A survey of data sharing and security issues in P2P networks.
P2P architecture is a network architecture where each member of the network contributes and shares resources with other members. There is no hierarchy level that differentiates members because they are all considered to be equal (peer). Without any server to manage the network, all members of P2P network need to maintain the network structure and resource management. In order to do this, one member is required to have network connection with several or all other members. Generally, a "Peer-to-Peer" system is a decentralized system: each node in the network possesses the same role as any other node. The opposite is a client-server system, which is centralized on the server. An immediate advantage of Peer-to-Peer architecture is scalability and fault tolerance.
The different types of P2P networks classified based on their topology are discussed below:
* Centralized P2P:
A centralized P2P system is one in which participants can freely join the network, but contact a central authority for certain required information. Probably the best example of a centralized P2P system used on a large scale is BitTorrent. BitTorrent emerged as a protocol for content distribution that leverages the decentralized nature and low cost of P2P. With BitTorrent, a single distribution source (the central authority) releases content for download. As peers in the network download the content, they also upload parts of it to other peers. At the start, there is only a single source from which to download the content, but each downloading peer also acts as a source.
* Pure P2P:
A pure P2P system is a distributed system in which all participants are equal. This generally means that responsibility is evenly distributed among the members, and that no single node is more powerful than any other. A pure P2P system is a network in which peers can participate without the need for any central authority and peers are supposed to operate according to the same protocol. The central server or router is completely absent and each peer acts as client and server at the same time. This is also sometimes referred to as "serverless" P2P.
* Hybrid P2P:
A central server is present which maintain information about the network. These information are stored by peers which is their responsibility. A peer contacts another peer through the query system with the server to get the location.
* Super-peer P2P:
A super-peer P2P system is one which combines elements of centralized and pure P2P designs. Super-peer designs are made up mostly of peers with relatively few connections. However, there are some peers that have many more connections than normal peers; these peers are known as super-peers or ultra-peers. The main advantage of these super-peers is that routing requires fewer hops due to the high level of connectivity.
* Mixed P2P:
Between "hybrid" and "pure" P2P networks. An example of such a network is Gnutella which has no central server but clusters its nodes around so-called "supernodes".
Napster-like networks (Napster Inc.) are the first generation p2p networks. Those networks did not possess any complex implementation and regularly depend on a central server (hybrid P2P). The central server technique has been utilized since it allows retaining control over the network and searches, though it also means there is a single point of malfunction.
Gnutella was considered as the second main P2P network. After Napster's downfall, Gnutella was created as a decentralized network, i.e. one that could not be shut down by simply turning off a server. Initially, the proposed model did not become popular due to some bottlenecks created during file searching.
The recently emerging P2P networks are considered as Third generation networks. They are a response to the legal attention P2P networks have been receiving for a few years and have built-in anonymity features. The basic architecture of a P2P network is shown in Fig. 1.
The advantages of P2P network architecture over CS are as follow:
* There is no need for a server to maintain the network. Without server, all burden and resource requirements are shared within the members of the network.
* There is no single point of failure in the network. If there is one broken connection in a P2P network, there are other connections with others that could replace or serve as a router to perform data communication.
* With no single point of failure, a P2P network implementing special content addressing such as CAN  or Pastry  will be able to recover the data in the network during a major disruption.
* The nature of P2P network encourages data duplications amongst the member of the network. Therefore, when there is a major network disconnection, it is still possible to reconstruct the network and the data inside it.
* It has higher scalability, i.e. more number of members in the network means more available resources.
In this paper, various designs and topologies proposed in different conventional systems have been discussed. Each model has its own advantage and systematic routing to increase the performance of communication and data sharing over the P2P networks.
II. Literature Survey:
There are several protocols for distributed peer-to-peer networks available for decades. Yet, each design was unable to work in real-world networks with restricted routes. Some of such techniques are discussed in this work.
