CN101902459B - Trust selection method and system for nodes in P2P network by applying P4P - Google Patents

Abstract

The invention relates a trust selection method and system nodes in a P2P (peer-to-peer) network by applying P4P. The method comprises the following steps: step 1, grouping the nodes in the P2P network by applying P4P; step 2, presenting the evaluation value of each node upon the completion of the interaction between the node and the other node, and reporting the evaluation value to the management node in the group to which the node belongs; step 3, calculating the level of trust of the virtual node corresponding to the group on the virtual nodes in the network according to the received evaluation value by regarding each group as the virtual node by the management node; step 4, acquiring the strategy matrix via the P4P by the management node, and adjusting the strategy matrix by using the level of trust, so that the higher the level of trust of the virtual node is, the higher the selection proportion corresponding to the virtual nodes in the strategy matrix is; and step 5, selecting the nodes for interaction according to the adjusted selection proportion obtained by the nodes. The invention can increase the level of trust of the group and further increase the probability that the nodes are selected for interaction in the group, thus reducing the network risk.

Description

Node trust selection method and system thereof in the P2P network of application P4P

Technical field

The present invention relates to computer network field, relate in particular to node trust selection method and system thereof in the P2P network of using P4P.

Background technology

Along with the development of P2P (peer-to-peer, end-to-end) technology, P2P uses has become one of main application on the Internet, especially aspect file transfer and streaming media service.But P2P uses to take in a large number between backbone network flow and the nodes and lacks trusting relationship, has restricted further developing that P2P uses.

P4P (Proactive Provider Participation in P2P, operator initiatively adds among the P2P) be a kind of existing art of up-to-date proposition, it utilizes ISP (ISP) that network is monitored, network is divided into different territories, and then utilizes the strategy matrix guiding node to select.The division in territory, simple method can consideration of regional, and complicated method need to be considered the factors such as the network bandwidth, Link State.The generation of strategy matrix, simple method can only consider can provide in each territory the number of the node of service, complicated also needs to consider the factors such as the network bandwidth, Link State.In the time of providing the node of respective service abundant in the territory, strategy matrix instructs node selection as much as possible territory interior nodes mutual, and the overseas node of the least possible selection is mutual, thereby greatly reduces the taking of backbone network flow, and has promoted the development of P2P network.But the P2P network of using P4P remains the P2P network in essence, still lacks trusting relationship between the node in the network, has the behavior that selfish node is uncooperative, the mutually deception of being unwilling between resource uploading and node provides invalid even harmful resource.

Solve the method that lacks trusting relationship between the P2P nodes in the prior art and be divided into two kinds: a kind of PKI of being based on (Public Key Infrastructure, PKIX) trust method of technology, another kind are based on the each other trust method of trust evaluation of node.

Utilize central server to come node users in the supervising the network based on the trust method of PKI technology, need to be in central server registration and authentication when each node users adds P2P network.Afterwards, central server is provided a certificate to each node users.When the requesting node user applies for the certificate of oneself to be issued the receiving node user when mutual, whether the receiving node user utilizes corresponding encryption and decryption technology to judge whether this certificate is legal after receiving certificate, and then determine mutual with it.The requesting node user adopts to use the same method and judges receiving node user's legitimacy.This method realizes simple, convenient management.But shortcoming is, because the center relies on, exist the problem of performance bottleneck and single point failure, and the method is uncontrollable to the user behavior that adds network.

Trust method based on trust evaluation is estimated mutually by the node of mutual mistake in the network, calculates the degree of belief of node according to these evaluation informations, and node is selected the high node of trust value when selecting mutual node, and then reduces network risks, obtains better service.This method takes full advantage of node self in the network and mutually retrains separately behavior, for show always good node to good evaluation, give and poor evaluation for the bad node of performance.At present, existing trust evaluation model can be divided into two kinds: the model of ungrouped model and grouping.Ungrouped typical model is: EigenTrust model and evidence model.The EigenTrust model calculates local degree of belief according to the transactions history of node, and passes through the iteration of mutual trust degree between neighbours, is that each node in the network calculates a unique global trusting value.Node is collected first neighbor node to the evaluation information of respective nodes in the evidence model, then utilizes D-S evidence composition rule.Wherein, EigenTrust model such as document The EigenTrust algorithm for reputationmanagement in P2P networks, Proceedings of the 12 ^ThDescribed in InternationalConference on World Wide Web (WWW ' 03) .2003:640-651, evidence model such as document An evidential model of distributed reputation management, described in Proceedings of the ACM International Conference Autonomous Agents andMulti-Agent Systems (AAMAS ' 02) .2002:82-93, D-S evidence composition rule such as document A Mathematical Theory of Evidence, Princeton NJ:PrincetonUniversity Press is described in the 1976:10-28.Calculate the trust value of respective nodes.A kind of typical thought of grouping model is the trust value height according to node, is divided into different groups according to certain gradient, and in the computing node trust value, the evaluation of other nodes and the evaluation of self calculate the trust value of respective nodes in the utilization group.

