Peer-to-peer computing is defined as the sharing of computer resources and information through direct exchange. The most distinct characteristic of P2P computing is that there is symmetric communication between the peers, each peer hasboth a client and a server role. The advantages of the P2P systems are multi-dimensional, they improve scalability by enabling direct and real-time sharing of services and information, enable knowledge sharing by aggregating information and resources from nodes that are located on geographically distributed and potentially heterogeneous platforms, and provide high availability by eliminating the need for a single centralized component. The basic characteristic of the P2P network is that there is a group of nodes with the same type of interests connected over the same communication system. The P2P network is self-organized and self-administrative as the nodes autonomously discover their peers, and self-healing as the nodes automatically try to find new peers if their current peers are (temporarily or permanently) disconnected from the network. The Peer-to-Peer (P2P) architectures that are most prevalent in today's Internet are decentralized and unstructured.
When the routing state maintenance mechanism is decided at the design stage, the maintenance operation is typically optimised to perform well in the com- mon scenario. As a result, under uncommon scenarios the maintenance opera- tion may not be ideal. If the scope of adaptation incorporates the algorithmic structure of the routing state maintenance mechanism, not only are designers freed from having to make decisions about unknown operating environments, but there is also a greater opportunity to adapt to unexpected operating condi- tions for the adaptation mechanism. There may be a maintenance mechanism that is highly effective in a particular scenario and not as effective in others. An algorithmic adaptation mechanism can detect such cases and gracefully switch between different maintenance strategies to suit the current environment. Fur- ther, when the system is faced with a scenario that is not anticipated by the system designers, the system may be able to generate its own maintenance mech- anism and can potentially offer superior adaptation to unexpected conditions. In order to algorithmically adapt the routing state maintenance in a structured overlay, a set of challenges must be overcome. The following section discusses these challenges.
Peer-to-peer network is a well established area in lit- erature. The main two types of them are Structured and Unstructured peer-to-peernetworks. We refrain from using the unstructured topology because it uses flooding techniques as a search mechanism. Although the search process has been improved using for example random walks; this type of a topology still requires expensive search process. We use structuredpeer-to-peer topology in MashChord. Some of the most popular structuredpeer-to-peer platforms are Chord , CAN , and Pastry . Chord  is a key lookup protocol that works by having a logical structured arrangement of peers and resources on a virtual ring topol- ogy. A hash function(SHA1) is used to generate identifiers for peers and resources. More information about Chord is provided in the next section because we adopt its topology and protocol in our work. CAN  is another structuredpeer-to-peer key lookup system. CAN arranges peers and resources in a virtual dimensional coordinate space such that each peer resides in a zone specific to it. Therefore, when a resource is mapped to a given zone, the peer responsible of that zone is the one that hosts and maintains that resource. Pastry  is a similar work to Chord where each node is assigned a unique identifier from 128 bit space and Pastry protocol routes each message and key to the nodeID numerically closer to the given message key. Structuredpeer- to-peernetworks can be used in different domains such as in  which surveys simulators built on top of structured and unstructured peer-to-peernetworks. Another work  proposes a scheme that converts static network topology into a dynamic one built on top of structuredpeer-to-peer network. OE-P2RSP  is a structuredpeer-to-peer system built on top of Pastry. It adds enhancements over Pastry such as avoiding centralized object ID generation. It also uses objects group to make sure that objects that belong to the same group reside on the same node. The work in  targets the problem of free riding which happens when users make use of the peer-to-peer network without contributing with resources to the network.
Abstract: The huge popularity of recent Peer-to-Peer (P2P) file sharing systems has been mainly driven by the scalability of their architectures and the flexibility of their search facilities. Such systems are usually designed as Unstructured P2P networks. So, designing an efficient search algorithm is a key challenge in unstructured peer-to-peernetworks due to the unstructured paradigm. In this study, we proposed a Search Result Path Caching algorithm (SRPC). The proposed scheme combines the DynamicSearch (DS) algorithm and result path caching for effective search. This proposed algorithm takes the advantages from dynamicsearch and path caching technique works with the queried results. After the successful discovery of results the queries are returned to originator, according to our proposed algorithm the originator will stores the results for future references. In future stored references are used to search the information without querying the overall network. We analyze the performance of our algorithm based on some performance metrics including the success rate, search time and search efficiency. The numerical results shows that the proposed SRPC algorithm performs about 5 times better than DS, 125 times better than flooding and 275 times better than Random Walk (RW) in power-law graphs.