2.1 Reputation Management Systems:
Kamvar et al. 2003 proposed the malicious node detection procedure in larger peer to peer networks based on Eigen trust. The trust value of each node in peer to peer networks were computed and further eigen vector were computed for a set of nodes in trust evaluated system. The global reputation metrics were determined based on the reputation algorithm in global mode discrimination method. It accumulated very low overhead value for the trust estimation of the nodes in P2P networks.
Xiong and Liu  developed an algorithm for static and dynamic peer to peer routing formation in larger area peer networks which were based on peer trust estimation process. The trust value of each node in larger peer to peer networks was based on the transaction factor where, each trust was based on trust metrics. They were able to tolerate the middle tampering attacks in centralized with respect to decentralized system in larger P2P network environments.
Zhou and Hwang  developed an algorith for trust estimation technique which were based on routing scalability in P2P reputation system in overlay networks based on trust estimation in node as peers. Ebay transaction were checked for the capability of larger users such as ten thousnd users at a time by finding the trust values in each node in network. The power-law distribution based on certain impacts in structural networks and unstructural networks. Power trust establishment method was used to compute the trust value of each peers in network and based on these trust values, the peers are able to transfer the packets through the network. The reputation parameter was improved by utilizing the power trust value in P2P networks.
Zhang et al.  analyzed three different problems in previously proposed trust systems and proposed a new system. They are, (1) no strong incentives designed to stimulate honest participation in the trust system; (2) a binary QoS differentiation method that classifies a service as either good or bad without any interim state, thus limiting the potential for use by P2P networks in which servers have diverse capabilities and clients have various QoS demands and, (3) failure to protect the privacy of references, which is important for obtaining honest feedback.
2.2 Attack-Defense Systems:
File sharing between numbers of peer nodes in network is a critical task due to various numbers of attacks such as collision attack. These type of attack mainly degardes peer to peer network performance interms of packet delivery ratio. Another kind of attack in P2P network is Sybil attack. The collision attack is very similar to Sybil attack with respect to its characteristics of the peer nodes in network. These kind of networks produced malicious nodes, which further degraded the reputation of the network in peer manner. The collision attack makes collision between the numbers of users in larger P2P networks. Eventhough, the Sybil attack have similar characteristics with collision attack, it generates fake accounts and produces large amount of fake datas and spreads these datas throughout the network.
Ozkasapa et al.  framed a methodology for low energy consumption for larger peer to peer networks. The authors utilized greedy methodology for energy reduction for transferring larger amount of information between peers in network. Emrah Qema et al.  also developed an efficient methodology using gossip-based model for discovering the accuracy false errors which also affects the performance of the network.
Ozkasapa et al.  also presented fair-share buffering method to improve the reliability of dissemination. They used stepwise fair-share buffering to provide uniform load distribution thereby reducing the overall buffer usage. They analyzed the comparative performance results with conventional buffering approaches and random buffering which serves as a standard unit.
Richard Lin et al.  presented a Flexible data sharing algorithm for cloud-enabled extended peer system using Extended Bestpeer. They proposed a Distributed Hash Table with Round Trip Time (RTT) to provide data availability and data accessibility in bio-informatics, thereby ensuring scalability.
With an immense increase in the peer-to-peer computing paradigm, PeerShark , a novel method was introduced to detect botnet traffic in peer to peer network. However, distributed model remained unaddressed.
Monitoring of File sharing in peer to peer network was introduced in Timpanaro et al.  to efficiently analyze the anonymity of the application and to improve the I2P statistics. A statistic model was introduced in Zhou et al.  using generic replication algorithm to solve the issues related to video on demand.
