FORM 2

THE PATENTS ACT, 1970

(39 of 1970)
&

THE PATENT RULES, 2003

COMPLETE SPECIFICATION

(See Section 10 and Rule 13)

Title of invention:

System and Method for Secure Searching in Peer To Peer Networks

Applicant:
TATA Consultancy Services Limited
A company Incorporated in India under The Companies Act, 1956
Having address:
Nirmal Building, 9th Floor,
Nariman Point, Mumbai 400021,
Maharashtra, India

The following specification particularly describes the invention and the manner in which it is to be performed.

FIELD OF THE INVENTION
The present invention generally relates to the field of computer networking, and more particularly to a secure and efficient method and framework for searching resources in a unstructured peer to peer networking environment which does not have any centralized point of administration.

BACKGROUND OF THE INVENTION

The term peer-to-peer (P2P) system encompasses a broad set of distributed applications which allow sharing of computer resources by direct exchange between systems. The goal of a peer-to-peer (P2P) system is to aggregate resources available at the edge of the Internet and to share it cooperatively among users. Specially, the file sharing peer-to-peer (P2P) systems have become popular as a new paradigm for information exchange among large number of users in a web based environment. They are more robust, scalable, fault tolerant and offer better availability of resources.

Depending on the presence of central server, peer-to-peer (P2P) system can be classified as centralized or decentralized. In decentralized architecture, both resource discovery and download are distributed. Decentralized peer-to-peer (P2P) application may be further classified as structured or unstructured network. In a structured network, there is a restriction on the placement of content and network topology. While in an unstructured P2P network, the placement of content is unrelated to topology. Unstructured P2P networks perform better than their structured counterparts in a dynamic environment. However, they need efficient search mechanisms and suffer from fake content distribution, free riding (peers who do not share, but consume resources), and whitewashing (peers who leave and rejoin the system in order to avoid penalties) etc.

Despite the existence of a number of trust management schemes for combating against free riding and distribution of malicious files, these mechanisms are not scalable due to their high computational and storage overhead. Moreover, they do not consider the quality of service (QoS) of the search conducted to download authentic files.

Distributed reputation based trust management systems have been proposed to provide protection against malicious content distribution. These schemes are of two types: (i) gossip-based and (ii) topology adaptation-based. In gossip-based schemes, a peer node evaluates trustworthiness of the resource provider from past transactions and recommendations from its neighbors. Some examples of gossip-based protocols are XREP, P-Grid and Credence. Credence employs object reputation approach for combating content pollution in P2P files sharing systems. The peer collects votes from others in the phase of gossip and calculates correlation among votes.

In topology-adaptation-based schemes, a peer node connects to those peers from whom there are high probabilities to get authentic files and drops malicious peers from its neighborhood. Trust management via dynamic topology was first proposed in Gnushare to reduce free riding over the peer to peer network overlay. Guo et al. have proposed trust-aware adaptive P2P topology to control free-riders and malicious peers in order to reduce spurious file download and banish free riders from the community.

Tang et al. have proposed trust-based incentive mechanisms to encourage sharing of large number of authentic files in a P2P network. Zhuge et al. have proposed trust-based probabilistic search algorithm called P-Walk to improve search efficiency and to reduce unnecessary traffic in P2P networks. In P-Walk, neighboring peers assign trust scores to each other. During routing, peers preferentially forward queries to the highly ranked neighbors. Tang has proposed trust-based broadcast searching protocol (TBS), where peers with higher reputation values have higher broadcast scales so as to provide incentives to these peers for sharing their contents.

De Mello et al. have proposed a searching mechanism that is based on discovery of trust path among peers in a P2P network. The authors have argued that since the trust path discovery does not allow resource replication, it is very sensitive to parameter choices in selective forwarding algorithms. A global trust model based on distance-weighted recommendations has been proposed by Li et al. to quantify and evaluate the peers in a P2P network.

Further, a mechanism named adaptive peer-to-peer technologies (APT) for the formation of adaptive topologies has been proposed by Condie et al. to reduce spurious file downloads and free ridings, where a peer connects to those peers from whom it is most likely to download satisfactory content. It adds or removes neighbors based on local trust and connection trust which are decided by its transaction history. The scheme follows a defensive strategy for punishment where a peer equally punishes both malicious peers as well as neighbors through which it receives response from malicious peers.

This strategy is relaxed in the reciprocal capacity-based adaptive topology protocol (RC-ATP) proposed by Tian et al., where a peer connects to those peers having higher reciprocal capacity. Reciprocal capacity is defined based on peer’s capacity of providing good files and of recommending source of download. In addition, a response selection mechanism is proposed to reduce the probability of download from malicious peers. Due to the adequate connections between reciprocal peers, network connectivity is better in RC-ATP than APT. It also reduces cost incurred by unsatisfactory download. However, overhead of topology adaptation is significantly higher in RC-ATP. In the light of above foregoing, required is a system capable of adapting the honest peers in a topologically better position at reduced network bandwidth occupancy and low storage overhead.

In the US Patent No. 7743044, a mechanism for information retrieval in fully decentralized, distributed, peer-to-peer network systems has been presented. Peer profiles are aggregated and collected in real-time by each peer. Each peer uses and integrates knowledge that it collects during query-reply cycles for each future query retrieved, thereby learning over time and making information retrieval a more intelligent and rapid process. Each peer then autonomously decides which of its peers are most likely to have an answer to a given query.

The US Patent No. 7089301 presents a method and system for intelligently directing a search of a peer-to-peer network, in which a user performing a search is assisted in choosing a host which is likely to return fast, favorable results to the user. A host monitor monitors the peer-to-peer network and collects data on various characteristics of the hosts which make up the network. Thereafter, a host selector ranks the hosts using the data, and passes this information to the user. The user then selects one or more of the highly-ranked hosts as an entry point into the network. Additionally, a cache collects a list of hosts based on the content on the hosts. In this way, a user may choose to connect to a host which is known to contain information relevant to the user’s search.

The US Patent Publication No. 2007/0016587 A1 presents a system and method for representing and rate the trustworthiness of peers as providers of content and data (codats) relevant to the peers’ interests. In one embodiment, trust is propagated through transaction pipes (paths) along which codats located in search for codats relevant to an area of interest may be accessed by the requestor.
Similarly, US Patent Publication No. 2007/0016587 A1 presents a method of searching across a peer-to-peer network, where each peer has searchable content and is connected to a plurality of neighboring peers via links. The links constitute a contributor pipeline to pass a query around the peers, to contribute any available result content at respective peers, and to establish a read pipeline via the links to pass contributed results back to a query originator. The read pipeline allows results to be read by any intervening peer that is interested. The pipelines are setup using a subscribe and publish mechanism.

