThierry Debels
Op 24 september 2013 richt Jacques van Ypersele de stichting Centres Ste-Thérèse op.
Uit een officieel document van de VS blijkt dat de stichting een geldsom kreeg van rijke Amerikaanse geldschieters.
Het gaat meer bepaald om The Stephen F & Camilla T Brauer Charitable Trust.
Hieruit blijkt dat Van Ypersele nog steeds erg goede relaties onderhoudt met onderdanen uit de Verenigde Staten.
Onder het mom van COREPER worden door de Europese Unie diverse snoepreisjes georganiseerd. OPvallend is dat de leden van COREPER hun partner kunnen meenemen op een dergelijke snoepreis.
Admission control and routing in multi hop wireless networksambitlick
This document summarizes a research paper about admission control and routing algorithms for multi-hop wireless networks that provide quality of service guarantees for flows requesting a pre-specified bandwidth. The paper develops an optimal admission control and routing algorithm that has performance close to an offline algorithm with complete future knowledge. The algorithm makes no assumptions about flow arrival patterns and can be implemented in a distributed manner. It also proves this algorithm is asymptotically optimal with respect to competitive ratio, a metric that measures performance compared to an offline algorithm. Finally, the paper discusses how the algorithm can be modified to allow use of standard shortest-path algorithms for distributed implementation.
Architecture for reliable service discoveryambitlick
The authors provide an overview of standardized service discovery and delivery solutions for mobile ad hoc networks (MANETs). They propose a novel architecture that allows selecting a service provider based on metrics like the power supply of the provider and the path to the destination. This architecture is based on a decentralized approach using Service Location Protocol (SLP) extensions. It aims to provide reliable service discovery and delivery in MANETs by considering energy constraints affecting the network topology and connectivity.
TCP Fairness for Uplink and Downlink Flows in WLANsambitlick
The document discusses simulation parameters for a wireless network simulation in NS2 including the simulator used, simulation time, packet interval, background data traffic types, packet size, transmission range, routing protocol, and MAC protocol. It also briefly discusses queue management for wireless networks and how it differs from wired networks by not being visible. Client-server computing and networking is defined as well as packet scheduling policies and comparing FIFO to RENO scheduling. Bandwidth sharing approaches are examined for heterogeneous congestion control protocols. Finally, end-to-end throughput is defined and its decrease with more nodes in a chain topology is explained due to relaying overhead at intermediate nodes.
A novel pause count backoff algorithm for channel accessambitlick
The document summarizes a proposed novel backoff algorithm called Pause Count Backoff (PCB) for channel access in IEEE 802.11 wireless networks. PCB observes the number of pauses in a node's backoff procedure to estimate the number of active stations and set an appropriate contention window size. Simulation results show PCB outperforms other algorithms like DCF, EIED, and AEDCF in terms of goodput, fairness index, and end-to-end delay under different network conditions.
Hoe rijk is eurocommissaris Marianne Thyssen?Thierry Debels
Nieuwbakken eurocommissaris Marianne Thyssen moet een verklaring afleggen over haar vermogen. Uit dat document blijkt dat ze een aanzienlijk roerend en onroerend vermogen heeft.
On Multihop Distances in Wireless Sensor Networks with Random Node Locationsambitlick
The document analyzes the distribution of maximum multihop distances in wireless sensor networks with random node locations. It proposes a greedy method to maximize the multihop distance in 2D networks by restricting propagation direction outward from the source in each hop and searching for the furthest neighbor. This differs from prior work modeling 1D networks that used a Gaussian distribution, which is shown to not accurately model 2D distances. The paper transforms the Gamma distribution to effectively approximate maximum distances in 2D and provides a more consistent representation of the multihop distance distribution compared to the Gaussian model. It derives the expected value and standard deviation of distances using the Gamma approximation and compares to simulation results.
1) This document proposes a new anomaly detection scheme for mobile ad hoc networks (MANETs) based on dynamic learning.
2) The scheme uses a statistical decision theory to calculate projection distances between the current and normal states of nodes using weighted coefficients and a forgetting curve.
3) Simulations evaluate the scheme's effectiveness in detecting five types of attacks on the ad hoc on-demand distance vector (AODV) routing protocol for MANETs of different sizes.
The document proposes an energy-efficient service discovery protocol for wireless sensor networks. It uses a lightweight clustering algorithm to build a distributed directory of service registrations. Each cluster head maintains information about services in its cluster. The protocol aims to minimize communication costs during service discovery and maintenance of the distributed directory. It constructs disjoint tree clusters where high-capability nodes become cluster heads and advertise their roles, while other nodes select parents based on capability grades.
Energy efficient cluster-based service discovery in wireless sensor networksambitlick
1) The document proposes an energy-efficient service discovery protocol for wireless sensor networks that exploits a cluster-based network overlay.
2) Clusterhead nodes form a distributed service registry to minimize communication costs during service discovery and maintenance.
3) The performance of the proposed integrated clustering and service discovery solution is evaluated through simulations under different network conditions.
Waarom proces van Delphine wellicht met sisser aflooptThierry Debels
Op 23 september 2014 wordt het proces van Delphine Boël verdergezet. Hierop krijgt de rechtbank een maand de tijd om een uitspraak te doen. Tegen deze uitspraak kan vervolgens door alle partijen hoger beroep aangetekend worden.
De rechtszaak met dossiernummer 2013 / A / 9225 bevat enkele merkwaardige elementen.
Jean-Jacques De Gucht heeft een passie voor kunst. In interviews heeft hij al eens gezegd dat hij er stiekem van droomt om galeriehouder te worden.
Die droom heeft hij nu waargemaakt. In het voorjaar van 2014 heeft De Gucht met zijn echtgenote de bvba De filatuur opgericht, een kunstgalerie.
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.
Java Abs Peer To Peer Design & Implementation Of A Tuple Sncct
final Year Projects, Final Year Projects in Chennai, Software Projects, Embedded Projects, Microcontrollers Projects, DSP Projects, VLSI Projects, Matlab Projects, Java Projects, .NET Projects, IEEE Projects, IEEE 2009 Projects, IEEE 2009 Projects, Software, IEEE 2009 Projects, Embedded, Software IEEE 2009 Projects, Embedded IEEE 2009 Projects, Final Year Project Titles, Final Year Project Reports, Final Year Project Review, Robotics Projects, Mechanical Projects, Electrical Projects, Power Electronics Projects, Power System Projects, Model Projects, Java Projects, J2EE Projects, Engineering Projects, Student Projects, Engineering College Projects, MCA Projects, BE Projects, BTech Projects, ME Projects, MTech Projects, Wireless Networks Projects, Network Security Projects, Networking Projects, final year projects, ieee projects, student projects, college projects, ieee projects in chennai, java projects, software ieee projects, embedded ieee projects, "ieee2009projects", "final year projects", "ieee projects", "Engineering Projects", "Final Year Projects in Chennai", "Final year Projects at Chennai", Java Projects, ASP.NET Projects, VB.NET Projects, C# Projects, Visual C++ Projects, Matlab Projects, NS2 Projects, C Projects, Microcontroller Projects, ATMEL Projects, PIC Projects, ARM Projects, DSP Projects, VLSI Projects, FPGA Projects, CPLD Projects, Power Electronics Projects, Electrical Projects, Robotics Projects, Solor Projects, MEMS Projects, J2EE Projects, J2ME Projects, AJAX Projects, Structs Projects, EJB Projects, Real Time Projects, Live Projects, Student Projects, Engineering Projects, MCA Projects, MBA Projects, College Projects, BE Projects, BTech Projects, ME Projects, MTech Projects, M.Sc Projects, Final Year Java Projects, Final Year ASP.NET Projects, Final Year VB.NET Projects, Final Year C# Projects, Final Year Visual C++ Projects, Final Year Matlab Projects, Final Year NS2 Projects, Final Year C Projects, Final Year Microcontroller Projects, Final Year ATMEL Projects, Final Year PIC Projects, Final Year ARM Projects, Final Year DSP Projects, Final Year VLSI Projects, Final Year FPGA Projects, Final Year CPLD Projects, Final Year Power Electronics Projects, Final Year Electrical Projects, Final Year Robotics Projects, Final Year Solor Projects, Final Year MEMS Projects, Final Year J2EE Projects, Final Year J2ME Projects, Final Year AJAX Projects, Final Year Structs Projects, Final Year EJB Projects, Final Year Real Time Projects, Final Year Live Projects, Final Year Student Projects, Final Year Engineering Projects, Final Year MCA Projects, Final Year MBA Projects, Final Year College Projects, Final Year BE Projects, Final Year BTech Projects, Final Year ME Projects, Final Year MTech Projects, Final Year M.Sc Projects, IEEE Java Projects, ASP.NET Projects, VB.NET Projects, C# Projects, Visual C++ Projects, Matlab Projects, NS2 Projects, C Projects, Microcontroller Projects, ATMEL Projects, PIC Projects, ARM Projects, DSP Projects, VLSI Projects, FPGA Projects, CPLD Projects, Power Electronics Projects, Electrical Projects, Robotics Projects, Solor Projects, MEMS Projects, J2EE Projects, J2ME Projects, AJAX Projects, Structs Projects, EJB Projects, Real Time Projects, Live Projects, Student Projects, Engineering Projects, MCA Projects, MBA Projects, College Projects, BE Projects, BTech Projects, ME Projects, MTech Projects, M.Sc Projects, IEEE 2009 Java Projects, IEEE 2009 ASP.NET Projects, IEEE 2009 VB.NET Projects, IEEE 2009 C# Projects, IEEE 2009 Visual C++ Projects, IEEE 2009 Matlab Projects, IEEE 2009 NS2 Projects, IEEE 2009 C Projects, IEEE 2009 Microcontroller Projects, IEEE 2009 ATMEL Projects, IEEE 2009 PIC Projects, IEEE 2009 ARM Projects, IEEE 2009 DSP Projects, IEEE 2009 VLSI Projects, IEEE 2009 FPGA Projects, IEEE 2009 CPLD Projects, IEEE 2009 Power Electronics Projects, IEEE 2009 Electrical Projects, IEEE 2009 Robotics Projects, IEEE 2009 Solor Projects, IEEE 2009 MEMS Projects, IEEE 2009 J2EE P
Peer-to-peer (P2P) networks are a type of computer network architecture where individuals form a loose group to share resources directly with others in the group without a centralized server. There are two main types of P2P network structures - unstructured and structured. Unstructured networks do not use algorithms to organize the network, while structured networks use algorithms to optimize routing. Popular applications of P2P networking include file sharing, media streaming, grid computing, instant messaging, and voice over internet protocol.
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
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.
A NEW ALGORITHM FOR CONSTRUCTION OF A P2P MULTICAST HYBRID OVERLAY TREE BASED...csandit
In the last decade Peer to Peer technology has been thoroughly explored, because it overcomes many limitations compared to the traditional client server paradigm. Despite its advantages over a traditional approach, the ubiquitous availability of high speed, high bandwidth and low latency networks has supported the traditional client-server paradigm. Recently, however, the surge of streaming services has spawned renewed interest in Peer to Peer technologies. In addition, services like geolocation databases and browser technologies like Web-RTC make a hybrid approach attractive.
A NEW ALGORITHM FOR CONSTRUCTION OF A P2P MULTICAST HYBRID OVERLAY TREE BASED...cscpconf
In the last decade Peer to Peer technology has been thoroughly explored, because it overcomes many limitations compared to the traditional client server paradigm. Despite its advantages over a traditional approach, the ubiquitous availability of high speed, high bandwidth and low latency networks has supported the traditional client-server paradigm. Recently, however, the surge of streaming services has spawned renewed interest in Peer to Peer technologies. In addition, services like geolocation databases and browser technologies like Web-RTC make a hybrid approach attractive. In this paper we present algorithms for the construction and the maintenance of a hybrid P2P overlay multicast tree based on topological distances. The essential idea of these algorithms is to build a multicast tree by choosing neighbours close to each other. The topological distances can be easily obtained by the browser using the geolocation API. Thus the implementation of algorithms can be done web-based in a distributed manner. We present proofs of our algorithms as well as practical results and evaluations.