The work on peer-to-peer networks presented in El-Ansary et al.  made use of small-world algorithms based on the proposition by Watts & Strogatz  on "rewiring" the network. In El-Ansary et al. , the idea of rewiring is applied to a Chord  overlay. Pandurangan et al.  and Pandurangan et al.  created a low-diameter peer-to-peer network but rely heavily on a central server that is needed to coordinate the connections between peers. This proposal creates a potential single point of failure in the overlay network. The authors also do not address the resilience of such a network in the event of targeted node removal, various attacks, or misbehaving nodes. Under such conditions the performance of the network would likely degrade and deviate from the low-diameter design goal.
On the other hand, in Bernsteing et al.  the authors suggest the use of machine learning techniques to help a peer pick a serving node among the ones carrying its desired object, instead of aggregating all the bandwidth and taking advantage of all serving nodes. The focus of their work was to find the most reliable node to download from in terms of its offered bandwidth and time spent in the network and does not take advantage of the aggregated bandwidth. In addition, Byers et al.  studied the benefits of using cooperative nodes in order to increase the storage capacity of the whole system, which is mainly targeted towards applications where nodes are cooperative.
Other research works  have investigated the performance of peer-to-peer systems but not for the case of parallel download scenario from a node's perspective. There has been little or no work on the analysis of parallel downloads for peer-to-peer networks. Adler et al.  studied the parallel downloads determining the optimal peers to download from, minimizing the cost associated with the download, assuming guaranteed bandwidth between clients and servers, and a cost for downloads directly proportional to the transfer from every serving node.
2.3 Inter-system content sharing:
Most of the previous works in this category, except Konishi et al. , only addressed the problem of routing a request between overlays which we hereby call inter-overlay routing. In Konishi et al. , they proposed a scheme not only for inter-overlay routing but also for exchanging data in pure flooding P2P file-sharing networks.
2.4 Inter-overlay routing:
There are lot of research works in the field of inter routing mechanism in large connected servers. Konishi at al. proposed inter routing mechanism where the links were established inside the routing elements and Ciancaglini et al.  proposed connected nodes for routing the datas between multiple nodes in the network. In Cheng, distributed hash tables (DHT) were used for inter routing connection establishment where the internal regions of the network router were connected with common mode point. The senders DHT were able to transmit and receive the datas from the receiver DHT where the connection was established through the multiple connected nodes in their path. These multiple connected nodes were acted as the peers and formed the peer to peer network.
Liquori et al.  proposed inter routing connection protocol named as Babelchord. This protocol was used to overlay the peers between the intermediate nodes. These peers were requested to change the information contents between multiple nodes in large P2P networks. Liquori et al.  developed a protocol Synapse for connection establishment between peer nodes in the larger network. The authors used white and black box method to route establishment between various DHT where they were connected in serial manner. The authors were also used Control Overlays to obtain the information through various look up results. The authors were not exposed black box method for the passage of information between connected nodes in P2P networks.
Synapse protocol was developed by Ciancaglini et al.  for P2P networks and further it was upgraded by Ciancaglini et al.  for increasing the numbers of peers in networks which increases the coverage area of the network. Liquori et al.  proposed the technique of overlay in peer network model. The authors further analyzed the throughput and packet delivery ratio for analyzing the capability of the developed protocol for their proposed system.
This paper discusses the P2P network related issues in detail. In summary, traditional reputation management systems only consider transaction history of nodes in calculating reputation values, which cannot effectively prevent the aforementioned common threats. P2P systems and the applications built upon them make up a large portion of the usage of the Internet today, yet they tend to lack good security and efficiency or trade off one for the other. There is a demand for a secure and efficient routing design which works well in the kind of networks used today.
[1.] Adler, M., R. Kumar, K. Ross, D. Rubenstein, D. Turner and D. Yao, 2004. Optimal Peer Selection in a Free-Market Peer-Resource Economy. Second Workshop on the Economics of Peer-to-Peer Systems (P2P ECON), Cambridge, Massachusetts.
[2.] Bernsteing, D.S., Z. Feng, B.N. Levine and S. Zilberstein, 2003. Adaptive Peer Selection. Proceedings of the 2nd International Workshop on Peer-to-Peer Systems (IPTPS), Berkeley, California.