Though the above mentioned prior arts have attempted to address the technical problem of efficient resource searching in a peer to peer network , the existing searching mechanisms lack efficiency and suffer from the problems of fake content distribution, free riding (peer nodes which do not share, but consume resources), whitewashing (peer nodes who leave and rejoin the system in order to avoid penalties) etc. In fact, three main problems which still exist in unstructured P2P networks are: (i) fake content distribution, (ii) lack of scalability in searching techniques, and (iii) free riding. Each of these major problems is discussed briefly hereon.
(i) Fake content distribution - Open and anonymous nature of P2P applications lead to complete lack of accountability of the content a peer puts in the network. The malicious peers often use these networks to do content poisoning and to distribute harmful programs such as Trojan Horses and viruses. Distributed reputation based trust management systems have been proposed to provide protection against malicious content distribution. These schemes are of two types: gossip-based and topology adaptation-based. The main drawbacks of gossip-based schemes are their high message exchange overheads and their susceptibility to misrepresentation by malicious peers. The XREP trust management intended for Gnutella system uses several rounds of message exchanges which consume high bandwidth and lack scalability. P-Grid trust model proposed by Aberer et al. uses specialized topology for the placement of data, and hence it is not suitable for unstructured P2P networks. Eigen Trust model proposed by Kamvar et al. has high computational overhead, suffers from misrepresentation by malicious peers, and cannot cope up with the transient nature of real world P2P applications. Credence protocol has high computational and communication overheads due to encryption and decryption of votes.
(ii) Search scalability – Another major drawback of P2P system is poor search scalability. Usually controlled flooding, random walker or more recently topology evolution is used to locate resource in unstructured network. However, these approaches are not very efficient. This makes efficient and scalable searching in an unstructured P2P network a challenging task.
(iii) Free Riding - Though in an ideal P2P system, the role of each peer is the same (as producer as well as consumer of resources), in reality peers are heterogeneous entities with varying capacities, most of them trying to maximize the gain from the system rather than trying to fulfill the system goal. Various trust-based incentive schemes currently existing to encourage sharing of files are not very effective in real-world deployment scenarios. Experiments conducted on Gnutella, a popular content sharing P2P network, reveal the presence of significant percentage of free riders. According to studies conducted by Saroju et al. about 70% of users share no files at all and 1% of the hosts answer nearly 50% of all the queries and the top 25% users account for 99%.

Accordingly, an effective and quick search conducting systems and methods for aggregating local trust information among the available nodes within an unstructured peer to peer communicating network to establish semantic community structures would be highly desirable. Also, desirable is a mechanism to punish the malicious peers and reward the honest peers to prevent fake content distribution in order to make the searching process more scalable.

OBJECT OF THE INVENTION:
In accordance with the present invention, a trust management framework for secure searching across an unstructured peer to peer network is provided.

It is an object of the present invention to provide efficient search mechanism free from fake content distribution or distribution of malicious/ spurious files by utilizing topology adaptation approach.

Another objective of the present invention is to prevent whitewashing across the communicating network wherein the peers leave and rejoin the system in order to avoid penalties.
It is an object of the invention to enable quick search of authentic resources by forming semantic community structures of trustworthy peers as a result of network topology approach which assists in resolving most of the queries within the community.

One of the other objectives of the present invention is to aggregate resources available at the edges of the communicating network and put them in topologically advantageous positions so that they can be shared cooperatively by the honest peers in the network.

Yet another objective of the present invention is to keep the connectivity of the network link intact by allowing the community edges to join the network but never deleting the connecting links in the original overlay network. This prevents any possibility of network partitioning due to deletion of connectivity edges.

In accordance with the aspect of the invention, the system is enabled to self adjust and adapt as the community grows to provide best performance in terms of increased query hits and reduced control message overhead in a dynamic resource sharing environment.

In another aspect of the present invention, an efficient searching penalizes malicious peers by increasing their search times and blocking the search queries initiated by them. This makes the framework robust even in the presence of large proportion of malicious peers.

Yet another aspect of the present invention rewards the good peers to share high quality files by bringing them closer in a network thereby reducing their searching time.

In one of the other aspect, the present invention enables quick and scalable search by initiating effective query propagation and reliable network topology adaptation by utilizing the least recently used data structures or initiating directed depth first search approach.

In yet another aspect, a scalable trust management system requires less computational storage and low storage overhead, optimizing simultaneously the quality of service (QoS) of the search operations.

It is another object of the present invention to ensure high quality and effective search mechanism by analyzing the performance of the proposed trust management framework by a set of defined metrics as determining factors.

One of the objectives of the present invention is to present a light weight yet efficient and robust searching operation by designing a trust management engine and evaluating its various performance metrics such as attempt ratio, effective attempt ratio, query miss ratio, hit per message, relative increase in connectivity, closeness centrality and largest connected component value.
SUMMARY OF THE INVENTION
The present invention provides a method and system, to prevent fake content distribution and free riding by malicious peers while providing an efficient and scalable searching mechanism within an unstructured peer to peer (P2P) network.

Accordingly, an overlay of trusted peers is constructed and implicit semantic community structures are established where neighboring peer nodes are selected based on trust rating information and content similarity. The invention constructs such an overlay by utilizing technology adaptation approach to combat the problem of inauthentic downloads and inefficient search in unstructured P2P networks.

The adaptive trust-aware framework used in the present invention is robust, scalable and light-weight and minimizes spurious file downloads, discourages free riding and optimizes quality of services of search operations. It increases search efficiency by taking advantages of implicit semantic community structures formed as a result of topology adaptation since most of the queries can be resolved within the community.

The search strategy is self-adjusting which adapts as community grows based upon local information only to provide best performance (in terms of increasing query hits and reduced message complexity). The system also provides incentives to share high quality files and punish malicious peers who provide fake files and discourage free riding.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure1 is a principle block diagram of a system representing trust management framework and formation of semantic community structures according to a preferred embodiment of the present invention.

Figure 2 illustrates neighbor selection process according to various embodiments of the present invention.

Figure 3 depicts a query propagation mechanism initiated by a peer for executing search according to an embodiment of the present invention.

Figure 4 is a representation of topology adaptation based on search executed in accordance with the embodiment of the present invention.

Figure 5 illustrates the attempt ratio metric for downloading authentic files for honest and malicious peers for varying percentages of malicious peers in the network (a) 10% malicious nodes, (b) 20% malicious nodes.

Figure 6 illustrates the effectiveness of effective attempt ratio metric in analyzing the authentic file downloads by honest peer relative to malicious peers.

Figure 7 is a comparative depiction of the average effective attempt ratio for authentic file download for various percentages of malicious peers with and without using the trust management framework.

Figure 8 shows Query miss ratio variation with varying percentages of malicious peers within a network.

Figure 9 represents hit per message to measure search efficiency for various search simulation cycle with 10% malicious peers in the network according to an aspect of the present invention.