Documents are assumed to fall into a number of thematic categories. Each node stores an approximate number of documents from every category that can be retrieved through any link attached to that particular node. The query termination condition always relates to a minimum number of hits. The forwarding process is similar to DFS: A node that cannot satisfy the query stop condition with its local repository will forward it to the neighbor with the highest “goodness” value. Goodness of a node depends upon number of related documents present with that particular node and with the nearby nodes. In RI, the destination of a packet is based upon the content of query. They give direction towards the document rather than the location of the document. This is another keyword-search approach , which trades index maintenance overhead for increased accuracy. RIs require flooding in order to be created and updated, so the method is not suitable for highly dynamicnetworks. Moreover, stored indices can be inaccurate due to thematic correlations and relocation of objects in a dynamic environment, which is the most major problems with this technique .
While the structuredness of DHTs enables P2P systems to search in a diameter- bounded, content addressable overlay network, the consequence is that peer nodes are re- quired to select and maintain a nearly fixed number of overlay neighbors deterministically based on their hash value obtained from a consistent hash function, regardless of the het- erogeneity among nodes’ service capacity, Internet connectivity, lifetime in the system, or even willingness to serve. As a result, more capable and/or more reliable peers cannot take more responsibility in routing and connecting of the overlay network because of the design philosophy inherent in DHTs. This limits systems’ ability to exploit the full potential of its peers, and the resultant overlay network is generally susceptible to high performance degradation in the presence of high node dynamics such as frequent node join, departure and failures. DHT literature also refer node dynamics in a DHT system as churn, and we will use these two terms interchangeably through this dissertation.
networks a new system named Gossip Trust is presented in . An approach proposes a new experimental weighting method for picking the best-expected route to create a role centered trust series. They put on past profound heuristics to measure the route complication and to measure the connecting competence . Hybrid search in P2P networks is presented in . Range query processing is one of the major applications in the domain of P2P networks; an approach handling this prospect is presented in . An improved Hybrid P2P approach that proposes and controls the concept of boot net in P2P networks and the results showed that this approach is much harder to shutdown . The concept of clustering through the use of k-mean approach in P2P networks is presented in .Liu et al. inoffered a principal approach intended for the topology discrepancy problem among unstructured P2P systems. The problem of resolving the mismatch problem can be analyzed .P2P streaming through a distributed protocol is presented in . Management of multidimensional past data in unstructured P2P networks procedure through a novel method is presented in . The concept of probabilistic flooding in generalized form in unstructured P2P networks is presented in . Heterogeneous search in the P2P networks can be analyzed in .The structured P2P networks show a more efficient response to the fact of fault tolerance . The design and analysis of a protocol named Ulysses can be analyzed in . The analysis of P2P network properties through the use of a graph is presented in . Analysis of P2Psystems in the prospect of error recovery is another important factor in the prospect of P2P networks . The impact of free riding in P2P networks resulted that, with the presence of free riding concept, P2P networks operates more efficiently . Routing or direction finding with secure parameters in structured P2P networks is presented in .The analysis of the contribution of each peer in P2P networks is important, an DNS
these requirements they introduced a basic set of group management style. They analyze the requirements that social groups have with regard to groups in distributed system. Here they discuss the problem of group management in fully decentralized social networking systems. This study is intended towards various group management models of social networking system to make them decentralized (P2P) systems rather than for P2P computing systems. Sing Jin Choi, et al.  proposed a new group based dynamic computational replication mechanism in a P2P grid computing system. Their proposed mechanism adaptively replicates computation on the basis of volunteer properties. The proposed mechanism dynamically adjusts the number of redundancy, and selects the volunteers that execute the replicated computations accordingly to the respective volunteer group properties. The replication mechanisms not only calculate the number of redundancy, but also select replicas on the basis of volunteer group. Their simulation results shows that the proposed mechanism reduce the number of redundancy and therefore complete more tasks. Yung Ming Li, et al.,  developed an analytical model to evaluate the impacts of network scale and system parameters on the performance of P2P file sharing network. In this study by analytical and simulation analysis they shows that increasing network scale has a positive effect on the expected content availability and transmission cost, but a negative effect on the expected provision and search costs. They also propose the optimal sizing and grouping decisions. However effects on P2P computing network is not given.
then check the path length and rewards the path if it is shorter than the average length, the path will be penalised otherwise . LARD eases the problem of network flooding. However, by using TTL like Mashayekhi and Habibi’s technique, LARD also has an issue of fault positive error . Improved Adaptive Probabilistic Search (IAPS) utilises ant-colony optimisation to search space and search overhead . The technique is proven to be better than popular random walk and Adaptive Probabilistic Search (APS) . DHMCF is a resource discovery technique for a pure unstructured P2P network that responds to dynamic requests . In this technique, there are 5 units to gather information, make decisions, find resources and balance the load. This technique, however, suffers from low robustness and security .