[3.] Byers, J., J. Considine, M. Mitzenmacher and S. Rost, 2002. Informed Content Delivery across Adaptive Overlay Networks. ACM SIGCOMM.
[4.] Ciancaglini, V., L. Liquori and G. Ngo, 2011. Towards a Common Architecture to Interconnect Heterogeneous Overlay Networks. The 17th International Conference on Parallel and Distributed Systems, ICPADS 2011, Tainan, Taiwan, pp: 817-822.
[5.] Ciancaglini, V., L. Liquori, G. Ngo and P. Maksimovic, 2012. An Extension and Cooperation Mechanism for Heterogeneous Overlay Networks. Networking Workshops.
[6.] El-Ansary, S., L.O. Alima, P. Brand and S. Haridi, 2003. Efficient broadcast in structured P2P networks. 2nd International Workshop on Peer-to-Peer Systems (IPTPS '03), Berkeley, CA.
[7.] Emrah Qema, and Oznur Ozkasapb, 2012. ProFID: Practical frequent items discovery in peer-to-peer networks. Future Generation Computer Systems, Elsevier, 6: 1544-1560.
[8.] Kamvar, S., M. Schlosser and H. Garcia-Molina, 2003. The EigenTrust algorithm for reputation management in P2P networks. In Proc. of WWW.
[9.] Konishi, J., N. Wakamiya and M. Murata, 2006. Proposal and evaluation of a cooperative mechanism for pure p2p file sharing networks. Proc. of BioADIT.
[10.] Liquori, L., C. Tedeschi and F. Bongiovanni, 2009. Babelchord: a social tower of DHT-based overlay networks. ISCC.
[11.] Liquori, L., C. Tedeschi, L. Vanni, F. Bongiovanni, V. Ciancaglini and B. Marinkovic, 2010. Synapse: A Scalable Protocol for Interconnecting Heterogeneous Overlay Networks. Networking.
[12.] Margolin, N.B. and B.N. Levine, 2005. Quantifying and discouraging sybil attacks. Technical report, U. Mass.
[13.] Napster Inc. (Formerly Roxio, Inc.). Napster. http://www.napster.com/.
[14.] Narang, P, Subhajit Ray, Chittaranjan Hota, Venkat Venkatakrishnan, 2014. PeerShark: Detecting Peer-to-Peer Botnets by Tracking Conversations. IEEE Computer Society, Security and Privacy Workshops.
[15.] Ozkasapa, O., M. Qaglar, E. ule Yaz and S. Kugukgifgi, 2010. An analytical framework for self-organizing peer-to-peer anti-entropy algorithms. Performance Evaluation, Elsevier, 67(3): 141-159.
[16.] Ozkasapa, O., E. Cema, S.E. Cebeci and Tugba Koca, 2013. Flat and hierarchical epidemics in P2P systems: Energy cost models and analysis, Future Generation Computer Systems, Elsevier, 36: 257-266.
[17.] Pandurangan, G., P. Raghavan and E. Upfal, 2001. Building P2P networks with good topological properties. Technical Report.
[18.] Pandurangan, G., P. Raghavan and E. Upfal, 2003. Building low-diameter P2P networks. IEEE Journal on Selected Areas in Communications, 21: 995-1002.
[19.] Ratnasamy, S., P. Francis, M. Handley, R. Karp and S. Shenker, 2001. A scalable content-addressable network. Proceedings of the 2001 conference on Applications, Technologies, Architectures, and Protocols for Computer Communications, pp: 161-172.
[20.] Richard Lin, C.H., C.H. Wen, Y.C. Lin, K.Y. Tung, R.W. Lin and C.Y. Lin, 2013. A P2P Framework for Developing Bioinformatics Applications in Dynamic Cloud Environments. International Journal of Genomics, Hindawi, pp: 1-9.