Figure 10 shows variation of relative increase in connectivity with respect to varying percentages of malicious peers for threat model A in accordance with one embodiment of the present invention, (a) 20% malicious nodes, (b) 40% malicious nodes

Figure 11 shows variation of relative increase in connectivity with respect to varying percentages of malicious peers for threat model B in accordance with another embodiment of the present invention (a) 20% malicious nodes, (b) 40% malicious nodes.

Figure 12 is a graph representing largest connected components assessing strength of community links formed for different content categories (a) 10% malicious nodes, (b) 30% malicious nodes.

Figure 13 presents the comparison of the average largest connected components for the current invention and an equivalent network graph with varying percentages of malicious peers in accordance with the embodiment of the present invention.

Figure 14 is a graphical representation of assessing closeness centrality for various percentages of malicious peers in the network (a) 20% malicious nodes, (b) 40% malicious nodes.

Figure 15 shows average shortest path length for honest and malicious nodes over generations of search for various percentages of malicious peers according to an embodiment of the present invention (a) 20% malicious nodes, (b) 40% malicious nodes.

Figure 16 depicts the fluctuations in topology adaptation overhead for varying percentage of malicious peers over generations of search (a) 20% malicious nodes, (b) 40% malicious nodes.

DETAILED DESCRIPTION OF THE INVENTION

The following description is presented to enable any person skilled in the art to make and use the invention, and is provided in the context of a particular application and its requirements. Various modifications to the disclosed embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein. As one skilled in the art will appreciate, most digital computer systems can be installed with the present invention. To the extent that a particular computer system configuration is programmed to implement the present invention, it becomes a digital computer system within the scope and spirit of the present invention. That is, once a digital computer system is programmed to perform particular functions pursuant to computer-executable instructions from program software that implements the invention described heretofore, it in effect becomes a special purpose computer particular to the present invention. The necessary programming-related techniques are well known to those skilled in the art and thus are not further described herein for the sake of brevity.

The present invention presents a trust management framework for peer to peer (P2P) networks that utilizes topology adaptation to minimize distribution of spurious files. It also reduces search time since most of the queries are resolved within the community of trustworthy peers. The invention provides an efficient searching means for good peers while it penalizes the malicious peers by increasing their search times. The invention provides a framework that is robust even in presence of large percentage of malicious nodes in a network.

In the embodiment of the present invention, the concept of topology adaptation is utilized to combat the problem of inauthentic downloads and inefficient search in unstructured peer to peer (P2P) networks. The invention uses an adaptive trust-management framework that is robust, scalable and light-weight and minimizes spurious file downloads, discourages free riding and optimizes quality of services of search operations.

The concept of topology adaptation has been used for efficient search in P2P networks by forming semantic communities of peers. The word ‘community’ here refers to a group of peers having contents belonging to same category of interest where peers of a certain community will have dense connection among them but between different communities connections are sparse. The community based search has been presented in the present invention to enhance the scalability of the search mechanism and to reduce network bandwidth occupancy of the peer to peer applications, so that users can conduct effective search and download authentic files rapidly to gain useful information.

Referring to Fig. 1, the formation of semantic community structures as a result of topology adaptation within an unstructured peer to peer (P2P) network, according to a preferred embodiment of the present invention is presented. The referral figure shows the effect of trust management and semantic community on a peer to peer (P2P) overlay. The peers sharing the same contents are shown with crosses while the free-riders are shown in white colors. The peers sharing a particular content type are marked with same color. Within each content category there are malicious peers who provide fake files, shown with crosses.

According to an embodiment of the present invention, the ideal trust management framework segregates good peers from the malicious ones as shown in Fig. 1. However, it does not bring the peers sharing similar content on the neighborhood of each other to form interest-based communities. Semantic communities adapt topology to form cluster of peers sharing similar contents, but they ignore trustworthiness of peers and hence do not punish malicious nodes. The current invention does both the functions. It adapts the topology to bring good peers with matching interest to the vicinity of each other as illustrated with dotted lines in Fig. 1. The peers who provide good community service are rewarded and offered topologically better positions. In addition, the malicious peers and free riders are punished and banished to fringe of the communicating network.

The embodiment of the present invention is a framework which supports peer to peer network formed of computing devices designed to permit users to share resources and download authentic files in a real time environment. Resources can include without being limited to: data, digital files, memory, processing power, and storage space while a computing device would be understood by a worker skilled in the art to include any electronic device with storage and computing capability and a communication means with which to communicate with other computing devices.

The topology of the network, as illustrated, plays an important role for the analysis of trust management and search procedure. The network has been modeled as a power law graph. In a power law network, degree distribution of nodes follows power law distribution, i.e. fraction of nodes having degree L is L-k where k is a network dependent constant. Prior to each simulation cycle a fixed fraction of peers chosen randomly is marked as malicious. As the search mechanism proceeds, the peers adjust topology locally to connect those peers which have better chance to provide good files in future and drop malicious peers from their neighborhood.

The network links, for the purposes of present invention, are categorized into two types: connectivity link and community link. The connectivity links are the edges of the original power law network which provide seamless connectivity among the peers. To prevent the network from being fragmented they are never deleted. On the other hand, community links are added probabilistically between the peers who know each other. A community link may be deleted when perceived trustworthiness of a peer falls in the perception of its neighbors. A constricted limit is put on the additional number of edges that a node can acquire to control bandwidth usage and query processing overhead in the network. This increase in network load is measured relative to the initial network degree (corresponding to connectivity edges). Let final_degree(x) and initial_degree(x) be the initial and final degree of a node x. The relative increase in connectivity (RIC) is constrained by a parameter known as edge_limit.

In accordance with the preferred embodiment of the present invention, the semantic community structures are established based on trustworthiness of connecting nodes. The dynamics of a P2P network are highly dependent on the volume and variety of files each peer chooses to share. Hence a model reflecting real-world P2P networks is required. It has been observed that peers are in general interested in a subset of the content on the P2P network. Also, the peers are often interested only in files from a few content categories. Among these categories some are more popular than others. In case of the present invention, both content categories and file popularity within each category is modeled with zipf distribution with α = 0.8.

The content distribution model proposed by Schlosser et al. is followed for simulation purpose. In this model, each distinct file fc ,r is abstractly represented by the tuple (c, r), where c represents the content category to which the file belongs, and r represents its popularity rank within a content category c. Let content categories be C = {c1, c2,…,c32}. Each content category is characterized by its popularity rank. That is, rank of c1 = 1, c2 = 2, c3 = 3 etc meaning that c1 is more popular than c2 and hence replicated more than c2 and so on. Also there are more files in category c1 than c2.