This document summarizes an research paper that proposes a new approach called Selfish Neighbor Selection (SNS) for optimizing overlay networks. SNS limits the number of neighbors for each peer to reduce overhead. It also considers both cooperative and selfish behavior among peers. The paper describes EGOIST, a prototype overlay routing system implemented based on SNS that expresses each node's "best response" wiring strategy as a k-median problem. EGOIST evaluations on PlanetLab show it performs better than heuristic overlays on metrics like delay, bandwidth, and node utilization. The paper also proposes using an Optimized Spanning Tree Protocol to allow traffic on the full-mesh portion to take shortest paths, and describes how EGOIST
Study on security and quality of service implementations in p2 p overlay netw...eSAT Publishing House
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
This document discusses peer-to-peer systems and middleware for managing distributed resources at a large scale. It describes key characteristics of peer-to-peer systems like nodes contributing equal resources and decentralized operation. Middleware systems like Pastry and Tapestry are overlay networks that route requests to distributed objects across nodes through knowledge at each node. They provide simple APIs and support scalability, load balancing, and dynamic node availability.
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.
This document discusses computer networking and network topologies. It defines a network as a set of interconnected computer systems that allow sharing of resources and communication using common protocols. Networks use packets to transmit data over digital connections. Common network topologies include bus, star, ring, mesh, fully connected, and tree networks. Overlay networks are virtual networks built over an existing underlying network. The document also discusses different types of network links that can be used to connect devices, such as electrical cable, optical fiber, wireless, and power lines.
Throughput Maximization using Spatial Reusability in Multi Hop Wireless Networkijtsrd
Energy is a valuable resource in wireless networks. For many multi hop networking scenarios, nodes require power for performing their operation, so requiring capable power management to make certain connectivity across the network. Though when wireless networks are attached outside power source due to obstruction between lively links the network may demand excessive energy per unit time Power due to this the overall performance is reduced. Since network life time or network capacity is depend on the power efficiency, many efforts to study energy efficient networks in the wireless network community. In multi hop wireless networks well organized routing algorithms are significant for network performance. We dispute that by carefully considering spatial reusability of the wireless communication media, we can reduce the overhead in multi hop wireless networks. To support our argument, propose Broadcast Tree Construction BTC and compare them with existing routing and multi path routing protocols, respectively. Our estimate outcome shows that proposed protocols significantly improve the end to end throughput compared with existing protocols. We also introduce the sleep scheduling approach for energy consumption and hybrid cryptography for security of data that can be prevent the data leakage and jammer attacks. Twinkal P. Dudhagawali | Prof. L. K. Gautam | Prof. V. P. Vaidya ""Throughput Maximization using Spatial Reusability in Multi Hop Wireless Network"" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-3 | Issue-3 , April 2019, URL: https://www.ijtsrd.com/papers/ijtsrd23489.pdf
Paper URL: https://www.ijtsrd.com/engineering/information-technology/23489/throughput-maximization-using-spatial-reusability-in-multi-hop-wireless-network/twinkal-p-dudhagawali
This document proposes a framework for managing and querying multidimensional historical data distributed across an unstructured peer-to-peer network. When data is published, it is summarized into a synopsis containing an index and sub-synopses storing compressed data portions. The index and sub-synopses are distributed across the network with suitable replication to provide coverage for queries. The framework supports data publication and querying while ensuring peer autonomy and handling volatile peers through prompt reaction to disconnections to prevent broken index references.
This document discusses web and social computing and peer-to-peer networks. It provides an overview of peer-to-peer network types including unstructured and structured networks. It also describes PeerSim, a peer-to-peer network simulator. The document outlines implementing maximum and minimum functions in PeerSim and analyzing the results. New methods were designed, run, and graphs of the outputs were generated to study how the maximum and minimum values changed over simulations.
Peer to Peer Approach based Replica and Locality Awareness to Manage and Diss...IJCNCJournal
Distributed Hash Table (DHT) based structured peer-to-peer (P2P) systems provide an efficient method of disseminating information in a VANET environment owing to its high performance and properties (e.g., self-organization, decentralization, scalability, etc.). The topology of ad hoc vehicle networks (VANET) varies dynamically; its disconnections are frequent due to the high movement of vehicles. In such a topology, information availability is an ultimate problem for vehicles, in general, connect and disconnect frequently from the network. Data replication is an appropriate and adequate solution to this problem. In this contribution, to increase the accessibility of data, which also increases the success rate of the lookup, a method based on replication in the Vanet network is proposed named LAaR-Vanet. Also, this replication strategy is combined with a locality-awareness method to promote the same purpose and to avoid the problems of long paths. The performance of the proposed solution is assessed by a series of in-depth simulations in urban areas. The obtained results indicate the efficiency of the proposed approach, in terms of the following metrics: lookup success rate, the delay, and the number of the logical hop.
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.
Similar to An efficient hybrid peer to-peersystemfordistributeddatasharing (20)
Bulk Projects For sale
IEEE 2009-10-11-12-13 PAPERS AVILABLE.
We are providing low cost project for final year student projects.
Solved 2010 -2011 -2012 - 2013 IEEE in all the domain
Mobile : 8940956123
E-Mail : ambitlick@gmail.com,
INNOVATIVE TITLES ARE ALSO WELLCOME TO DO WITH US
For All BE/BTech, ME/MTech, MSC/MCA/MS , and diplamo graduates
PROJECT SUPPORTS & DELIVERABLES
•Project Abstract
•IEEE Paper
•PPT / Review Details
•Project Report
•Working Procedure in Video
•Screen Shots
•Materials & Books in CD
•Project Certification
This document lists 15 potential 2013 IEEE NS2 project titles related to wireless networks and sensor networks. It includes projects on topics like capacity of hybrid wireless mesh networks, delay-optimal broadcast in multihop wireless networks, detection of spoofing attackers, and harvesting-aware energy management for wireless sensor networks. The document provides contact information for a company called Ambitlick Solutions that offers support and deliverables for IEEE projects, including project abstracts, papers, presentations, reports, and certification.
This document lists over 40 potential 2013 IEEE Java Dotnet project titles across various domains including wireless networks, mobile computing, network security, data mining, cloud computing, parallel and distributed computing, and multimedia/image processing. The projects focus on technical topics such as wireless sensor networks, wireless mesh networks, cognitive radio networks, mobile ad hoc networks, network coding, video streaming, machine learning, data warehousing, and more. Project deliverables include an abstract, IEEE paper, presentation, report, working prototype or proof of concept, and certification. Bulk older and new projects can be provided at a low cost.
Handling selfishness in replica allocationambitlick
The document discusses techniques for handling selfish nodes in replica allocation over mobile ad hoc networks. It aims to reduce traffic overhead while maintaining high data accessibility. The techniques include a selfish node detection algorithm that considers partial selfishness and novel replica allocation methods to address issues caused by selfish nodes hoarding replicas for their own benefit instead of sharing memory space. Simulations are used to evaluate the performance of these techniques in improving data delivery rates.
Mutual Distance Bounding Protocols enable entities to determine an upper bound on their physical distance and authenticate each other. They have been actively researched due to distance-based attacks on wireless systems like RFID. While most protocols provide unilateral authentication of a tag to a reader, one was proposed to provide mutual authentication with a lower false acceptance rate. However, this analysis is shown to overestimate security, as a new attack achieves a higher false acceptance rate. A method is also introduced to modify existing unilateral authentication protocols into mutual authentication protocols.
Moderated group authoring system for campus wide workgroupsambitlick
This paper describes a distributed authoring system for campus workgroups that allows group members to modify any document type using their own devices. Each member maintains an updatable copy of shared content, and read-only copies are distributed asynchronously based on wireless availability. Group members manually reconcile updates through moderation, merging changes from others into their copy. Over time, successive moderation converges all copies into a single version. An evaluation found the asynchronous update model effective and the moderation process intuitive for students.
Efficient spread spectrum communication without pre shared secretsambitlick
This document proposes a new mechanism called Time Reversed Message Extraction and Key Scheduling (TREKS) that allows for efficient spread spectrum communication without pre-shared secrets. TREKS is four orders of magnitude faster than previous solutions to this problem. It enables long-term spread spectrum communication with optimal energy costs, minimal storage overhead, and a computation cost at most twice traditional spread spectrum. The approach was evaluated through simulations and experiments sustaining 1Mbps communication spread over 100 Megachips per second using modest hardware.
Adaptive weight factor estimation from user review 1ambitlick
This document proposes a novel technique called Adjacent Pair Priorities (APP) to estimate weight factors for quality of service parameters in vertical handoff decision algorithms. The APP technique allows users to set relative priority levels for adjacent pairs of QoS parameters in descending order using an exponential mapping. This adaptive approach controls the width of the weight distribution to provide flexibility for users. The document outlines the system requirements, block diagram, modules and references several research papers on vertical handoff decision schemes and network selection algorithms.
The document proposes an Integrated Institutional Portal that allows all colleges and institutions within a university or district to share information. [1] The portal would allow students and staff from different colleges to discuss and request information from one another through blogs, forums and by publishing notices. [2] Currently, each college maintains separate portals without a common forum for communication. [3] The proposed centralized portal managed by a super administrator would make all college information like results, events and departments accessible to benefit students and staff across institutions.
This document describes an Embassy Administration portal that aims to centralize and automate manual processes at a college. [1] The portal allows separate login access for staff, parents, students and other members of the college. [2] It displays student results, attendance, and performance for parents to view as well as enables communication between parents and faculty. [3] The system conducts model and unit examinations.
The document proposes a customer relationship management system (CRMS) to help space marketing executives, managers, and management interact and share information online. The existing CRMS is manual and DOS-based, which has disadvantages like a distributed database, obsolete technology, and low efficiency. The proposed system is a web-based online CRMS designed for the space marketing department. It allows monitoring executive calls, tracking performance, and sharing information among departments to improve customer relationships and business operations. The system has modules for corporate administration, regional management, center management, and executives to organize work and monitor progress at different levels.
Mutual Distance Bounding Protocols enable entities to determine an upper bound on their physical distance and authenticate each other. They have been actively researched due to distance-based attacks on wireless systems like RFID. While most protocols provide unilateral authentication of a tag to a reader, one was proposed to provide mutual authentication with a lower false acceptance rate. However, this analysis is shown to overestimate security, as a new attack achieves a higher false acceptance rate. A method is also introduced to modify existing unilateral authentication protocols into mutual authentication protocols.
Moderated group authoring system for campus wide workgroupsambitlick
This paper describes a distributed authoring system for campus workgroups that allows group members to modify any document type using their own devices. Each member maintains an updatable copy of shared content, and read-only copies are distributed based on wireless availability. Group members manually reconcile updates through moderation, merging changes from others into their copy. Over time, successive moderations converge the multiple versions into a single version. An evaluation found the asynchronous update propagation and moderation process intuitive for students.
Efficient spread spectrum communication without pre shared secretsambitlick
This document proposes a new mechanism called Time Reversed Message Extraction and Key Scheduling (TREKS) that allows for efficient spread spectrum communication without pre-shared secrets. TREKS is four orders of magnitude faster than previous solutions to this problem and enables long-term spread spectrum communication without establishing keys. It was evaluated through simulation and on a testbed and can sustain 1Mbps communication spread over a 100 Megachips bandwidth in real-time, with provably optimal energy cost and minimal storage overhead.