[21.] Rowstron, A.I.T. and P. Druschel, 2001. PASTRY: Scalable, decentralized object location, and routing for large-scale peer-to-peer systems. In Middleware 2001: Proceedings of the IFIP/ACM International Conference on Distributed Systems Platforms Heidelberg, London, UK, pp: 329-350.
[22.] Stoica, I., R. Morris, D. Karger, M.F. Kaashoek and H. Balakrishnan, 2001. Chord: a scalable peer-to-peer lookup service for internet applications. Proceedings of the 2001 conference on applications, technologies, architectures, and protocols for computer communications (ACM Sigcomm 2001), pp: 149-160.
[23.] Timpanaro, J.P., Isabelle Chrisment, Olivier Festor, 2013. Monitoring Anonymous P2P File-Sharing Systems. 13th IEEE International Conference on Peer-to-Peer Computing.
[24.] Watts, D.J. and S.H. Strogatz, 1998. Collective dynamics of 'small-world' networks'. Nature, 393: 440442.
[25.] Xiong, L. and L. Liu, 2004. PeerTrust: Supporting Reputation-Based Trust for Peer-to-Peer Electronic Communities. IEEE TKDE.
[26.] Yang, X. and G. de Veciana, 2004. Service Capacity of Peer to Peer Networks. IEEE Infocom 2004, Hong Kong, China.
[27.] Yu, H., M. Kaminsky and A.D. Flaxman, 2006. Sybilguard: Defending against Sybil attacks via social networks. In Proc. of SIGCOMM.
[28.] Zhang, Y. and Y. Fang, 2007. A Fine-Grained Reputation System for Reliable Service Selection in Peer-toPeer Networks, IEEE TPDS.
[29.] Zhou, R. and K. Hwang, 2007. PowerTrust: A Robust and Scalable Reputation System for Trusted Peer-toPeer Computing, IEEE TPDS, 18(4).
[30.] Zhou, Y., Z. Tom, J. Fu and Dah Ming Chiu, 2011. Statistical Modeling and Analysis of P2P Replication to Support VoD Service. IEEE, Proceedings on INFOCOM.
(1) Dr. CG. Ravichandran and (2) J. Lourdu Xavier
(1) Principal, S C A D Institute of Technology, Palladium, India-641658
(2) Department of MCA, RVS College of Engineering, Dindigul, India-624005
Received 28 January 2017; Accepted 22 May 2017; Available online 28 May 2017
Address For Correspondence:
Dr. CG. Ravichandran, Principal, S C A D Institute of Technology, Palladam, India-641658
Caption: Fig. 1: Basic architecture of P2P network
Table 1: Comparison of literature in performance improvements of P2P networks Methodology Author & Year Technique used Implementation area fair-share Ozkasapa et al. stepwise fair- P2P network buffering  share buffering gossip-based Emrah Qema et Protocol for P2P network protocol al.  Frequent Items Discovery (ProFID) where Energy cost Ozkasapa et al. Matrix P2P network model  Partition Flexible data Richard Lin et Extended cloud-enabled sharing al.  Bestpeer / extended peer algorithm Distributed system Hash Table with Round Trip Time -- Bernsteing et machine P2P network al.  learning techniques cooperative Byers et al. cooperative P2P network nodes (to  nodes increase the storage capacity) Performance Yang & de parallel P2P network improvement in Veciana  downloads parallel downloads Determining the Adler et al. parallel P2P network optimal peers  downloads at low cost low-diameter Pandurangan et central server P2P network/ peer-to-peer al.  to coordinate overlay network the peers networks
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||peer to peer|
|Author:||Ravichandran, C.G.; Xavier, J. Lourdu|
|Publication:||Advances in Natural and Applied Sciences|
|Date:||May 1, 2017|
|Previous Article:||Cloud based result analysis and management system using rest API.|
|Next Article:||An efficient constant bit rate information bit approach for Video Steganography.|