The above model is described in Table 1 with the help of a hypothetical scenario wherein the content distribution is shown among the peer nodes.
Peers Content categories
P1 {C1, C2, C3}
P2 {C2, C4, C6, C7}
P3 {C2, C4, C7, C8}
P4 {C1, C2}
P5 {C1, C5, C6}

Each peer randomly chooses between 3 to 6 content categories to share files and shares more files in more popular categories. Table 1 shows a fictitious content distribution for illustration purpose. The category c1 is more replicated as it is most popular. The Peer 1 shares files in three categories: c1, c2, c3 where it shares maximum number of files in category c1, followed by category c2 and so on. On the other hand, Peer 3 shares maximum number of files in category c2 as it’s the most popular among the categories chosen by it, followed by c4 and so on.

It is observed that the peers usually query for files that exist on the network and are in the content category of their interest. In each cycle of simulation, active peers issue queries. However number of queries a peer issues may vary from peer to peer, modeled by Poisson distribution as follows. If M is the total number of queries to be issued in each cycle of simulation and N is the number of peers present in the network, query rate is the mean of the Poisson process. The expression gives the probability that a peer issues K queries in a cycle. The probability that a peer issues query for the file fc,r depends on the peer’s interest level in category c and rank r of the file within that category.
The present invention presents a trust management engine which is entrusted with the task of obtaining trust rating information of peer nodes from past transaction history as well as recommendation from its neighbor. It allows a peer to join the network with default trust level and gradually build its reputation by providing good files to other peers. In the proposed scheme, each peer maintains a least recently used (LRU) data structure to keep track of recent transactions with almost 32 peers at a time. Each time peer i downloads a file from peer j, it rates the transaction as positive (trij=1) or negative (trij=-1) depending on whether downloaded file is authentic or fake. is the fraction of successful downloads peer i had made from peer j, where TD is the total number of downloads. Peer i considers peer j as trustworthy if , and malicious if . If , peer i considers peer j is average trustworthy. Peer i may seek recommendations from other peers about peer j only when information is not locally available. It is the trust management engine which makes the proposed scheme robust and light-weight.

The proposed peer to peer network in the present system is transient in nature. A large number of peers join and leave the network at any time. This activity is termed as node churning. To simulate node churning, prior to each generation (a set of consecutive searches), a fixed percentage of nodes are chosen randomly as inactive. These peers neither initiate nor respond to a query in that generation and join the system latter with their LRU structure cleared. Since in a real world network, even in presence of churning, the approximate distribution of content categories and files remain constant, content of nodes undergoing churn is exchanged which in effect assigns each of them new content as well as keeps content distribution model of the network unchanged.

Malicious peers adopt various strategies, popularly known as threat model to conceal their behavior and disrupt system activity. Two threat models are considered in the proposed scheme. The peers who share good quality files enjoy better topological due to topology adaptation. In threat model A, malicious peers attempt to circumvent this by providing good file occasionally with probability, known as degree of deception to lure other peers to form communities with them. In threat model B, a group of malicious peer joins to the system and provides good files until their connectivity reaches to edge limit, and then start spreading fake content in the network.

Therefore, the preferred embodiment of the present invention presents a highly scalable, robust and light weight trust management framework for optimizing searching operation across an unstructured peer to peer network utilizing network topology adaptation approach to segregate trusted peers from malicious ones and forming content based semantic community structures of peers in a real time networking environment by querying for desired content , wherein the said framework comprises of a trust management engine that utilizes trust rating information derived from maintained data structures to dynamically adapt the network topology by probabilistically inserting community edges among trusted peers of structured semantic communities to a constricted limit for reduced bandwidth usage while keeping the connectivity edges in original overlay network unchanged to avoid network fragmentation.

The network overlay learns trust information through the search and updates trust information derived using the trust management engine and adapts topology based on the outcome of the search. The following criteria are kept in mind while designing the protocol: (a) It should improve search efficiency as well as search quality by way of authentic file downloads. (b) It should have minimal overhead in terms of computation, storage and message passing. (c) It should provide incentive to share large number of high quality files. (d) It should be self policing in the sense that a peer can adjust search strategy based on local estimate of network connectivity.

The major steps therefore involved in designing the trust aware protocol using the trust management framework are: (i) search, (ii) trust checking and (iii) topology adaptation. Each of these steps is discussed in detail as follows:

The method initiates with the establishment of the network topology environment by way of connecting the computing devices with the trust management framework.
(i) Searching: The time to live (TTL) bound search is used which evolves along with topology. At each hop, query is forwarded to a fraction of neighbors, the number of neighbors is decided based on the local estimate of network connectivity, known as probability of community formation, Probcom, computed at the query initiating node, using relative increase in degree. It is defined at node x as follows:

When Probcom is low, the peer has the capacity to accept new community edges and expand community structures. Higher the value of Probcom, lesser the neighbors choose to disseminate queries. As simulation proceeds, connectivity of good nodes increases and reaches a saturation level. So, they focus on directing queries to appropriate community which may host the specific file rather than expanding communities. For example, if peer i can contact at most 10 neighbors and Probcom of j is 0.6, it forwards query to 10 x (1 – 0.6) = 4 neighbors only. The search strategy modifies itself from initial TTL limited BFS (breadth first search) to directed DFS (depth first search) with the restructuring of the network.

The search is carried out in two steps– query initiation and query forward.
A. Query initiation: Initiating peer node forms a query packet containing the name of the file (c, r) and forwards it to a fixed fraction of neighbors along with Probcom and Time to live (TTL) value. The query is disseminated using the following neighbor selection rule. The neighbors are ranked based on both trustworthiness and the similarity of interest. Preference is given to the trusted neighbors sharing similar contents. Among the trusted neighbors, community members having content matched to the query are preferred. When there is insufficient number of community links, query is forwarded through connectivity links also.

The various cases of neighbor selection are illustrated in Figure 2. The community edges and the connectivity edges are drawn with solid and dotted lines respectively. The nodes that dispatch the query are shaded. It is assumed that in each case only two neighbors are selected. When the query (c2, f4) reaches node P, following cases may occur. In first case, P has adequate community neighbors sharing file in category c2, hence they are chosen. In Case 2 of Figure 2, there is insufficient number of community neighbors sharing file in the requested category, the community neighbors sharing c2 and c6 preferred to the connectivity neighbor c2 to forward query. In Case 3 of Figure 2, only one community neighbor who share file is c2. Hence it is chosen. From the remaining connectivity neighbors, most trusted c6 is selected. In Case 4 of Figure 2, only connectivity neighbor are there, assuming all of them at the same trust level, the matching neighbor c2 is chosen and from the rest c5 is selected randomly.
When a query reaches peer i from peer j, following actions are performed by peer i. These actions constitute the query forward step. The query initiation protocol can be summarized as:

Query Initiation:
1) Form a query packet containing tuple (c,r)and Time To Live value.
2) Calculate number of neighbors to whom to forward query using Probcom
3) Select neighbors using Neighbor Selection Rule.
4) Forward query packets to the neighbor selected.