Comments on “mabs multicast authentication based on batch signature”ambitlick
This document summarizes and critiques the MABS-DSA protocol proposed by Zhou et al. for multicast authentication using batch verification. While MABS-DSA was intended to increase efficiency and security over other implementations, the author finds through reexamination of the arithmetic that the algorithm is actually incorrect and batch signature verification would fail almost always, even when individual packets were properly signed by an honest sender. The key issue is a flaw in protocol correctness rather than the intended security improvements.
Energy-Efficient Protocol for Deterministic and Probabilistic Coverage In Sen...ambitlick
The document proposes a new probabilistic coverage protocol (PCP) for sensor networks that can employ different sensing models. PCP aims to address the costly task of designing and testing different coverage protocols for each sensing model. It works with common disk sensing models as well as probabilistic sensing models with minimal changes. Simulation results show that PCP outperforms other deterministic and probabilistic protocols in terms of number of activated sensors, total energy consumed, and network lifetime while being robust against failures and inaccuracies.
Energy efficient protocol for deterministicambitlick
The document describes a new probabilistic coverage protocol (PCP) for sensor networks that can employ both deterministic and probabilistic sensing models. PCP works by activating sensors to construct an approximate triangular lattice over the monitored area. It is more energy efficient than previous protocols by reducing the number of activated sensors needed for coverage. Simulation results show PCP outperforms other protocols in terms of energy consumption and network lifetime while maintaining coverage under various conditions.
Estimating Parameters of Multiple Heterogeneous Target Objects Using Composit...ambitlick
This article proposes a method for estimating parameters of multiple heterogeneous target objects (objects with different sizes and shapes) using networked binary sensors. The sensors are simple and only report detections, but no individual sensor location is known. The method introduces "composite sensor nodes" containing multiple sensors in a fixed arrangement. This provides relative location information to help distinguish individual target objects. As an example, the article considers a composite node with two sensors on a line segment. Measures from these nodes can identify target shapes and estimate object parameters like radius and side lengths. Numerical tests demonstrate networked composite sensors can estimate parameters of multiple target objects.
A Privacy-Preserving Location Monitoring System for Wireless Sensor Networksambitlick
This document proposes a privacy-preserving location monitoring system for wireless sensor networks. The system uses two in-network location anonymization algorithms:
1) A resource-aware algorithm that aims to minimize communication and computational costs by having each sensor node find a cloaked area containing at least k persons and report only aggregate location information.
2) A quality-aware algorithm that aims to maximize accuracy by iteratively refining cloaked areas reported by the resource-aware algorithm to minimize their size, while still maintaining k-anonymity.
The system collects anonymous aggregate location information to build a spatial histogram for estimating person distributions and answering queries about aggregate locations, while preserving individuals' location privacy against potential attacks from untrusted
This presentation was provided by Steph Pollock of The American Psychological Association’s Journals Program, and Damita Snow, of The American Society of Civil Engineers (ASCE), for the initial session of NISO's 2024 Training Series "DEIA in the Scholarly Landscape." Session One: 'Setting Expectations: a DEIA Primer,' was held June 6, 2024.
How to Fix the Import Error in the Odoo 17Celine George
An import error occurs when a program fails to import a module or library, disrupting its execution. In languages like Python, this issue arises when the specified module cannot be found or accessed, hindering the program's functionality. Resolving import errors is crucial for maintaining smooth software operation and uninterrupted development processes.
This presentation includes basic of PCOS their pathology and treatment and also Ayurveda correlation of PCOS and Ayurvedic line of treatment mentioned in classics.
Walmart Business+ and Spark Good for Nonprofits.pdfTechSoup
"Learn about all the ways Walmart supports nonprofit organizations.
You will hear from Liz Willett, the Head of Nonprofits, and hear about what Walmart is doing to help nonprofits, including Walmart Business and Spark Good. Walmart Business+ is a new offer for nonprofits that offers discounts and also streamlines nonprofits order and expense tracking, saving time and money.
The webinar may also give some examples on how nonprofits can best leverage Walmart Business+.
The event will cover the following::
Walmart Business + (https://business.walmart.com/plus) is a new shopping experience for nonprofits, schools, and local business customers that connects an exclusive online shopping experience to stores. Benefits include free delivery and shipping, a 'Spend Analytics” feature, special discounts, deals and tax-exempt shopping.
Special TechSoup offer for a free 180 days membership, and up to $150 in discounts on eligible orders.
Spark Good (walmart.com/sparkgood) is a charitable platform that enables nonprofits to receive donations directly from customers and associates.
Answers about how you can do more with Walmart!"
A review of the growth of the Israel Genealogy Research Association Database Collection for the last 12 months. Our collection is now passed the 3 million mark and still growing. See which archives have contributed the most. See the different types of records we have, and which years have had records added. You can also see what we have for the future.
A workshop hosted by the South African Journal of Science aimed at postgraduate students and early career researchers with little or no experience in writing and publishing journal articles.
Strategies for Effective Upskilling is a presentation by Chinwendu Peace in a Your Skill Boost Masterclass organisation by the Excellence Foundation for South Sudan on 08th and 09th June 2024 from 1 PM to 3 PM on each day.
LAND USE LAND COVER AND NDVI OF MIRZAPUR DISTRICT, UPRAHUL
This Dissertation explores the particular circumstances of Mirzapur, a region located in the
core of India. Mirzapur, with its varied terrains and abundant biodiversity, offers an optimal
environment for investigating the changes in vegetation cover dynamics. Our study utilizes
advanced technologies such as GIS (Geographic Information Systems) and Remote sensing to
analyze the transformations that have taken place over the course of a decade.
The complex relationship between human activities and the environment has been the focus
of extensive research and worry. As the global community grapples with swift urbanization,
population expansion, and economic progress, the effects on natural ecosystems are becoming
more evident. A crucial element of this impact is the alteration of vegetation cover, which plays a
significant role in maintaining the ecological equilibrium of our planet.Land serves as the foundation for all human activities and provides the necessary materials for
these activities. As the most crucial natural resource, its utilization by humans results in different
'Land uses,' which are determined by both human activities and the physical characteristics of the
land.
The utilization of land is impacted by human needs and environmental factors. In countries
like India, rapid population growth and the emphasis on extensive resource exploitation can lead
to significant land degradation, adversely affecting the region's land cover.
Therefore, human intervention has significantly influenced land use patterns over many
centuries, evolving its structure over time and space. In the present era, these changes have
accelerated due to factors such as agriculture and urbanization. Information regarding land use and
cover is essential for various planning and management tasks related to the Earth's surface,
providing crucial environmental data for scientific, resource management, policy purposes, and
diverse human activities.
Accurate understanding of land use and cover is imperative for the development planning
of any area. Consequently, a wide range of professionals, including earth system scientists, land
and water managers, and urban planners, are interested in obtaining data on land use and cover
changes, conversion trends, and other related patterns. The spatial dimensions of land use and
cover support policymakers and scientists in making well-informed decisions, as alterations in
these patterns indicate shifts in economic and social conditions. Monitoring such changes with the
help of Advanced technologies like Remote Sensing and Geographic Information Systems is
crucial for coordinated efforts across different administrative levels. Advanced technologies like
Remote Sensing and Geographic Information Systems
9
Changes in vegetation cover refer to variations in the distribution, composition, and overall
structure of plant communities across different temporal and spatial scales. These changes can
occur natural.
An efficient hybrid peer to-peersystemfordistributeddatasharing
1. 1158 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
An Efficient Hybrid Peer-to-Peer System
for Distributed Data Sharing
Min Yang, Student Member, IEEE, and Yuanyuan Yang, Fellow, IEEE
Abstract—Peer-to-peer overlay networks are widely used in distributed systems. Based on whether a regular topology is maintained
among peers, peer-to-peer networks can be divided into two categories: structured peer-to-peer networks in which peers are
connected by a regular topology, and unstructured peer-to-peer networks in which the topology is arbitrary. Structured peer-to-peer
networks usually can provide efficient and accurate services but need to spend a lot of effort in maintaining the regular topology. On the
other hand, unstructured peer-to-peer networks are extremely resilient to the frequent peer joining and leaving but this is usually
achieved at the expense of efficiency. The objective of this work is to design a hybrid peer-to-peer system for distributed data sharing
which combines the advantages of both types of peer-to-peer networks and minimizes their disadvantages. The proposed hybrid peer-
to-peer system is composed of two parts: the first part is a structured core network which forms the backbone of the hybrid system; the
second part is made of multiple unstructured peer-to-peer networks each of which is attached to a node in the core network. The core
structured network can narrow down the data lookup within a certain unstructured network accurately, while the unstructured networks
provide a low-cost mechanism for peers to join or leave the system freely. A data lookup operation first checks the local unstructured
network, and then, the structured network. This two-tier hierarchy can decouple the flexibility of the system from the efficiency of the
system. Our simulation results demonstrate that the hybrid peer-to-peer system can utilize both the efficiency of structured peer-to-
peer network and the flexibility of the unstructured peer-to-peer network and achieve a good balance between the two types of
networks.
Index Terms—Peer-to-peer systems, P2P, structured peer-to-peer, unstructured peer-to-peer, hybrid, overlay networks.
Ç
1 INTRODUCTION
L AST few years have witnessed a rapid development of
peer-to-peer networks. Research has shown that a large
fraction of traffic in the Internet is occupied by peer-to-peer
are group members and the peer-to-peer overlay network
is a multicast tree. It can also be used for distributed
computing which utilizes the idle resources in the network
applications [2]. A peer-to-peer (P2P for short) network is a for a huge computing task. Finally, it can be used to
logical overlay network on top of a physical network. Each provide communication anonymity in which the sender’s
peer corresponds to a node in the peer-to-peer network and identity is concealed.
resides in a node (host) in the physical network. All peers Based on whether a regular topology is maintained
are of equal roles. The links between peers are logical links, among peers, peer-to-peer networks can be divided into
each of which corresponds to a physical path in the physical two categories: structured peer-to-peer networks in which
network. The physical path is determined by a routing peers are connected by a regular topology, and unstructured
algorithm and composed of one or more physical links. peer-to-peer networks in which the network topology is
Logical links can be added to the peer-to-peer network arbitrary. Structured peer-to-peer networks build a distrib-
arbitrarily as long as a corresponding physical path can be uted hash table (DHT) on top of the overlay network. The
found, that is, the physical network is connected. hash table supports efficient data insertion and lookup.