Above-mentioned trust evaluation model directly uses can not satisfy two challenges that P4P brings to the foundation of faith mechanism in the P2P network of using P4P: first, the bad behavior of territory interior nodes will affect the generation of P4P strategy matrix, and then the misleading node selects more node mutual in the poor territory of service ability, and in the good territory of service ability, select on the contrary node still less mutual, faith mechanism is not considered strategy matrix is revised in the prior art, reduces the node bad behavior to the impact of strategy matrix; The second, the impact of a flow that divided into groups, node is in the good behaviour of grouping interior nodes and not mean that node also can show when mutual between the territory good, needs to distinguish the calculating of trust value between the trust value and grouping intermediate node between the same grouping interior nodes.

Summary of the invention

For addressing the above problem, the invention provides node trust selection method and system thereof in the P2P network of using P4P, the strategy matrix that obtains according to P4P by the evaluation of estimate adjustment that provides between application node, can be so that the degree of belief of the grouping of node be higher, selected to carry out mutual probability larger for node in this grouping, thereby reduce network risks.

The invention discloses the node trust selection method in the P2P network of a kind of P4P of application, comprising:

Step 1 is used P4P the P2P nodes is divided into groups;

Step 2, in the P2P network, each node provides the evaluation of estimate to described another node after finishing alternately with another node, and described evaluation of estimate is reported the interior management node of place grouping;

Step 3, management node is grouped into empty body node with each, calculates empty body node corresponding to place grouping to the degree of belief of empty body node in the network according to the evaluation of estimate that the place grouping interior nodes that receives reports;

Step 4, described management node is by P4P acquisition strategy matrix, determine that empty body node corresponding to place grouping is to the degree of belief of all empty body nodes in the network, use described degree of belief and adjust described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger;

Step 5, the selection percentage after the management node that node divides into groups from the place is adjusted selects every group of interior high node of degree of belief to carry out alternately according to described selection percentage.

Described step 2 further is,

Step 21, node are judged alternately whether success after finishing alternately with another node, according to more corresponding being successfully completed mutual number of times or not being successfully completed mutual number of times of peer node in the timing statistics of new record of described judgement;

Step 22 is calculated as follows evaluation of estimate corresponding to peer node,

S _Ij-μ F _Ij＜0 or S _Ij=0 o'clock, p _Ij=0

p _IjThat described node i is to the evaluation of estimate of peer node j, S _IjThat described node i is successfully completed mutual number of times, F from peer node j in the timing statistics _IjBe that described node i is not successfully completed mutual number of times from peer node j in the timing statistics, μ is the size preset greater than 1 penalty factor;

Step 23, described node reports described management node with evaluation of estimate.

Described step 3 further is,

Step 31, described management node is grouped into empty body node with each, and the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _iTo the evaluation of estimate of empty body node in the network,

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{p}_k$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kFor described management node in timing statistics, receive for empty body node D _jK evaluation of estimate of the node in the corresponding grouping;

Step 32, described management node is with empty body node D _iEvaluation of estimate to Dummy node in the network is carried out normalization by following formula, and income value is empty body node D _iTo the degree of belief of Dummy node in the network,

$D{T}_{\mathrm{ij}}=\frac{D{P}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}D{P}_{\mathrm{ik}}}$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is by empty body node D _iProvide total number of the empty body node of evaluation of estimate.

Described step 4 further is,

Step 41, described management node is by P4P acquisition strategy matrix, and described strategy matrix is expressed as,

sp ₁₁ sp ₁₂ … sp _1n

sp ₂₁ sp ₂₂ … sp _2n

… … … …

sp _n1 sp _n2 … sp _nn

Wherein, empty body node D _iThe ratio vector be [sp _I1Sp _I2Sp _In], element is empty body node D in the ratio vector _iNode in the corresponding grouping is selected the selection percentage of node in each grouping;

Step 42, described management node determine that the place divides into groups corresponding empty body node to the degree of belief of all empty body nodes in the network, and described degree of belief is formed the degree of belief vector of described empty body node, are expressed as [DT _I1DT _I2DT _In], element is empty body node D in the degree of belief vector _iDegree of belief to each empty body node;

Step 43 with the corresponding element addition in the ratio vector sum degree of belief vector of same empty body node or multiply each other, obtains the new vector of described empty body node;

Step 44 is carried out the new ratio vector that normalization obtains described empty body node with each element in the new vector of described empty body node.