When a query reaches for peer i to peer j, the following actions are performed by the peer I which constitute query forward step.

B. Query forward: (i) First check the trust level of peer j: Peer i checks trust rating of peer j through a check trust rating mechanism executed by the trust management engine. Trust rating is used at various stage of the search mechanism to make decision about possible download source, to stop a query forwarded from a malicious node and to adapt topology. A least recently used (LRU) data structure is used at each peer to keep track of 32 most recent peers it has interacted with. When no transaction history is available, a peer seeks recommendation from its neighbors using trust query. When peer i doesn’t have trust score of peer j in its LRU history, it first seeks recommendation about j from all of its community neighbors. If none of its community neighbors possesses any information about j, peer i initiates directed depth first search (DFS) search.

Accordingly decision regarding further propagation of the query is taken. (ii) Check the availability of file: If the requested file is found, response is sent to peer j. If TTL value has not expired, the following steps are executed. (iii) Calculate the number of messages to be sent: It is calculated based on the value of Probcom. (iv) Choose neighbors: Neighbors are then chosen in using the neighbor selection rule.

The search process for obtaining the desired content file is shown in Figure 3. It is assumed that the query is forwarded at each hop to two neighbors. The matching community links are preferred over connectivity links to dispatch query. Peer 1 initiates query and forward it to two community neighbors 3 and 4. The query reaches peer 8 via peer 4. However, peer 8 knows that peer 4 is malicious from previous transactions. Hence it blocks the query. The query forwarded by peer 5 is also blocked by peer 10 and 11 as both of them know that peer 5 is malicious. The query is matched at four peers: 4, 6, 9 and 13.
The query forward protocol is presented is summarized below.

Query Forward:
1) Input the query packet containing tuple (c.r) from peer j
2) Computationally checking for the availability of file if locally available: A Boolean operation returns true if (c,r) is available.
3) Calculate trust score of peer j.
4) If score <=0 then
5) Exit
6) Else
7) If available (c,r ) then
8) Send response to peer j
9) End if
10) TTL TTL---1
11) If TTL>=0 then
12) Calculate number of neighbors using the formula (1-Probcomm)XN where N is the total number of edges
13) Selecting Neighbors using Neighbor Selection Rule.
14) Forward query packets to the neighbors selected.
15) End if
16) End if

Once the query is initiated and forwarded among the peer nodes, the availability of desired content file is checked. The Responses are sorted by the initiating peer i based on the reputation of resource providers and peer having highest reputation is selected as source of download. The requesting peer checks the authenticity of downloaded file. If the file is found to be fake, peer i attempts to download from other sources until it finds the authentic resource or no more sources exist and updates the trust rating and possibly adapts topology after failed or successful download, to bring trusted peers to its neighborhood and to drop malicious peers from its community.

The restructuring of network is controlled by a parameter known as degree of rewiring which is the probability with which a link is formed between two peers. This parameter allows trust information slowly to be propagated as happens in real network. Topology adaptation consists of the following operations: (i) link deletion: Peer i deletes the existing community link with peer j if it finds peer j as malicious. (ii) link addition: Peer i probabilistically forms community link with peer j if resource is found to be authentic. If, for both peers i and j, only then an edge can be added subject to the approval of resource provider j. If peer j finds that peer i is malicious, it doesn’t approve the link. The restructuring of network topology by bringing honest nodes topologically in a better position for quick communication while banishing the fake or malicious nodes at the fringes of the network link follows a defined protocol described as follows:

Topology adaptation (as executed by initiator i)
1) A set S containing responses
2) Sort S based on the trust score of resource provider
3) Repeat
4) Download file from source j having highest trust score among the rest
5) Check file status and update trust score of source
6) Rewire network (I,j)
7) Until file is authentic or no more source exist in S

The rewiring of network is a controlling parameter for topology adaptation which is executed by following a series of steps defined below:

Rewire Network
1) i is initiator, j is source provider
2) if source j is malicious then
3) Delete existing community edge (i.j)
4) Else
5) R= rand (0,1)
6) If rand≤ Probcomm. Then
7) If Approve link (I,j) then
8) Form community edge(i,j)
9) End if
10) End if
11) End if

The approval of community link in the rewiring of network link further follows a sequence of determining steps as follows:
Approve Link
1) i is initiator, j is resource provider
2) Return: Boolean Value
3) If RIC (i) or RIC (j)≥ edge_ limit then
4) Return 0
5) Else if score (i)<0 then
6) Return 0
7) Else
8) Return 1
9) End if

In the example shown in Figure 4, peer 1 downloads the file from peer 4 and finds that the file is spurious. It reduces the trust score of peer 4 and deletes the community link 1-4. It then downloads the file from peer 6 and gets an authentic file. Peer 1 now sends a request to peer 6, and the latter grants the request after consulting its Least recently used (LRU) data structure and the community edge 1-6 is added. The malicious peer 4 loses one community link and peer 6 gains one community edge. However, the network still remains connected by connectivity edges, shown in dotted lines.

It is therefore established that how the preferred embodiment of the present invention utilizes the concept of topology adaptation for detecting false content distribution, free riding by malicious peers and also achieves scalability of searching process for large communicating networks.

The advantageous embodiment of the present invention constructs an overlay of trusted peers using a trust management framework where the neighbor peers are selected based on trust rating and content similarity. It increases the search efficiency by taking advantage of implicit semantic community structures formed as a result of topology adaptation wherein most of the queries can be resolved within the community. The search strategy is self-adjusting which adapts as the community grows based upon local information only to provide best performance in terms of increased query hits and reduced message complexity.

The method used for locating the resources dynamically using the technically designed trust management framework assures that the links in the original power law network never gets deleted while the peer nodes reorient themselves in a topologically better position and the network never gets partitioned or fragmented. This provides for reduced bandwidth occupancy and low storage overheads which places the invention in a better position to solve the technical problem of efficient utilization of computational resources within large communicating networks.

In accordance with one of the preferred embodiments of the present invention, a method for optimizing searching operation by querying the peer nodes contained on a computing device connected to the trust management framework across an unstructured peer to peer network, wherein the said method comprising the processor implemented steps of:
establishing a network of computing devices connected to the trust management framework;
automatically generating the search query packet at an initiating node along with the time to live value;
iteratively performing the task of forwarding the defined number of said search query packets within a network link to selected number of neighboring peer nodes determined by the neighbor selection rule if the time to live value is greater than the predetermined value, wherein the said neighbors are pre-selected based on content based similarity;
computationally evaluating the trust value of selected neighboring peer node from maintained data structures to determine the need for query propagation;
sorting the response received from the trusted neighboring peer node based on the gained reputation from previous transactions and accordingly updating the trust rating information; and
dynamically restructuring the topology of peer nodes based on the sorted responses by probabilistically inserting the community edges among the trusted nodes and deleting the fake nodes from network link while keeping the connectivity edges in original overlay network unchanged to avoid network fragmentation.