The flexibility of the overlay topology and the decen- Given a key of the data item, the corresponding value of the
tralized control of the peer-to-peer network make it suitable data item can be inserted or found by transforming the key
for distributed applications. For example, it can be used for to a hash value by a hash function. The hash value is the
distributed data (file) sharing, where peers announce the index of the data item and all the hash values form the key
data (files) they have and exchange data (files) from each space. In DHT, the key space is divided among peers. Each
other through a loosely formed peer-to-peer network, or for peer is responsible for one partition of the key space. Peers
collaborative Web caching in which Web pages are cached are connected by an overlay network through which the
in collaborative peers to reduce network delay for URL requests of data insertion and lookup are delivered.
requests, or for application layer multicast in which peers Structured peer-to-peer networks can provide efficient and
accurate query service but need a lot of efforts to maintain
the DHT, which makes it vulnerable to frequent peer joining
. The authors are with the Department of Electrical and Computer
Engineering, Stony Brook University, Stony Brook, NY 11794. and leaving, also known as churn. Churn is a common
E-mail: {myang, yang}@ece.sunysb.edu. phenomenon in peer-to-peer networks. Measurement stu-
Manuscript received 14 July 2009; revised 20 Oct. 2009; accepted 29 Oct. dies of deployed peer-to-peer networks show a high rate of
2009; published online 17 Nov. 2009. churn [21], [22]. Unstructured peer-to-peer networks orga-
Recommended for acceptance by A. Zomaya. nize peers into an arbitrary network topology, and use
For information on obtaining reprints of this article, please send e-mail to:
tc@computer.org, and reference IEEECS Log Number TC-2009-07-0324. flooding or random walks to look up data items. Each peer
Digital Object Identifier no. 10.1109/TC.2009.175. receiving the flooding packets or random walk packets
0018-9340/10/$26.00 ß 2010 IEEE Published by the IEEE Computer Society
2. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1159
checks its own database for the data item queried. This Internet-like scales. Both peers and data are hashed to a
approach does not impose any constraint on the network virtual d-dimensional Cartesian coordinate space. The
topology. It can perform complex data lookup and support entire space is partitioned to distinct zones such that each
peer heterogeneity. Unstructured peer-to-peer networks are peer is in charge of one zone. Every peer maintains a
resilient to churn while they usually achieve this goal by routing table which holds the IP address of its neighbors in
sacrificing the data query efficiency and accuracy. the coordinate space. The data is stored in and retrieved
Hence, neither structured peer-to-peer networks nor from the peer that owns the zone covering the data. CAN
unstructured peer-to-peer networks can provide efficient, takes advantage of the ordering of the Cartesian coordinate
flexible, and robust service alone. The motivation of this space in the routing algorithm. Packets are forwarded along
paper is to combine the two types of peer-to-peer networks the straight line connecting the source and the destination in
and provide a hybrid solution which can offer efficiency the Cartesian coordinate space. When a new peer joins the
and flexibility at the same time. To achieve this goal, the system, some existing zone will be split into two zones one
solution should inherit the advantages of both types in such of which is assigned to the new peer, and all the related
a way that their disadvantages are minimized. peers need to update their neighbor lists. When a peer
In this paper, we propose a hybrid peer-to-peer system leaves the system, a neighboring peer will take over the
for distributed data sharing which combines the structured zone by running a takeover algorithm, and all the related
and unstructured peer-to-peer networks. In the proposed peers need to update their neighbor lists again.
hybrid system, a structured ring-based core network forms Chord [11] organizes the peers into a circle which is
the backbone of the system and multiple unstructured peer- called a chord ring, where each peer is assigned an ID. Peers
to-peer networks are attached to the backbone and com- are inserted into the ring in the order of their IDs. Each peer
municate with each other through the backbone. Data is has two neighbors: successor and predecessor. When a peer
generated and distributed among the peers. The core joins the system, it first finds the position to insert the new
structured network provides an accurate way to narrow peer. Then, the successor pointers of both the new peer and
down the queried data within a certain unstructured an existing peer must be changed. The correctness of Chord
network, while the unstructured networks provide a low- relies on the fact that each peer is aware of its successor. To
cost mechanism for peers to join or leave the system freely. guarantee this, each peer maintains a successor list of size r
The main contributions of this paper can be summarized which contains the peer’s first r successors. Each data item
as follows: also has an ID and is stored in a peer such that the ID of the
. Propose a hybrid peer-to-peer system for distributed data item is between the ID of the peer and its predecessor.
data sharing. It utilizes both the efficiency of the Packets are forwarded along the circle. In order to accelerate
structured peer-to-peer network and the flexibility of the search, each peer maintains a finger table, where each
the unstructured peer-to-peer network, and achieves finger points to a peer with a certain distance from the
a good balance between the efficiency and flexibility. current peer. Chord uses a “stabilization” protocol running
. Give a theoretic analysis on the system performance, in the background to update the successor pointers and
and derive quantitative results on the improvement finger tables. Compared to CAN, Chord is simpler as the
of the join latency and data lookup latency. key is hashed in a 1D space.
. Evaluate the proposed scheme through extensive Gnutella [13] is a decentralized unstructured peer-to-peer
simulations. Simulation results match the theoretic network. The network is formed by peers joining the
analysis and show that the proposed scheme achieves network following some loose rules. There is no constraint
the original design goal. on the network topology. To look up a data item, a peer
The rest of the paper is organized as follows: In Section 2, sends a flooding query request to all neighbors within a
we briefly describe some related work. We present the new certain radius. As Gnutella has no requirement on the
hybrid peer-to-peer system in Section 3 and analyze its network topology and data placement, it is extremely
performance in Section 4. We give some enhancements to resilient to peer joining and leaving the system frequently.
the hybrid peer-to-peer system in Section 5. In Section 6, we However, flooding is not scalable and consumes a lot of
present the simulation results and discuss the performance network bandwidth. Also, it is difficult to find a rare data
of the system. Section 7 gives the concluding remark and item as it has to flood the query request to most of the peers.
future work. There has been some work [20] on improving the efficiency
when looking up a rare data item.
BitTorrent [14] is a centralized unstructured peer-to-peer
2 RELATED WORK network for file sharing. A central server called tracker
Many peer-to-peer networks have been proposed for keeps track of all peers who have the file. Each file has a
different applications in the literature, see, for example, [1], corresponding torrent file stored in the tracker which
[2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]. In contains the information about the file, such as its length,
this paper, we focus on peer-to-peer networks for efficient name, and hashing information. When receiving a down-
distributed data (file) sharing among peers. load request, the tracker sends back a random list of peers
The Content Addressable Network (CAN) [10] was which are downloading the same file. When a peer has
proposed to provide a scalable indexing mechanism for file received the complete file, it should stay in the system for
sharing over a large network. As a distributed infrastruc- other peers to download at least one copy of the file from it.
ture, CAN provides hash-table-like functionality over Since BitTorrent uses a central server to store all the
3. 1160 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
information about the file and the peers downloading the
file, it suffers so called “single point of failure” problem
which means that if the central server fails, the entire system
is brought to a halt. Note that in some literatures, hybrid
peer-to-peer networks were used to refer to the centralized
peer-to-peer systems such as BitTorrent. In this paper,
“hybrid” is used to refer to the combination of structured
and unstructured peer-to-peer network topologies.
YAPPERS [15] combines both structured peer-to-peer
networks and unstructured peer-to-peer networks to
provide a scalable lookup service over an arbitrary
topology. Each peer has an immediate neighborhood (IN) Fig. 1. Overview of the proposed hybrid peer-to-peer system.
which contains all peers within h hops of the peer and an
extended neighborhood (EN) which contains all peers
their p ids. Each t-peer maintains two pointers which point
within 2h þ 1 hops of the peer. Both data keys and peers are
to its successor and predecessor, respectively. A finger table
hashed to different buckets or colors. Data is stored in the
is also used to accelerate the search. A stub network, called
peer in the same color. For a data lookup, a peer first checks
s-network, is a Gnutella-style unstructured peer-to-peer
the peers in the IN in the same color, then these nodes will
network. We call the peers in an s-network s-peers except
forward the request to the peers in their IN, and so on.
for the t-peer attached to this s-network. The topology of an
Finally, all the peers in the same color will be checked.
s-network is arbitrarily formed. Each s-network is attached
However, YAPPERS is designed for efficient partial lookup
to a t-peer and this t-peer belongs to both the t-network and
which only returns partial values of data. For a complete
the s-network. One thing to mention about the s-network is
lookup, YAPPERS still needs to flood the request to all
that the topology of an s-network is a tree instead of a mesh.
peers which are in the same color as the data. Compared to
We will describe the topology construction in more details
YAPPERS, the hybrid peer-to-peer system proposed in this
later in this section. Fig. 1 shows the overview of the
paper can conduct accurate complete lookups in a more
proposed hybrid peer-to-peer system.
efficient way.
We define a system parameter ps as the proportion of
In [16], the authors propose a hybrid peer-to-peer system
s-peers in the system. ps is a tunable system parameter
which treats rare and popular data items differently. Some
which has a great impact on the system performance. In
“ultrapeers” form a structured peer-to-peer network which
particular, when ps equals to 0, the system degenerates to a
is responsible for caching the rare data. Each ultrapeer has
ring-based structured peer-to-peer network; when ps
multiple attached leaf peers. Data lookup is first performed
equals to 1, the system becomes a Gnutella-style unstruc-
through the conventional flooding method. If not success-
tured peer-to-peer system; and when ps equals to 0.5, a half
ful, the query is reissued to an ultrapeer as a DHT data
of the peers in the system are t-peers while the other half of
lookup. It is somewhat similar to the hybrid system
the peers are s-peers. By tuning the system parameter, we
proposed in this paper. However, the main difference is
can observe how the system parameter affects the system
that in [16], the structured overlay was used as a
performance so that we can select an optimal value for ps to
supplement for unsuccessful flooded data lookup, while
maximize the performance in different applications. In
in this paper, the structured overlay is responsible for
Section 4, we will give a theoretical analysis on the impact
connecting all the unstructured networks and transmitting
of the system parameter on the system performance for
queries between them.
peer joining and data lookup. We will verify our analysis
by simulations in Section 6.
3 THE HYBRID PEER-TO-PEER SYSTEM The basic idea behind the hybrid peer-to-peer system is
that the t-network is used to provide efficient and accurate
In this section, we first give an overview of the new hybrid
service while the s-network is used to provide approximate
peer-to-peer system we propose, in which a tunable system
best-effort service to accommodate flexibility. Peers can join
parameter ps is defined to characterize the system perfor- either t-network or s-network directly. An s-network is
mance. Then, we describe how to maintain the peer-to-peer composed of peers that serve the data of some common
network topology when peers join and leave the system. properties. A data lookup is confined within an s-network if
Finally, we describe how to insert and look up data items in the queried data has the common properties served by the
the system. s-network. The lookup request is passed around the
s-network through flooding or random walk. Although
3.1 System Overview
flooding may generate a lot of network traffic, it can greatly
The new hybrid peer-to-peer system is composed of two simplify peer joining and leaving process, which makes the
parts: a core transit network and many stub networks, each system robust to churn. On the other hand, since the
of which is attached to a node in the core transit network. s-network contains only a small proportion of the total
The core transit network, called t-network, is a structured number of peers, flooding is confined within a small
peer-to-peer network which organizes peers into a ring number of peers. When the queried data is served by
similar to a chord ring. We call peers in the t-network another s-network, the data query request is first forwarded
t-peers. Each t-peer is assigned a peer ID (p id), which is a to the t-network through the t-peer in the s-network
positive integer. Peers are inserted to the ring in the order of generating the request. Then, in the t-network, the request
4. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1161
will be forwarded along the ring until it reaches the system are based on which network they belong to. Next, we
s-network serving the queried data. Finally, in the will discuss these two types of networks separately.
s-network, the request will be delivered to the s-peers by
flooding again. The t-network links all the s-networks 3.2.1 T-Peer Join/Leave
together and provides an efficient way to locate the desired The join operations of t-peers are similar to that in Chord
s-network. The stableness of the t-network is critical to the [11]. Each t-peer maintains two pointers that point to its
system performance because all the communications successor and predecessor along the ring, respectively.
between different s-networks are through the t-network. After the joining peer obtains the arbitrary t-peer, it sends a
As a structured peer-to-peer network, the t-network is
join request containing its p id to the t-peer. This join
vulnerable to churn mainly because the t-network needs to
request will be forwarded along the ring until it reaches the
recalculate the pointers in the finger tables whenever a
t-peer joins or leaves. However, the hybrid system can t-peer such that the p id of the joining peer is between the
effectively reduce the topology maintenance overhead p id of the t-peer and its successor. The t-peer will initiate a
caused by peer joining or leaving. This is because that, on join process and the joining peer is inserted between the
one hand, a large portion of peers join the s-networks t-peer and its successor.
directly without disturbing the t-network; and on the other The p id of a new peer is generated at the server. The
hand, an s-peer can be selected to substitute the leaving server has several options to generate the p id. One way is
t-peer in the same s-network, i.e., the selected s-peer will to generate the p id by hashing the IP address of the new
become a t-peer. In this case, the total number of t-peers is peer. Another way is to generate the p id based on the
unchanged. Therefore, there is no need to recalculate the location of the new peer. This can make the peers that are
pointers in the finger tables, and only a simple update is close to each other in the physical network also close to each
needed. other in the overlay network. Moreover, the server can
In this paper, we focus on applying the hybrid peer-to- generate a random p id for the new peer. However, the
peer system to distributed data sharing. A data item is p idgeneration process does not guarantee the uniqueness
represented by a (key, value) pair. A key is a label or name of of p id. In the case of a conflict, the t-peer initiating the join
the data, such as a file name, while a value is the content process will generate a new p id which lies in between the
associated with the key, such as a file. A peer uses operation
p id of itself and its successor. The new p id can be random
store (key, value) to insert the data item into the system and
or simply the midpoint for load balancing purpose.
operation lookup (key) to obtain the value of the data item.