In order to use Internet resources efficiently we need to search and locate information efficiently. System performance diminishes by either duplicating a large quantity of data on each and every node or flooding query to all the nodes in the network. Firstly, this paper reviews various searching algorithms. Search techniques can be classified as blind search in which information about neighbors is not kept by the peer and informed search where peers store information for routing queries to other nodes. It discusses how range queries can be processed efficiently by rotating scheme over structured P2P systems and secure searching algorithm based on topology adaptation which penalizes the malicious peers. Genetic algorithm providing parallel search are also covered in the paper. Lastly, it focuses on merits, demerits and applicability of these algorithms in different situations.
Distributed Search. In order to determine what makes a social network search- able by a distributed algorithm using only local information, we used Kleinberg’s idea of distance-dependent connection probability  to generalize a promising random graph model, Kronecker graphs . Specifically, we developed a mathematically tractable random network model (distance-dependent Kronecker graphs) incorporat- ing searchability, where individuals use their local and long-distance connections to their peers to optimally route a message through the network. This model general- izes Kronecker graphs, using a family of “distance”-dependent matrices and a new Kronecker-like operation, as opposed to the static generator matrix used in the origi- nal Kronecker graph model. As a result, our network model defines both local regular structures and global distance-dependent connections. While this model is more com- plicated than the original model, it is more general, as it can generate existing social network models, and more importantly, networks that are searchable. These prop- erties emerge naturally from the definition of the embedding of the nodes and the probability of connection within the family of generator matrices. More generally, any lattice-based network model with distance-dependent connection probabilities can be analyzed using the framework described in Chapter 2 for exploring degree dis- tribution, diameter, and searchability. In particular, our searchability analysis shows how to make any network model searchable by defining the appropriate probability of connection. Using this analysis, we are able to show a particularly nice aspect of our model; i.e., it is rather robust to changes in the long-distance connection probability, as opposed to earlier models.
Dedicated clusters, in many cases, can be measured using benchmarks derived from parallel machines. However, wide area distributed computing, peer-to- peer and grass-roots volunteer computing, where participating nodes are highly transient, does provide numerous new challenges. These challenges are not based in performance metrics alone: the voluntary and participatory nature of these new designs require a benchmark to consider some of the more qualitative aspects of the exercise as computing efficiency alone can be less important, to the outcome of the project, than the ability to attract participants. The costs of joining and the effort required to start and host a distributed computing exercise can have a large effect on the success or outcome of a modern project as it tries to attract participating nodes. Thus it is the case that this benchmark also introduces an evaluation criteria for distributed computing systems that can serve as a basis for evaluation, comparison and selection.
to 0.0 at other peers in the system, and then starts sending infected files. We have to admit, though, that this is a remote possibility. For starters, in real-life P2P systems the main concern of poisoners is to destabilize the P2P network with the minimum possible effort. Therefore, poisoners who are willing to send genuine content to other peers (and hence dedicate the resources to do such thing) are extremely rare. We see this often in the real-life BitTorrent P2P file sharing network, for example . In addition, since each peer in the network has its own P c m for the poisoner peer m, the
This paper presented a semantic search algorithm for the collaborative Open Hypermedia System that creates a semantic overlay of related entities and uses clustering to optimise the search. Our algorithm performs very well in controlled environments with static content. The number of hops required to achieve the same percentage of recall varies in direct proportion to the cache rate between topologies. The search algorithm also performs satisfactorily when peers update their topics randomly and has been proven suitable for locating information in an environment where peers change their contents randomly. However, clustering of related entities at times leads to the formation of information islands and the ways to reorganise the topology in terms of published contents also form a part of the future study.
The messages arriving from the peers in the overlay are processed to remove redundancy and remove entries whose owner peers are unavailable. The redundancy is removed by seeing to that the set of files have unique FIDs. This is achieved here by using hash table with FIDs as key mapping to corresponding PIDs. When a file is selected for downloading, a message containing FID is sent to the owner peer. Owner peer responds to this download request by sending the file to the querying peer. As soon as the download is finished a certificate is sent to the downloader peer from the owner peer marking it (the certificate) as a proof of download. The downloader peer then could tag the associated file within a user defined time span. A similar certificate is sent to the owner peer along with the time span. Digital certificates are present only in the owner peer and downloader peer in the period of time span. Time span is needed as the user might not be interested in tagging the file immediately after download and leaving them without tagging would waste memory resources heavily.
A peer-to-peer network is a type of decentralized and distributed network architecture in which individual nodes in the network act as both suppliers and consumers of resources, in contrast to the centralized client–server model where client nodes request access to resources provided by central servers . In this network, tasks are shared amongst multiple interconnected peers who make a portion of their resources directly available to other network participants, without the need for centralized coordination by servers . Below figure provides a conceptual representation of the P2P overlay topology. In this, every machine plays the role of client and server at the same time. Although a P2P network has a number of advantages over the traditional client-server model in terms of efficiency and fault-tolerance, additional security threats can be introduced. Users and IT administrators need to be aware of the risks from propagation of malicious code, the legality of downloaded content, and vulnerabilities within peer-to-peer software .