Also, the trust management framework used in the present invention is robust, scalable and lightweight which makes the system ideally suitable for deployment in real –world unstructured peer to peer networks. The P2P system encompasses a broad set of distributed applications which allow sharing of computer resources (documents, music, videos etc) by direct exchange between systems. The goal of a P2P system is to aggregate resources available at the edge of the Internet and to share it cooperatively among users. Specifically, the file sharing P2P systems have become popular as a new paradigm for information exchange among large number of users in the Internet. They are more robust, scalable and fault-tolerant, and offer better availability of resources. In unstructured P2P networks, placement of content is unrelated to topology, as opposed to structured P2P networks and under dynamic environment, unstructured P2P networks perform better than their structured counterparts. The current invention provides an efficient and scalable searching mechanism that also prevents false content distributions and free ridings by malicious peers by way of employing an incentive scheme for honest peers to share high quality good files and blocking the queries initiated by malicious peers. Hence, it has an immense potential for applications.

Most significantly, the present invention takes the advantage of semantic community formation to improve Quality of Searches for locating the resources without flooding the peer nodes with queries, as the case with existing techniques of APT or RCATP which lacks an effective mechanism to disseminate trust information in large networks. Further, the techniques of least recently used data structures and depth first search makes the process more scalable and yet robust.

Finally, one of the exemplary embodiments of the present invention provides various performance metrics such as attempt ratio (AR), effective attempt ratio (EAR), query miss ratio (QMR), hit per message (HM), relative increase in connectivity (RIC), closeness centrality (CC), largest connected component (LCC) etc., to evaluate the performance of the current invention, show that it is a very light-weight yet efficient and robust searching scheme for unstructured P2P networks.

However, to analyze the performance of the proposed trust management framework to execute the process of efficient and scalable searching, a set of metrics are defined that measures the efficiency and quality of searches. Each of these metrics is defined below.

(1) Attempt ratio (AR): A peer keeps on downloading the file from various sources, based on their trust rating till it gets the authentic file. AR is the probability that in the first attempt, the authentic file is downloaded. Ideally, AR should be high.

(2) Effective attempt ratio (EAR): The malicious peers may also increase their search quality in order to hide their true nature. Hence, the search quality achieved by good peers should be measured relative to that provided by malicious peers. EAR measures the cost of downloading an authentic file by good peers relative to malicious peers. EAR is measured against various percentage of malicious peers and the percentage of malicious peers for which EAR drops to zero is noted. If P(i) be the total number of attempts made by peer i to download an authentic file, EAR is given by:

M and N are the number of good and malicious peers issuing queries in a particular generation. For example, EAR= 50 means that if a good peer, on the average, needs one attempt to download an authentic file, a malicious peer needs two attempts.

(3) Query miss ratio (QMR): Since there is no semantic community initially, and a low value of TTL is used for file searching, there will be a high rate of query misses in the first few generations of search. However, as the search process continues, the query miss is expected to fall down for good peers. For malicious peers the rate of decrease of query miss will be very slow since queries from malicious peers are blocked. QMR is defined as the ratio of the number of search failures to the total number of searches in a generation.

(4) Hit per message (HM): Due to the formation of semantic community, number of messages required to get a hit is expected to fall down. HM measures the search efficiency achieved by the proposed system and method, and is defined as the number of query hit per message irrespective of the authenticity of the file being downloaded.

(5) Relative increase in connectivity (RIC): After a successful download, a requesting peer attempts to connect to the resource provider by forming a community edge if approved by the resource provider. This ensures that peers providing good community services are rewarded by having increasing number of community neighbors. The metric RIC measures the number of community neighbors a peer gains relative to its connectivity neighbors in the initial network topology. If Dinit(i) and Dfinal(i) are the initial and final degrees of the peer i, and N is the number of peers, then RIC for peer i may be computed using : . Due to the incentive scheme in the proposed mechanism, the connectivity of good peers is expected to increase significantly with time.

(6) Closeness centrality (CC): Since the topology adaptation effectively brings the good peers closer to each other, the length of the shortest path between a pair of good peer decreases. This intrinsic incentive for sharing authentic files is measured by the metric CC. The peers with higher CC values are topologically better positioned. If Pij is the length of the shortest path between peer i and peer j through the community edges, and if V denotes the set of peers, then CC for peer i is given by:

(7) Largest connected components (LCC): The community edges connect peers with similar content interest and having mutual trust on each other. If we consider the peers which share a particular category of contents, then the community edges form a trust-aware community overlay. However, it will be highly probable that the trust-aware overlay graph will be a disconnected graph. LCC of the network can be taken as a measure of the goodness of community structure since it signifies how strongly the peers with similar contents and interests are connected with each other. It is expressed in terms of the percentage of nodes of a particular category that lies within the LCC.
One of the other preferred embodiment of the present invention, in view of the above defined metric units as determining factor provides a method of automatically evaluating the performance of a trust management framework configured to implement optimized search operation across an unstructured peer to peer network wherein the said method comprising the processor implemented steps of:
defining the set of determinant factors relevant to the process of query execution and restructuring of the network topology;
automatically initiating the search query packet at an initiating node along with the time to live value and accounting of the rate of query miss per simulation cycle;
disseminating the defined number of said search query packets within a network link to selected number of neighboring peer nodes determined by the neighbor selection rule if the time to live value is greater than the predetermined value, wherein the said neighbors are pre-selected based on content based similarity;
computationally evaluating the trust value of selected resource providing neighboring peer node and assessing the connectivity of resource providing community;
sorting the response received from the trusted neighboring resource providing peer node reporting the search quality achieved by honest peer nodes relative to malicious peers;
dynamically restructuring the topology of peer nodes based on the sorted responses and defined determinant factor by probabilistically inserting the community edges among the trusted nodes and deleting the fake nodes from network link while keeping the connectivity edges in original overlay network unchanged to avoid network fragmentation; and
estimating the connecting strength of semantic community of resource providing nodes formed as a result of restructuring of network topology and the associated overhead load .

A discrete time simulator written in C is used for simulation. In simulation, 6000 peer nodes, 18000 connectivity edges, 32 content categories are chosen. The degree of deception and the degree of rewiring are taken as 0.1 and 0.3 respectively. The value of the edge_limit is taken as 0.3. The TTL values for Breadth first search (BFS) and Depth first search (DFS) are taken as 5s and 10 s respectively. The discrete time simulator simulates the search process repeatedly on the power law network and outputs all the metrics averaged over generations. Barabasi-Alabert generator is used to generate initial power law graphs with 6000 nodes and approximately 18000 edges. The number of search per generation is taken as 5000 while the number of generations per cycle of simulation is 100.