Before performing the store or lookup operations, a peer After the join process completes, the segment of the id
hashes the data key to an integer d id which is in the same space represented by the successor has changed. The peers
range as p id. As mentioned earlier, s-peers are grouped in the successor’s corresponding s-network should transfer
into different s-networks such that each s-network serves part of its data load to the new peer, which is referred to as
the data of some common properties. In the hybrid system, load transfer. The peers check the data items it stores and
the p ids of t-peers divide the range of the d id into several transfer all the data items whose d id lie between the p id of
segments. Each s-network is responsible for the data whose the new peer and the predecessor of the new peer.
d ids lie in the same segment. Both store and lookup In a structured peer-to-peer network such as Chord,
operations try the local s-network first if the data item is when a peer leaves the system, all the pointers related to the
served by the local s-network; otherwise, they turn to the leaving peer must be updated. These pointers include the
t-network. Thus, such two-tier hierarchy structure can successor pointer in its predecessor, the predecessor pointer
provide efficient lookup in the top tier while maintaining in its successor, and the finger pointers in all the peers on
the flexibility in the bottom tier. the Chord ring. The hybrid peer-to-peer system can reduce
Here, we only consider exact-match data lookup. How- some update work as it does not need to recalculate new
ever, it is easy to extend it to support more complex data finger pointers. When a t-peer wants to leave the system, it
lookup such as regular-expression-based data lookup. The
transfers its role to one of the s-peers in its s-network. The
query message is first flooded within the same s-network. In
selected s-peer will change its role from an s-peer to a t-peer
the meanwhile, it is forwarded to other s-networks through
the t-network. and take over the neighbors and the pointers of the original
t-peer. By substituting the t-peer with an s-peer, the number
3.2 Peer Join/Leave and position of t-peers remain the same. Other t-peers only
Peer-to-peer networks are highly dynamic and autonomic need to substitute the leaving t-peer with the new t-peer in
systems in which peers can join or leave the systems at any the finger table.
time. Peers that want to join the system first contact a well- Sometimes, peers will leave the system without notice due
known server to obtain an arbitrary existing peer in the to peer crash. We call it abruptly leaving. To handle abruptly
system. The IP address of the server can be acquired by a leaving, peers send “HELLO” messages (also called heart-
DNS-like public service. Unstructured peer-to-peer net- beat messages) to their neighbors periodically. Each peer
works process join or leave requests in a more flexible way maintains a timer for each of its neighbors. The time is reset
than structured peer-to-peer networks largely because on receiving a “HELLO” message from the corresponding
structured peer-to-peer networks have to maintain the neighbor. Timeout indicates peer crash. The disconnected
network topology after peers join or leave. Since the hybrid s-peers will compete to replace the crashed t-peer by sending
peer-to-peer system contains both structured peer-to-peer messages to the server. The server will pick an s-peer to be the
network (t-network) and unstructured peer-to-peer network new t-peer. The selection can be random or choosing the peer
(s-network), the operations of peer joining or leaving the with the smallest IP address.
5. 1162 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
TABLE 1 adds the peer to its neighbor list. Then, it notifies the random
Join/Leave Algorithm for t-Peers peer to add itself to its neighbor list and the join operation is
completed. The p id of the s-peer is the same as its neighbor.
It depends on the type of applications that which
s-network an s-peer should be assigned to. In a system with
interest-based s-networks, as will be discussed in Section 5,
the assignment is determined by the interests of the s-peer. In
other systems, the assignment could be random to distribute
the load evenly. The server is responsible for assigning a
joining s-peer to some s-network with a smaller size.
In a Gnutella-style peer-to-peer network, the data lookup
is through flooding. The range of flooding is determined by
the search radius, that is, the time-to-live (TTL) value of the
packet. As the data is distributed around the network
randomly, the search radius is critical to the probability of
finding the desired data item. For the same topology and
the same peer that initiates the search, the longer the search
radius, the higher the probability of finding the desired
data item, but the longer the latency required. Note that if
we add some simple constraints on choosing the random
peer when a new peer joins, we may shorten the network
diameter, and thus, reduce the search radius without
sacrificing the success ratio of finding the desired data
item. Next, we will discuss how to add such constraints.
First, we restrict the random peer to be picked to only
t-peers. Thus, all s-peers in the same s-network are
connected with one t-peer. As a result, the diameter of an
s-network is at most two, and one data lookup can reach all
the peers within two hops. The topology of the s-network is
a star centered at a t-peer. Although the data lookup can
achieve short latency in such an s-network, there is a
notable disadvantage that the load is extremely unbalanced.
The t-peer maintains a long neighbor list while the s-peer
has a neighbor list with only one neighbor. Each data
lookup request has to be forwarded through the t-peer. In
order to alleviate this problem, we put another restriction
on the degree of peers. When the degree of a peer reaches a
threshold , it passes the join request to one of its neighbors
randomly. The join request will be passed until it arrives at
a peer whose degree is less than . In our simulation, we
use this scheme for s-peer joining. The new s-peer searches
from a t-peer along a random branch until it finds a peer
with a degree less than . This peer is called the connect
point (cp) of the new s-peer. Besides the neighbor list, each
s-peer maintains two pointers that store the address of the
t-peer of the s-network and its cp. The resulting topology of
an s-network is a tree. Here, we use a tree instead of a mesh
due to bandwidth efficiency consideration. A major draw-
back of an unstructured peer-to-peer network is that the
flooded query messages occupy a lot of network band-
width. In a mesh network, it is very likely that a peer
receives the same query message multiple times from
different neighbors. On the other hand, a tree structure
Table 1 lists the pseudocode for join and leave operations. guarantees that each peer receives the query message
exactly once.
3.2.2 S-Peer Join/Leave When an s-peer leaves the system, it should notify all its
Each s-peer belongs to an s-network and maintains a list of its neighbors about the leaving. The neighbors then delete it
neighbors. A neighbor can be either an s-peer or a t-peer. from their neighbor lists. The neighbor whose cp is the
After an s-peer acquires the IP address of the random peer, it leaving peer should rejoin the s-network by sending a join
6. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1163
request to the t-peer again. The leaving s-peer should also
choose a neighbor to transfer the load to.
Again, we need to handle the abruptly leaving when
peers crash without notice. To detect and recover from
these errors, we still use the periodic “HELLO” messages.
Each s-peer periodically broadcasts “HELLO” messages to
all its neighbors. “HELLO” messages will reset the timer
and timeout indicates peer crash. As the topology of an
s-network is a tree with a maximum degree of , it is
Fig. 2. Concurrent join/leave operation for t-peers.
vulnerable to abruptly leaving. When an s-peer crashes, all
the peers in the subtree which is rooted at one of the
should be inserted between x and its successor. Thus, the
children of the crash peer will be disconnected from the
t-peer will pass the join request to its new successor x.
system until the error is detected by the timeout. Short-
However, the join operation of x is not completed, and the
ening the period of the timer can reduce the time the peers
successor and predecessor pointers in x are not set correctly.
are disconnected. However, the drawback is that it
Therefore, the join request of y will not be handled correctly.
generates more network traffic from “HELLO” messages.
To overcome this problem, we let each peer respond a data We adopt the idea of the concurrency handling in the
query message with an acknowledgment message. There is database system for the concurrent joins or leaves in the
a timer for acknowledgment messages. Resetting the timer t-network. The join and leave requests are sequentialized
for acknowledgment messages also resets the “HELLO” such that the next request is not processed until the
timer so that the scheduled “HELLO” message is canceled previous request finishes. For a join request, it follows a
to save network bandwidth. Now a peer can detect a peer join triangle, as shown at the left of Fig. 2. When peer pre
crash by either receiving no “HELLO” message before receives the join request of the new peer, it sets a mutex
timeout or receiving no acknowledgment message before variable joining that indicates some peer is being inserted
timeout. There is another suppress timer. The peers send between peer pre and peer suc. Now peer pre will not
acknowledgment messages only when the suppress timer accept any leave requests including that from itself. If a new
is timeout and a data query message is received. This timer join request comes before the previous request is com-
is used to suppress the excessive acknowledgment mes- pleted, peer pre will insert the new request to a queue and
sages when the data query is frequent. In summary, the process the request queue after the previous join request
timers for the acknowledgment messages are used to finishes. Then, peer pre sends a packet including the
ensure that the peers respond to link failure or peer crash address of peer suc to the new peer. The new peer sets its
more quickly when there are data queries. The timer for successor and predecessor pointers to suc and pre, respec-
“HELLO” messages is used to ensure the peers maintain a tively, and sends another packet to peer suc. After receiving
consistent view of the overlay topology in the absence of the packet, peer suc updates its predecessor pointer and
data query messages. sends a packet back to peer pre. When peer pre receives this
Another effect of peer crash is that the load on the peer is packet, it sets its successor pointer to the new peer and
lost. As will be seen in the simulation results, we examined continues to process the next join request in the queue. If the
this scenario and observed its impact on the data lookup queue is empty, it resets the mutex variable.
failure ratio. When a peer leaves the system, it follows the leave
triangle which is shown on the right of Fig. 2. When a peer
3.3 Concurrent Join/Leave
is leaving, it also sets a mutex variable leaving. Now the
Peer-to-peer networks are highly dynamic systems since peer will not accept any new join request (if the join request
peers usually are end hosts that are in charge of different queue is not empty, the peer should process the join request
individuals or groups. Concurrent joins and leaves are very first) and leaving request. Then, the leaving peer sends a
common and can greatly degrade the performance if not packet to peer pre including the address of peer suc. Peer
handled carefully. suc sets its successor pointer to peer suc and sends a packet
The concurrency handling in the s-network is simpler to peer suc including the address of the leaving peer. After
than that in the t-network. When an existing peer receives receiving the packet, peer suc will check whether the
two join requests, it follows the First Come First Serve leaving peer included in the packet is what its predecessor
(FCFS) rule, that is, the second join request will be passed to pointer is pointing to. Only if they are the same peer, will
the next neighbor if the degree of the peer reaches the limit the peer suc set its predecessor pointer to peer pre and send
after receiving the first join request. When two or more a packet to the leaving peer to notify the completion of the
s-peers leave the system simultaneously, no special action leaving operation.
is needed because the disconnected parts will rejoin the
s-network, as described in the previous section. 3.4 Data Insertion/Lookup
For the t-network, concurrency handling is much more Data is generated and inserted to the system by peers. As
complicated as it involves three t-peers and the topology mentioned earlier, each s-network is responsible for a
constraint is strict. The concurrent joins or leaves may lead range of ID space. The peer generating the data item first
to an incorrect topology or break the t-network apart if not hashes the key into this space. If the d id lies in the range
handled carefully. For example, suppose a t-peer is inserting of the current s-network, the data item is inserted to its
a new peer, say, x, between itself and its successor. After database and the data insertion is completed. If the d id
setting its successor pointer to the new peer, it receives does not lie in the range, the data item is sent to the
another join request indicating that another new peer, say, y, t-peer of the current s-network. Then, it is forwarded
7. 1164 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
along the t-network until it reaches the t-peer in charge of
the ID range covering d id. Then, the data item is inserted
into the database of the t-peer.