To check the robustness of the proposed system against attack from malicious peers, the percentage of malicious peers is gradually increased. Figure 5 illustrates the cost incurred by each type of peers to download authentic files. As the percentage of malicious peers is increased, cost incurred by malicious peers to download authentic files decreases while that of good peers increases. This is illustrated in Figure 6 using EAR. As evident from Figure 6, with 10% malicious peers in the network, EAR is 80; i.e., on the average, if a good peer needs one attempt to download an authentic file, a malicious peer needs 5 attempts to do so. Up to 60% malicious peers, good peers have higher probability to get an authentic file in the first attempt than malicious peers. So the proposed system can withstand against malicious peers till the percentage of malicious peers is below 60.

The performance of the proposed trust management system is compared with an equivalent power law network with no such trust management system. Since the proposed topology adaptation approach allows for addition of community edges, to keep the number of edges in both the networks equal, addition of edges between similar nodes are done in the equivalent network.

Figure 7 shows the comparison of the average EAR values. In the network without trust management, EAR drops to zero with 50% malicious nodes. However, in the network with trust management, even with 60% malicious peers, EAR is maintained at 20. This clearly demonstrates the robustness of the proposed trust management system.

Further, Figure 8 shows Query Miss Ratio (QMR) experienced by both types of peers for varying percentages of malicious peers in the network. Initially, QMR is high as no interest-based communities are formed and the searching is essentially a blind one. As the simulation proceeds, peers with similar content interests come closer to each other and queries are forwarded through the community edges. As a result, QMR drops for good peers. It is observed from Figure 8 that steady state value of QMR for good peers is less than 0.2, and QMR is independent of the percentage of malicious peers. This is a significant performance achievement of the proposed system. For malicious peers, the steady state value of QMR is 0.4. The high QMR for malicious peers is due to the fact that the queries forming the malicious peers are blocked by the good peers. Thus the proposed approach effectively rewards the peers who share large number of good files.

Figure 9 shows variation of Hit per Message (HM) for both types of peers. Though HM for good peers reaches a steady state as a result of topology adaptation, for malicious peers, it fluctuates. HM for malicious peers are sometimes higher than good peers. It is due to the fact that the queries forwarded by malicious peers are blocked resulting in their higher HM values. The hit here does not mean authentic hit. The authentic hit of good peers is higher than that of malicious peers as they have higher AR.

Figure 10 shows the variation of Relative Increase in Connectivity (RIC) for each type of peers under threat model A. It may be observed that RIC for the good peers increases to 2.4 constrained by the edge limit, whereas for malicious peers, RIC does not increase beyond 1.2. With the increase in the percentage of malicious peers, the saturation rate slows down but the final value remains the same. This shows that the proposed system provides better connectivity to the peers who share large number of authentic files and isolates the malicious peers.

Figure 11 shows variation of RIC under threat model B. Since in this scenario, a malicious peer provides fake files when it has achieved higher connectivity independently and stops acting maliciously after it has lost sufficient number of edges, fluctuation in RIC persists throughout the simulation.

Figure 12 depicts the size of Largest Connected Components (LCC) for each of the 32 content categories. It may be observed that the average size of LCC for all content categories remains even if the percentage of malicious peers increases. This clearly shows that the community formation among the honest peers is not affected by the presence of malicious nodes in the network. In Figure 8, the average value of LCC in the proposed scheme is compared with that of an equivalent graph for various percentages of malicious peers. While there is a minor variation in the average LCC for the proposed scheme, the average value of LCC of equivalent network is found to fall rapidly with the increase in number of malicious peers. This shows the effectiveness of the rewiring of network in the proposed scheme in community formation among the peers.

Figure 13 presents how the closeness centrality (CC) of good and malicious peers varies in the community topology. In computation of CC, only the community edges have been considered. It may be observed that the steady state value of CC for honest peers is around 0.12, irrespective of the percentage of malicious peers in the network. However, for the malicious peers, the CC value is found to lies between 0.03 to 0.07. This demonstrates how effectively the malicious peers are driven to the fringe of the network while the good peers are rewarded by providing them topologically better positions and allowing them to form content-based communities.

Figure 14 presents how the closeness centrality (CC) of good and malicious peers varies in the community topology. In computation of CC, only the community edges have been considered. It may be observed that the steady state value of CC for honest peers is around 0.12, irrespective of the percentage of malicious peers in the network. However, for the malicious peers, the CC value is found to lie between 0.03 to 0.07. This demonstrates how effectively the malicious peers are driven to the fringe of the network while the good peers are rewarded by providing them topologically better positions and allowing them to form content-based communities.

Higher values of CC also indicate that good peers have smaller average shortest path length between them. In the simulation, the diameter of the initial network is taken as 5. At the end of one simulation run, if there is no path between a pair of peers using community edges, then the length of the shortest path between that pair is assumed to be arbitrarily long, say 15 (used in Figure 15). As shown in Figure 15, the average shortest path distance (ASPD) decreases form the initial value of 15 for both honest and malicious nodes. However, the rate and the extent of decrease for honest peers are much higher due to the formation of semantic communities. For malicious peers, after an initial fall, the value of ASPD increases consistently and finally almost reaches the maximum value of 15. On the other hand, the average value of ASPD for honest peers is observed to be around 6. Since the honest nodes are connected with shorter paths, the query propagations and their responses will also be faster among these nodes. This justifies the smaller value of TTL used in the simulation.

Finally, the overhead due to the topology adaptation mechanism is analyzed. The load due to topology adaptation is measured by the metric topology adaptation overhead (TAO), which is the number of community edges added or deleted in a generation. Figure 16 shows the variation of TAO for different percentages of malicious peers. It is observed that TAO starts falling from an initial high value and oscillates with small amplitudes. This is due to the fact that initially the edge capacities of peers are not saturated and they acquire community edges rapidly. As the simulation proceeds, good peers acquire relatively stable neighborhood resulting in steep decrease in TAO. For higher values of generations, TAO fluctuates slightly since good peers delete existing edges with malicious peers as soon as the malicious peers are discovered, and acquire new community edges with good peers. With increase in percentage of malicious peers, fluctuations of TAO also increase as there is more addition and deletion of community edges. However, in any cases, TAO becomes drastically low once the community topology becomes matured. This shows that the proposed system introduces negligible overhead to the system.

In conclusion, the system can support more efficient and scalable resource searching operation across a peer to peer network due to its adaptive, robust and light weight network structure.