Note that although this data placement scheme is simple
and easy to implement, the data load between different
peers may be imbalanced. Each t-peer corresponds to an
s-network. The data generated in all other s-networks except
the one a t-peer is in will always be stored in the t-peer;
therefore, the data load in t-peers is much heavier than that
in s-peers. To alleviate the load imbalance, we spread the
data load to the neighbors of the t-peer, i.e., the s-peers
connected to the t-peer. When a t-peer receives a data
insertion request, it picks a random s-peer from its directly
connected s-peers and itself, and then sends a data insertion
request to the chosen peer. The random peer will perform
the same random data load spreading until the data item is
finally inserted to the system. In the performance evaluation
section, we will implement both the original data placement
scheme and the improved one and study their impacts on
the probability density function of the number of data items
per peer.
When a peer looks up a specified data item, it first
obtains the d id of the data item by hashing the data key. If
the d id lies in the current s-network, the peer floods lookup
packets around the s-network and sets a timer for it. The
timer will be reset if the peer receives the data item or
expire, which indicates that the data item is not found.
Thepeer may choose to increase the TTL value and the
expiration duration of the timer and reflood the lookup
packets. If the d id does not lie in the current s-network, the
peer sends a lookup request to the t-peer and also sets a
timer for it. Similar to data insertion, the data lookup
request is forwarded along the t-network until it arrives at a
proper t-peer which will then flood data lookup packets
around its s-network. Each peer receiving the lookup Fig. 3. (a) The average join latency under different s. (b) The average
request will check its database for data item d id. If the lookup latency under different s.
data item is found in its database, the peer will stop
flooding and send the data item to the peer requesting the always a t-peer, the number of hops passed by the join
data item directly. request is always 1; thus, the join latency is a constant. If we
impose the degree constraint when an s-peer joins, the
4 PERFORMANCE ANALYSIS number of hops passed by a join request is determined by
the average number of s-peers in an s-network. Recall that
In this section, we analyze the performance of the proposed
hybrid peer-to-peer system and compare it with the an s-network has a tree topology where the t-peer is the root
structured peer-to-peer networks and the unstructured of the tree and the degree of the tree is constrained by . The
peer-to-peer networks. In particular, we vary the system join request is passed from the root to a leaf, which equals
parameter ps , the proportion of the s-peers in the system, the average height of the tree. As mentioned in the previous
and observe how it affects the system performance. section, the s-peers are distributed into the s-networks
evenly by the server. Therefore, the average number of
4.1 Performance of the Join Operation
s-peers in one s-network is ps =ð1 À ps Þ. Thus, the average
The efficiency of the join operation can be measured by the number of hops passed by the join request can be calculated
join latency, which is defined by the time difference
by log ðps =ð1 À ps ÞÞ.
between the joining peer sending the join request to the
We can see that the join latency of t-peers is mono-
server and the joining peer being inserted into the system.
To give a quantitative analysis on the join latency, we use tonically decreasing when the system parameter ps in-
the number of hops the join request passes to estimate the creases, while the join latency of s-peers is monotonically
join latency. Assume that the total number of peers in the increasing when the system parameter ps increases (here,
system is N. Then, the number of s-peers is ps à N and the we constrain the degree by ). Therefore, the average join
number of t-peers is ð1 À ps Þ Ã N. For t-peers, the join latency of peers is
request is passed along the ring. Thus, the average number
ð1 À ps Þ Ã logðð1 À ps Þ Ã N=2Þ þ ps à log ðps =ð1 À ps ÞÞ: ð1Þ
of hops passed by the join request is ð1 À ps Þ Ã N=2. Finger
tables can reduce this number to logðð1 À ps Þ Ã N=2Þ. For Fig. 3a plots the curve of the average join latency under
s-peers, there are two cases. If the selected random peer is different values when the system parameter ps increases
8. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1165
from 0 to 1. We can see that the hybrid peer-to-peer system We use the midpoint of the two numbers to approximate
can achieve a much shorter average join latency than the the average number, which is
structure peer-to-peer networks and the unstructured peer-
to-peer networks. For example, in the figure, the shortest ps =ð1 À ps Þ À ðttlþ2 À ttlþ1 À þ 1
join latency is achieved when ps is around 0.7 for ¼ 2. We þ 2þttl=2 À 2 À ð À 1Þ Ã ttl=2Þ=ð2 Ã ð À 1Þ2 Þ
can observe that given system parameter ps , the larger the ð2Þ
% ps =ð1 À ps Þ À ðttlþ1 Ã ð À 1Þ þ 2þttl=2
degree constraint , the shorter the join latency. This is
because that a larger can reduce the average height of the À ð À 1Þ Ã ttl=2Þ=ð2 Ã ð À 1Þ2 Þ:
tree of an s-network, and thus, reduce the number of hops From (2), we can draw the following conclusion: the lookup
passed by the join request. failure ratio increases if ps increases while it decreases when
ttl increases.
4.2 Performance of the Data Lookup Operation
We use the number of hops the lookup request packet
We use the data lookup latency and the data lookup failure passes to estimate the data lookup latency. If the s-network
ratio to evaluate the data lookup performance. Data lookup is constructed without the degree constraint, the average
latency is defined as the time difference between the time data lookup latency is
when the peer issues the data lookup request and the time
when the peer receives the data. If the peer does not receive ð1 À ps ÞN
p à 2 þ ð1 À pÞ Ã 2 þ log :
the requested data, the timer will expire and this data 2
lookup operation is considered as a data lookup failure.
In the above expression, the first term p à 2 represents the
When calculating lookup latency, we only consider the cases
average number of hops the flooding packets pass if the
of successful lookups. The failed lookups are evaluated by
data item is stored in the same s-network, and the second
the data lookup failure ratio which is defined by the number
term, ð1 À pÞ Ã ð2 þ logðð1Àps ÞN ÞÞ represents the average
2
of failed data lookups divided by the total number of data
number of hops the lookup request packet passes if the
lookups.
data item is stored in a different s-network.
Before we analyze the performance of the data lookup
Similarly, if the s-network is constructed with the degree
operation, we define a variable p which is the probability
constraint , the average data lookup latency is
that the requested data item is stored in the same s-network
as the s-peer issuing the data lookup. As mentioned earlier, p à ttl þ ð1 À pÞ
the average number of s-peers in an s-network is ps =ð1 À ps Þ.
1 ps ð1 À ps ÞN
Thus, we have p ¼ ps =ðN Ã ð1 À ps ÞÞ. For a structured peer- Ã max 0; log þ ttl þ log ;
2 1 À ps 2
to-peer network, each data item is stored at a specific peer.
ps
The data lookup request is forwarded to that peer so the where maxf0; 1 log ð1Àps Þg represents the average number of
2
lookup failure ratio is 0. For an unstructured peer-to-peer hops the lookup request packet passes from the s-peer
network like Gnutella, the location of a data item is starting the lookup to the t-peer, and ttl represents the
arbitrary, and it uses flooding to perform a best-effort number of hops the flooding packets pass in the s-network
search. The range of the flooding is confined by TTL value which contains the desired data item.
(ttl) in the flooding packets. To calculate the lookup failure Fig. 3b shows the curve of the average data lookup
latency under different values when the system parameter
ratio, we need to find the number of peers that are out of the
ps increases from 0 to 1. We can see that when ps is less than
search range of the flooding. In the case that the s-network is
0.5, i.e., the size of an s-network is less than or equal to 1, the
constructed without the degree constraint, the lookup
data lookup latencies are longer and are the same for
failure ratio can be zero if the ttl is set to 2. Now we
different . This is because most of the data lookup requests
consider the case that the s-network is constructed with the have to pass through the t-network. When ps is larger, the
degree constraint . In this case, the number of peers that is size of the s-network increases, more data lookup requests
out of the search range of the flooding is determined by the can be handled in the local s-network, and the data lookup
location of the peer initiating the flooding and the ttl. If a latency is shortened, though at the expense of a higher
t-peer initiates the flooding, the number of peers that are out lookup failure ratio. Given system parameter ps , the larger
of the search range is the degree constraint , the shorter the data lookup latency.
This can also be explained by the fact that a larger will
ps =ð1 À ps Þ À ð1 þ þ 2 þ Á Á Á þ ttl Þ have a smaller average height of the tree in an s-network.
¼ ps =ð1 À ps Þ À ðttlþ1 À 1Þ=ð À 1Þ: To summarize, both the join operation and data lookup
operation greatly depend on ps . The join latency is
If a leaf s-peer initiates the flooding, the number of peers minimized when ps ranges from 0.7 to 0.8. This indicates
that are out of the search range is that the hybrid system outperforms the two “pure” systems
in terms of join latency. For the data lookup, the selection of
ps is a tradeoff between data lookup latency and lookup
X
ttl=2þ1
ps =ð1 À ps Þ À ði À 1Þ=ð À 1Þ failure ratio. In practice, the selection of ps may have other
i¼2 factors to consider. For example, if the s-networks are
¼ ps =ð1 À ps Þ À ð2þttl=2 À 2 À ð À 1Þ Ã ttl=2Þ=ð À 1Þ2 : interest-based, as discussed in the next section, the data
lookup latency largely depends on the TTL instead of ps as
9. 1166 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
most data lookup can be conducted within the same One way to solve this mismatch problem is to cluster the
s-network. Different applications can adjust ps to meet their peers such that the peers in the same cluster are close to
respective requirements. each other with respect to the latency. We can adopt the
idea of binning scheme introduced in [17] to construct a
topology-aware overlay network. In the scheme, the well-
5 ENHANCEMENTS known server is responsible for choosing some peers as
In this section, we discuss some enhancements which can landmarks. Each new peer receives the list of landmarks
make the proposed hybrid peer-to-peer system more from the server as a response of its join request. The new
flexible and improve the system performance when applied peer then sends probe messages to the landmarks to learn
to different applications. the distances between them and itself. The landmark peers
are listed in an ascending order of distances. The ordered
5.1 Supporting Link Heterogeneity
list acts as the coordinate of the new peer in the system. The
Link heterogeneity is a common phenomenon in networks. coordinate is sent to the server, and then the server assigns
In peer-to-peer networks, link heterogeneity refers to the the new peer a cluster ID based on its coordinate. Peers with
fact that peers have different access link capability. In the same coordinates form a cluster. Intuitively, peers in the
general, access links include dial-up connections, ADSL, same cluster are close to each other. The server assigns
and high-speed cable. The ratio of the link capacity of the peers in the same cluster to the same s-network. In the case
fastest link to the slowest link can be more than 1,000. that the number of s-networks is more than the number of
Due to the imbalance of link capacity, if a peer with high clusters, the peers in the same cluster may be assigned to
link capacity is receiving messages from a peer with low different s-networks. The peers are assigned to several
link capacity, its download speed is upper bounded by the s-networks in a round robin manner such that the numbers
download speed of the low link capacity peer. In other of s-peers of different s-networks are balanced.
words, the bandwidth of the high link capacity peer is In addition, every two landmark peers should not be too
wasted. To maximize the usage of all the link capacities, we close to each other. A new peer cannot be a landmark if its
connect the peer with higher link capacity in the system to coordinate is the same as one of the existing landmark
multiple peers with lower link capacities so that the fast peers. A landmark peer is removed from the landmark list if
peer can download or upload data to the multiple slow it has the same coordinate as another landmark peer.
peers simultaneously.