The embodiments of the present invention may include receiving, sending or storing instructions and/or data implemented in accordance with the foregoing description upon a carrier medium. Generally speaking, a carrier medium may include storage media or memory media such as magnetic or optical media, e.g., disk or CD-ROM, volatile or non-volatile media such as RAM (e.g. SDRAM, DDR SDRAM, RDRAM, SRAM, etc.), ROM, etc. as well as transmission media or signals such as electrical, electromagnetic, or digital signals, conveyed via a communication medium such as network and/or a wireless link. Various modifications and changes may be made as would be obvious to a person skilled in the art having the benefit of this disclosure. It is intended that the invention embrace all such modifications and changes and, accordingly, the above description to be regarded in an illustrative rather than a restrictive sense.

WE CLAIM

1) A scalable, robust and light weight trust management framework for optimizing searching operation across an unstructured peer to peer network utilizing network topology adaptation approach to segregate trusted peers from malicious ones and forming content based semantic community structures of peers in a real time networking environment by querying for desired content, wherein the said framework comprises of a trust management engine that utilizes trust rating information derived from maintained data structures to dynamically adapt the network topology by probabilistically inserting community edges among trusted peers of structured semantic communities to a constricted limit for reduced bandwidth usage while keeping the connectivity edges in original overlay network unchanged to avoid network fragmentation.

2) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein, the unstructured peer to peer network is modeled as a power law graph and degree distribution of communicating nodes follows power law distribution.

3) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the network topology is dynamically adapted by orienting the good peers having better chances to provide authentic files to come closer and pushing the malicious peers at the fringes of network.

4) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the connectivity links are the edges of the original power law network which provide seamless connectivity among the peer nodes.

5) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the community links are amenable to changes and are deleted when perceived trustworthiness of peer nodes obtained from the trust rating information, is not above a threshold value.

6) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the community edges are inserted probabilistically between the trusted peer nodes to a constricted limit which is measured by determining the relative increase in connectivity from the ratio between initial and final degree of a connecting node.
7) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the semantic community structures are formed based on content distribution modeled as zipf distribution with α = 0.8.

8) A scalable, robust and light weight trust management framework as claimed in claim 7, wherein the content distribution between the semantic communities is characterized by content category from which the authentic file is downloaded and the popularity rank associated with said content category.

9) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the number of queries that can be initiated by the peer node to determine content category is modeled by Poisson distribution.

10) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the trust management engine computes trust rating information of each peer node by maintaining least recently used data structures to keep track of past transactions.

11) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the trust management engine rates the peer node as trust worthy if the fraction of successful transactions between the nodes is less than 0.5.

12) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the trust management engine rates the peer node as malicious if the fraction of successful transactions are less than 0.5.

13) A scalable, robust and light weight trust management framework as claimed in claim 1, wherein the selected inactive nodes are allowed to join the network only when their least recently used data structure record is made available.

14) A method for optimizing searching operation by querying the peer nodes contained on a computing device connected to the trust management framework across an unstructured peer to peer network, wherein the said method comprises the processor implemented steps of:
establishing a network of computing devices connected to the trust management framework;
automatically generating the search query packet at an initiating node along with the time to live value ;
iteratively performing the task of forwarding the defined number of said search query packets within a network link to selected number of neighboring peer nodes determined by the neighbor selection rule if the time to live value is greater than the predetermined value, wherein the said neighbors are pre-selected based on content based similarity;
computationally evaluating the trust value of selected neighboring peer node from maintained data structures to determine the need for query propagation;
sorting the response received from the trusted neighboring peer node based on the gained reputation from previous transactions and accordingly updating the trust rating information; and
dynamically restructuring the topology of peer nodes based on the sorted responses by probabilistically inserting the community edges among the trusted nodes and deleting the fake nodes from network link while keeping the connectivity edges in original overlay network unchanged to avoid network fragmentation.

15) A method for optimizing searching operation, as claimed in claim 14, wherein the query packet contains the desired content based file, time to live value and probability value of community formation.

16) A method for optimizing searching operation, as claimed in claim 14, wherein the number of query packets to be forwarded is determined by the probability of community formation.

17) A method for optimizing searching operation, as claimed in claim 14, wherein the network link can be either of community link or connectivity link based on the content similarity of nodes with the initiated query packet.
18) A method for optimizing searching operation, as claimed in claim 14, wherein the restructuring of network topology is controlled by a probability value governing the link between trusted peer nodes.

19) A method for optimizing searching operation, as claimed in claim 14, wherein the fake nodes are deleted from the community link by pushing them at the fringes of community network link.

20) A method for optimizing searching operation, as claimed in claim 14, wherein the depth first search is conducted by dissemination of trust query in case of non availability of previous transaction in the maintained data structures of the peer node.

21) A time efficient method of automatically evaluating the performance of a trust management framework configured to implement optimized search operation across an unstructured peer to peer network wherein the said method comprises the processor implemented steps of:
defining the set of determinant factors relevant to the process of query execution and restructuring of the network topology;
automatically initiating the search query packet at an initiating node along with the time to live value and accounting of the rate of query miss per simulation cycle;
disseminating the defined number of said search query packets within a network link to selected number of neighboring peer nodes determined by the neighbor selection rule if the time to live value is greater than the predetermined value, wherein the said neighbors are pre-selected based on content based similarity;
computationally evaluating the trust value of selected resource providing neighboring peer node and assessing the connectivity of resource providing community;
sorting the response received from the trusted neighboring resource providing peer node reporting the search quality achieved by honest peer nodes relative to malicious peers;
dynamically restructuring the topology of peer nodes based on the sorted responses and defined determinant factor by probabilistically inserting the community edges among the trusted nodes and deleting the fake nodes from network link while keeping the connectivity edges in original overlay network unchanged to avoid network fragmentation; and
estimating the connecting strength of semantic community of resource providing nodes formed as a result of restructuring of network topology and the associated overhead load .

22) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the defined determining factors for the query processing includes attempt ratio, effective attempt ratio, query miss ratio, hit per message and relative increase in connectivity.

23) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the defined determining factor for the restructuring of network topology includes closeness centrality which measures the incentive to be given for sharing desired content file by resource provider nodes.

24) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the first attempt ratio determines the probability of receiving the desired content file from the resource providing node in the network.

25) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the rate of query miss per simulation cycle determines the number of search failures by computing the ratio between the query misses and number of query propagated for search.

26) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the connectivity of resource providing community is calculated from the ratio between the initial and final degree of initiating node to determine number of community members joining the network.

27) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the resource providing nodes restructure themselves based on the average shortest path distance between the said nodes, as determined by calculating the closeness centrality value for quicker search.

28) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the connecting strength of semantic community formed as a result of restructuring of network topology is assessed by the largest connected component factor which expresses the percentage of nodes sharing a similar content category.

29) A time efficient method of automatically evaluating the performance of a trust management framework as claimed in claim 21, wherein the overhead load is estimated by stability of topology adaptation overhead factor which determines the number of resource providing community edges added or deleted.