As mentioned earlier, the t-peers carry more traffic than 5.3 Interest-Based s-Networks
the s-peers. Even after we add some restriction to the In previous sections, we assigned the new s-peer randomly
connect point selection, the load imbalance cannot be to existing s-networks. However, the assignment may be
removed completely. On the contrary, we can make use of determined by applications. For example, the s-network
this imbalance for link heterogeneity consideration. Since may be composed of peers that are interested in the same
t-peers are connected to more other peers than s-peers on type of data. The data generated in one s-network is looked
average, we assign peers with higher link capacities as up mostly by a peer in the same s-network. In this case, a
t-peers while peers with lower link capacities as s-peers. In new s-peer can indicate its interest when sending the join
practice, each peer attaches its link capacity value to the request to the server. The server will acknowledge with a
packet sent to the server. Based on the value, the server t-peer whose s-network matches the interest, and the new
decides whether the peer is a t-peer or an s-peer. s-peer will join that s-network. With the application-specific
The idea of link heterogeneity can be applied to the assignment, s-peers are more likely to generate data stored
s-network construction as well. When an s-peer joins an in their own database and look up data in their own
s-network, the connect point first checks its link usage s-network. Hence, the latencies of both data insertion and
which is defined as the ratio of the degree to the link lookup can be reduced. This way, most of traffic related to
capacity of the peer. If the link usage is low enough, the data operations is confined within one s-network and the
connect point will establish a link between itself and the burden of the t-network is alleviated.
joining s-peer. If the link usage is high, the connect point Interest-based s-network is also useful for partial/key-
will notify the joining s-peer and choose a random branch word search as well as exact search. The partial search first
for it, as discussed in Section 3.2.2. indicates a field of interest. Then, the partial search is
conducted in the corresponding s-network similar to that in
5.2 Supporting Topology Awareness other unstructured peer-to-peer networks.
Since overlay networks are logical networks on top of
physical networks, the overlay links are logical links. Each 5.4 Bypass Link
logic link is composed of one or more physical links. The As discussed in the previous section, an s-network can be
overlay links are added arbitrarily, as needed. As a result, composed of s-peers with the same interest. However, in
the topology of the overlay network may be different from some applications, an s-peer may be interested in several
the topology of the physical network. Two nodes which are types of data at the same time. In the hybrid peer-to-peer
close to each other in the overlay network may be far away system, an s-peer can only belong to one s-network at a
in the physical network. Such topology mismatch may time. Therefore, we add random bypass links between
increase the link stress and degrade the performance, where different s-networks to accelerate the data operations.
link stress is defined as the number of copies of a message Another reason to use bypass links is to alleviate the
transmitted over a certain physical link. burden of the t-network. Since the t-network is responsible
10. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1167
Fig. 4. (a)-(f) Probability density functions of the number of data items per peer under different ps values for the two data placement schemes.
for transmitting join or leave requests of t-peers and implemented the system on the NS2 simulator. The network
performing all the data operations among s-networks, the topologies used in the simulations are random transit-stub
t-network links bear much heavier load than that in network topologies generated by GT-ITM software [24]. Each
s-networks. With bypass links, part of the data operations network topology is composed of 1,000 nodes, and each node
can be diverted from the t-network to the bypass links. is assigned to be either an s-peer or a t-peer randomly. The
Bypass links can be added according to the following ratio between the number of s-peers and the total number of
three rules: first, a bypass link cannot be added to a peer peers is determined by the system parameter ps . Note that
unless the degree of the peer is less than the threshold ; when ps ¼ 0, the hybrid system degenerates to a ring-based
second, if peer a generates and inserts a data item in peer b, structured peer-to-peer network. When ps ¼ 1, the hybrid
and peer a and peer b are in different s-networks, a bypass
system becomes a Gnutella-style unstructured peer-to-peer
link should be added between peer a and peer b; third, if
peer a looks up and finds a data item in peer b, and peer a system. In the simulations, we let ps increase from 0 to 1 so
and peer b are in different s-networks, a bypass link should that we can compare the hybrid system with both the
be added between peer a and peer b. structured peer-to-peer network and the unstructured peer-
Peers maintain bypass links in a loose way. Each bypass to-peer network. We set the peers with heterogeneous link
link is attached with a timer. When the timer expires, the capacities such that 1=3 of the peers have the highest link
bypass link is deleted. Transmitting a packet through the capacities, 1=3 of them have the lowest link capacities, and
bypass link will refresh the attached timer. When a peer 1=3 of them have the medium link capacities. The highest link
leaves the system, the bypass links it maintains are lost. capacity is 10 times of the lowest link capacity. The landmarks
are predetermined so that they are uniformly distributed
5.5 BitTorrent-Style s-Network around the network. is equal to three in the simulations. We
Although the hybrid peer-to-peer system uses the Gnutella will first compare the data distribution under two different
network as the prototype of the s-network, it can also easily data placement schemes discussed in Section 3.4, and then,
deploy the BitTorrent-style s-network. In a BitTorrent-style compare the data lookup failure ratio and data lookup
s-network, the t-peer acts as the tracker. When a data item is efficiency by tuning the system parameter ps .
inserted to the database of an s-peer, it is reported to the
t-peer. The t-peer maintains all the information of data 6.1 Data Distribution
items stored in its s-network. A data lookup request is sent As discussed in Section 3.4, when a data item is generated
to the t-peer directly. T-peer responds to the data lookup
in an s-network and stored in another s-network, it will be
with the peer which has the data item. Then, the data item is
forwarded along the t-network until it arrives at a proper
delivered between the two peers directly. In a BitTorrent-
style s-network, the t-peer plays a more important role than t-peer. There are two options for the t-peer: insert the data
that in Gnutella-style s-network and no flooding is needed. item to its own database or insert it to the database of a
random neighbor. The first option will cause imbalanced
data load among peers while the second option can greatly
6 SIMULATION RESULTS alleviate the imbalance.
In this section, we evaluate the performance of the proposed Fig. 4 shows the different probability density functions
hybrid peer-to-peer system through simulations. We have of the number of data items per peer under the two
11. 1168 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
schemes. In order to display the curves in the center of the
figure, the y-axis starts from À0:01 and both x-axis and
y-axis are rescaled. Figs. 4a, 4b, and 4c are the probability
density functions for the first data placement scheme when
ps is equal to 0, 0.4, and 0.9, respectively. We can see that
when ps ¼ 0, more than 50 percent of the peers store less
than 10 data items, and the rest of the peers store data items
ranging from 10 to 80. When ps increases to 0.4, the peers
without any data occupy 38 percent of the total peers, and
the rest of the peers store data items ranging from 0 to 120.
When ps ¼ 0:9, the proportion of peers without any data
item is 85 percent, and the highest number of data item per
peer is more than 500. The peers are evenly distributed
along the x-axis. We can observe that with the increase of
ps , the data items are more likely to be stored in t-peers.
The data items generated by an s-peer are finally inserted
in some t-peer after traveling along the t-network. Most of
the s-peers have no data items stored. Figs. 4d, 4e, and 4f
show the probability density functions for the second data
placement scheme when ps is equal to 0, 0.4, and 0.9,
respectively. The number of peers without any data item
drops dramatically. In Fig. 4f, the number of peers with no
data item is only around 12 percent compared to 85 percent
in Fig. 4e, and 50 percent of the peers store less than 20 data
items. We can draw the following conclusion from the
figures: when ps is small, the two schemes can distribute
the data items evenly among the peers; when ps is large, the
first scheme stores most of the data items in a few peers
while the second scheme can balance the data distribution
so that each peer stores a similar amount of data items.
Fig. 5. (a) Lookup failure ratio under different TTL values. (b) Lookup
failure ratio when peers crash under different ps values.
6.2 Lookup Failure Ratio
Lookup failure ratio is defined as the the number of failed
as structured peer-to-peer networks. In the meanwhile, we
data lookups divided by the total number of data lookups.
can minimize the average join latency by setting ps to
Here, we assume that the data are inserted to the system
around 0.7 (assuming ¼ 2), as demonstrated in Section 4.
before it is looked up by some peer.
However, the lookup failure ratio will increase accordingly.
The lookup failure ratio is affected by the TTL value, the
We can increase the TTL value to lower the lookup failure
diameter of the network, and the position of the peer
ratio to some acceptable level. We will discuss this tradeoff
initiating the data lookup. Here, we compare the lookup
later and deduce the optimal ps value.
failure ratio of the hybrid peer-to-peer system by tuning We now consider the impact of peer crash on the
both the TTL and ps . lookup failure ratio. Peer crash means that the peer has no
Fig. 5a shows the lookup failure ratios under different time to transfer the data load to one of its neighbors. In the
TTL values when ps increases from 0 to 1. As illustrated in
simulation, the peers are chosen randomly to leave the
the figure, structured peer-to-peer networks achieve zero
system without transferring its data load. Fig. 5b shows
lookup failure ratio, while unstructured peer-to-peer net-
the lookup failure ratio when different proportions of
works have a greater than zero lookup failure ratio
peers leave the system. We can see that the lookup failure
depending on the TTL value. When ps 0:5, the lookup
failure ratio of the hybrid system is around 0 for all TTL ratio increases linearly to the number of peers leaving the
values. Since the number of s-peers is less than the number system. As more peers are leaving the system, the data
of t-peers, an s-network has less than one s-peer on the stored in the peers is no longer available. With the increase
average. Thus, the flooding packets from a t-peer can reach of ps , the lookup failure ratio remains at the same level.
all the s-peers within one hop on the average. When ps This again shows that the improved data placement
increases, the lookup failure ratio increases exponentially. scheme can distribute the data load among peers evenly
When TTL ¼ 1 and ps ¼ 0:9, the lookup failure ratio is regardless of the value of ps .
increased to 18 percent. From the figure, we can also Generally speaking, higher ps value causes higher
observe that increasing TTL value can reduce the lookup lookup failure ratio. Increasing TTL value can alleviate
failure ratio dramatically. When ps ¼ 0:9, the lookup failure the situation greatly. When peers crash, the increase of the
ratio is four percent if TTL ¼ 4, compared to 14 percent if lookup failure ratio is proportional to the number of
TTL ¼ 2 and 18 percent if TTL ¼ 1. If we set ps to less than crashed peers and changing ps has no impact on the
0.5, the lookup failure ratio is almost zero which is the same lookup failure ratio.
12. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1169
TABLE 2
Total connum Under Different ps Values
which results in a larger search radius and lookup latency.
This latency increase partially offsets the latency decrease so
that the slope of the curve is more gradual when ps is large.
When link heterogeneity is considered, the average
lookup latency is further decreased. When ps is between
0.4 and 0.8, this latency decrease is more visible. The lookup
latency decreases by about 20 percent when ps ¼ 0:7.
Fig. 6b shows the average lookup latency with and
without topology awareness consideration under different
ps values. We consider two possible numbers of the
landmarks: 8 and 12. From the figure, we can see the lookup
latencies are the same when ps ¼ 0. When ps increases, the
latency of topology awareness scheme decreases much faster
than the basic scheme. The more the landmarks, the less the
Fig. 6. (a) and (b) Average lookup latency under different ps values. lookup latency. When ps approaches 0.9, the three curves
tend to merge. This indicates that the topology awareness
6.3 Lookup Efficiency has little impact when the number of the s-networks is
In an ideal system, a high lookup success ratio should not large, because in this case, the peers close to each other are
sacrifice the lookup efficiency. We evaluate the lookup assigned to the same s-network with high probability.
efficiency in terms of time and space. For the time, the It can be seen that extending the basic scheme to support
lookup efficiency is simply the lookup latency defined link heterogeneity or topology awareness can shorten the
earlier. For the space, it is the bandwidth involved in the lookup latency, especially when ps is around 0.7 for link
lookup. We approximate the bandwidth by the number of heterogeneity or around 0.3 for topology awareness. It reveals
peers a lookup may contact, denoted as connum. The less that the hybrid system is more sensitive to the extensions.
connum is, the less the bandwidth is needed. Here connum Table 2 shows the total number of peers the data lookups
is the number of peers all the data lookup requests contact contact under different TTL values. Generally speaking,
during the simulation. structured peer-to-peer networks contact more peers to find
Fig. 6a shows the average lookup latency with and a data item than unstructured peer-to-peer networks. This
without link heterogeneity consideration under different does not mean that unstructured peer-to-peer networks are
ps values. As the figure shows, structured peer-to-peer more efficient than structured peer-to-peer networks be-
networks have larger lookup latency than unstructured cause the lookup failure ratios of unstructured peer-to-peer
peer-to-peer networks while a hybrid system has a lookup networks are much greater than that of structured peer-to-
latency in between. While ps increases, the average lookup peer networks. connum decreases linearly with the increase
latency decreases more quickly when ps is small. This is due of ps . When ps ¼ 0:9, we can see that connum is only about
to the fact that the number of t-peers decreases when ps 10 percent of that of the structured peer-to-peer networks.
increases. A typical data lookup is composed of three steps: This is also due to the three steps of a data lookup. When ps
an s-peer sends data lookup request to the t-peer; the is small, most of the peers are t-peers. A data lookup will
lookup request is forwarded along the t-network until it pass a half of the t-network on the average. When ps is large,
reaches a proper t-peer; the lookup request is flooded the number of t-peers involved in the second step decreases.
around the s-network. As both the first and third steps are Although the number of s-peers involved in the first and
confined by the TTL value in the lookup packet, the number third steps increases, the increase is very slow as the s-peers
of hops a lookup experiences is determined by the second are evenly distributed in different s-networks.
step, which is proportional to the total number of t-peers. We can also observe that the TTL value has no impact on
When ps is large, the average size of s-networks is large connum when ps is small. However, when ps is large, the
13. 1170 IEEE TRANSACTIONS ON COMPUTERS, VOL. 59, NO. 9, SEPTEMBER 2010
larger the TTL value is, the larger connum is. This is because REFERENCES
that a larger TTL value enlarges the range of the flooding [1] E.K. Lua, J. Crowcroft, M. Pias, R. Sharma, and S. Lim, “A
packet, thus increasing the number of peers involved. Survey and Comparison of Peer-to-Peer Overlay Network
Therefore, we have the following conclusion: the average Schemes,” IEEE Comm. Surveys and Tutorials, vol. 7, no. 2,
pp. 72-93, Mar. 2005.
number of peers a lookup may contact drops linearly as ps [2] S. Sen and J. Wang, “Analyzing Peer-to-Peer Traffic Across Large
increases. Also, increasing TTL causes slightly greater Networks,” Proc. Internet Measurement Workshop, Nov. 2002.
connum only when ps is larger than 0.5. [3] Y. Chu, S. Rao, and H. Zhang, “A Case for End System Multicast,”
Proc. ACM Sigmetrics ’00, pp. 1-12, June 2000.
From the above discussions, we can see that increasing ps
[4] E. Brosh and Y. Shavitt, “Approximation and Heuristic Algo-
helps reduce both lookup latency and connum, which is rithms for Minimum Delay Application-Layer Multicast Trees,”
consistent with the theoretical analysis in Section 4. It means Proc. IEEE INFOCOM ’04, Mar. 2004.
that the hybrid system can achieve less lookup latency and [5] M. Freedman and R. Morris, “Tarzan: A Peer-to-Peer Anonymiz-
ing Network Layer,” Proc. Ninth ACM Conf. Computer and Comm.
connum than structured peer-to-peer networks. However, a
Security, Nov. 2002.
higher ps value also causes the lookup failure ratio to [6] S. Banerjee, C. Kommareddy, K. Kar, B. Bhattacharjee, and S.
increase dramatically. Based on the results in Section 4 and Khuller, “Construction of an Efficient Overlay Multicast Infra-
the simulations, ps should be set to an optimal value which structure for Real-Time Applications,” Proc. IEEE INFOCOM ’03,
pp. 1521-1531, Mar. 2003.
is around 0.7 to maximize both the efficiency and the [7] S. Iyer, A. Rowstron, and P. Druschel, “Squirrel: A Decentralized,
effectiveness of data lookup. Peer-to-Peer Web Cache,” Proc. 21st Ann. ACM Symp. Principles of
Distributed Computing (PODC), 2002.
[8] A.R. Bharambe, S.G. Rao, V.N. Padmanabhan, S. Seshan, and H.
7 CONCLUSIONS AND FUTURE WORK Zhang, “The Impact of Heterogeneous Bandwidth Constraints on
DHT-Based Multicast Protocols,” Proc. Ann. Int’l Workshop Peer-to-
In this paper, we have proposed a hybrid peer-to-peer system Peer Systems (IPTPS ’05), pp. 115-126, Feb. 2005.
which combines both the structured peer-to-peer network [9] D. Milojicic, V. Kalogeraki, R. Lukose, K. Nagaraja, J. Pruyne, B.
and the unstructured peer-to-peer networks to form a two- Richard, S. Rollins, and Z. Xu, “Peer-to-Peer Computing,”
tier hierarchy to provide efficient and flexible distributed Technical Report HPL-2002-57, HP Labs, Mar. 2002.
[10] S. Ratnasamy, P. Francis, M. Handley, R. Karp, and S. Shenker,
data sharing service. The top tier is the t-network which is a “A Scalable Content Addressable Network,” Proc. ACM
structured ring-based peer-to-peer network providing effi- SIGCOMM ’01, pp. 161-172, 2001.
cient and accurate service. The bottom tier is composed of [11] I. Stoica, R. Morris, D. Karger, M.F. Kaashoek, and H.
multiple unstructured s-networks which provide approx- Balakrishnan, “Chord: A Scalable Peer-to-Peer Lookup Protocol
for Internet Applications,” IEEE/ACM Trans. Networking, vol. 11,
imate best-effort service to accommodate flexibility. By no. 1, pp. 17-32, Feb. 2003.
assigning peers to the t-network or the s-network, the hybrid [12] V. Vishnumurthy and P. Francis, “On Heterogeneous Overlay
peer-to-peer system can utilize both the efficiency of the Construction and Random Node Selection in Unstructured P2P
structured peer-to-peer network and the flexibility of the Networks,” Proc. IEEE INFOCOM ’06, 2006.
[13] Gnutella Development Forum, The Gnutella v0.6 Protocol, http://
unstructured peer-to-peer network and achieve a good groups.yahoo.com/group/the gdf/files/, 2009.
balance between them. Our simulation results show that [14] Bittorrent, http://www.bittorrent.com/, 2009.
compared to structured peer-to-peer networks, the hybrid [15] P. Ganesan, Q. Sun, and H. Garcia-Molina, “YAPPERS: A Peer-to-
Peer Lookup Service over Arbitrary Topology,” Proc. IEEE
system has less lookup latency and connum, thus has higher
INFOCOM ’03, pp. 1250-1260, 2003.
data lookup efficiency. The efficiency can be further [16] B.T. Loo, R. Huebsch, I. Stoica, and J.M. Hellerstein, “The Case for
increased when considering link heterogeneity and topology a Hybrid p2p Search Infrastructure,” Proc. Workshop Peer-to-Peer
awareness. Compared to unstructured peer-to-peer net- Systems (IPTPS ’04), pp. 141-150, Feb. 2004.
[17] S. Ratnasamy, M. Handley, R.M. Karp, and S. Shenker, “Topolo-
works, the hybrid system has much lower data lookup
gically-Aware Overlay Construction and Server Selection,” Proc.
failure ratio. By adjusting the system parameter ps and IEEE INFOCOM ’02, June 2002.
TTL value, the data lookup failure ratio can be lowered to an [18] H. Zhang, G. Neglia, D. Towsley, and G. Lo Presti, “On
acceptable level. Moreover, the join latency can be minimized Unstructured File Sharing Networks,” Proc. IEEE INFOCOM ’07,
pp. 2189-2197, 2007.
by selecting appropriate ps .
[19] M. Zaharia and S. Keshav, “Gossip-Based Search Selection in
Our future work includes how to design a caching Hybrid Peer-to-Peer Networks,” Proc. Ann. Int’l Workshop Peer-to-
scheme for the hybrid peer-to-peer system to improve the Peer Systems (IPTPS), Feb. 2006.
system performance. In the case that some extremely [20] Y. Chawathe, S. Ratnasamy, L. Breslau, N. Lanham, and S.
Shenker, “Making Gnutella Like p2p Systems Scalable,” Proc.
popular data are requested by a large amount of peers,
ACM SIGCOMM, Aug. 2003.
the peer hosting the data may be overwhelmed by the large [21] S. Saroiu, K. Gummadi, and S. Gribble, “A Measurement Study of
amount of requests. The goal of the caching scheme is to Peer-to-Peer File Sharing Systems,” Proc. Multimedia Computing
balance the load of the hosting peer when popular data are and Networking Conf. (MMCN), Jan. 2002.
[22] K.P. Gummadi, R.J. Dunn, S. Saroiu, S.D. Gribble, H.M. Levy, and
requested by many peers. The idea is to distribute the load
J. Zahorjan, “Measurement, Modeling and Analysis of a Peer-to-
among as many peers as possible so that no peer is Peer File-Sharing Workload,” Proc. ACM Symp. Operating System
overwhelmed. The challenges include how to choose some Principles (SOSP), Dec. 2003.
surrogate peers to redirect the requests to, which data [23] M.-T. Sun, C.-T. King, W.-H. Sun, and C.-P. Chang, “Attribute-
Based Overlay Network for Non-DHT Structured Peer-to-Peer
should be cached and how long the data should be cached. Lookup,” Proc. Ann. Int’l Conf. Parallel Processing (ICPP), Sept.
2007.
[24] http://www.cc.gatech.edu/projects/gtitm/, 2009.
ACKNOWLEDGMENTS
The research was supported in part by the US National
Science Foundation under grant numbers CCR-0207999 and
CCF-0744234.
14. YANG AND YANG: AN EFFICIENT HYBRID PEER-TO-PEER SYSTEM FOR DISTRIBUTED DATA SHARING 1171
Min Yang received the BEng and MS degrees in Yuanyuan Yang received the BEng and MS
computer science from Beijing University of degrees in computer science and engineering
Posts and Telecommuncations and Tsinghua from Tsinghua University, Beijing, China, and
University, China, in 1999 and 2002, respec- the MSE and PhD degrees in computer science
tively, and the PhD degree in electrical engineer- from Johns Hopkins University, Baltimore,
ing from Stony Brook University, New York, in Maryland. She is a professor of computer
2009. He is currently with FalconStor Software in engineering and computer science at Stony
Melville, New York. His research interests in- Brook University, New York, and the director of
clude multicast routing, network coding, and Communications and Devices Division at New
peer-to-peer networks. He is a student member York State Center of Excellence in Wireless
of the IEEE and the IEEE Computer Society. and Information Technology. Her research interests include wireless
networks, optical networks, high-speed networks, and parallel and
distributed computing systems. Her research has been supported by
the US National Science Foundation and the US Army Research
Office. She has published over 200 papers in major journals and
refereed conference proceedings and holds six US patents in these
areas. She is currently an associate editor for the IEEE Transactions
on Computers and a subject area editor for the Journal of Parallel and
Distributed Computing. She has served as an associate editor for the
IEEE Transactions on Parallel and Distributed Systems. She has
served as a general chair, program chair, or vice chair for several
major conferences and a program committee member for numerous
conferences. She is a fellow of the IEEE and the IEEE Computer
Society. More information about her and her research can be found at
http://www.ece.sunysb.edu/yang.
. For more information on this or any other computing topic,
please visit our Digital Library at www.computer.org/publications/dlib.