Described step 5 further is,

Step 51, the selection percentage after the management node that node divides into groups from the place is adjusted;

Step 52, node determine to select the number of node from each grouping according to described selection percentage;

Step 53, node comprise respectively the tabulation of the node that service can be provided of respective number to the management node request of each grouping;

Step 54, node carries out with the node in the described tabulation after obtaining tabulation alternately.

Described step 5 further is,

Step 61, the selection percentage after the management node that node divides into groups from the place is adjusted;

Step 62, described node determine to select the number of node from each grouping according to described selection percentage;

Step 63, described node can provide respectively the tabulation of the node of service to the management node request of each grouping, select the node of corresponding number from described tabulation;

Step 64, described node carries out with the node of selecting alternately.

From described tabulation, select the node of corresponding number further to be in the described step 63,

Step 71, described node calculates described node to the trust value of the selected node in the grouping at described place when selecting node from the grouping at place;

Step 72, described node calculates described node to the trust value of the selected node in the grouping at described non-place when selecting node from the grouping at non-place;

Step 73 by described trust value order from high to low, is selected the node of corresponding number from the tabulation of each grouping.

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping at described place further is in the described step 71,

Step 81, described node in the management node of each grouping obtains timing statistics each grouping node the evaluation of estimate of selecteed node is obtained described node to the trust value of the node that provides evaluation of estimate from record;

Step 82, described node are calculated as follows the trust value of node in the grouping of place from the management node of each grouping obtains that node provides each grouping in timing statistics the number to the evaluation of estimate of selected node,

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-N_1}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}$

Wherein, t _IjThat described node i is to the trust value of described selected node j; P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i is to providing evaluation of estimate P to node j in the grouping of place _KjThe trust value of node; P _LjBy l the evaluation of estimate of grouping exterior node in node i place to node j in the timing statistics; T _IlNode i to the place grouping outer node j is provided evaluation of estimate P _LjThe trust value of node; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes, N ₂The number to node j evaluation of estimate that provides for the node outside timing statistics interior nodes i place grouping; LIM is interaction times threshold value in the grouping of presetting.

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping of described place further is in the described step 71,

Step 91, described node sends the trust value to node in the grouping of management node place of record to the management node of each grouping, and asks described management node to be calculated as follows the place grouping to the reference trust value of selected node,

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m}$

g _MjEmpty body node D _mCorresponding grouping is to the reference trust value of selected node j, T _Ik ^mThat node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate P _Kj ^mThe trust value of node; P _Kj ^mBy empty body node D in the timing statistics _mK the evaluation of estimate to node j that corresponding grouping interior nodes provides; N _mEmpty body node D in timing statistics _mThe number to node j evaluation of estimate that corresponding grouping interior nodes provides;

Step 92, described node receive the reference trust value for selected node that each grouping management node calculates, and form with reference to the trust value vector with reference to trust value, are expressed as [g _1jg _2jG _IjG _Wj], w is the number of dividing into groups in the network;

Step 93, the management node that described node divides into groups from the place obtains the evaluation of estimate number of the selected node that node provides the grouping of place in timing statistics, be calculated as follows the trust value of described selected node,

Wherein, t _IjThat described node i is to the trust value of selected node j; g _IjThat the grouping of node i place is to the reference trust value of node j; g _KjTo divide into groups to the reference trust value of node j except other of node i place grouping; N ₁Be the number to node j evaluation of estimate that provides in timing statistics interior nodes i place grouping interior nodes; LIM is interaction times threshold value in the grouping of presetting.

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping at described non-place further is in the described step 72,

Step 101, described node in the management node of each grouping obtains timing statistics each grouping node the evaluation of estimate of selecteed node is obtained described node to the trust value of the node that provides evaluation of estimate from record;

Step 102, described node are calculated as follows the trust value of selected node,

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-(N_1+N_3)}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}+e^{\frac{-N_1}{\mathrm{LIM}}}(1-e^{\frac{-N_3}{\mathrm{LIM}}})\frac{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}{P}_{\mathrm{rj}}}{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}}$

Wherein, t _IjThat described node i is to the trust value of selected node j; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes; N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping; N ₃It is the number to the evaluation of estimate of node j that node during other divide into groups except node j place grouping and the grouping of node i place in timing statistics provides; P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i provides evaluation of estimate P _KjThe trust value of node; P _LjL the evaluation of estimate to node j that is provided by node j place grouping interior nodes in the timing statistics, T _IlThat node i is to providing evaluation of estimate P _LjThe trust value of node; P _RjR the evaluation of estimate to node j that is provided by node in other groupings except the grouping of node i place and the grouping of node j place in the timing statistics; T _IrThat node i is to providing evaluation of estimate P _RjThe trust value of node; LIM is interaction times threshold value in the grouping of presetting.

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping at described non-place further is in the described step 72,

Step 111, described node sends the described node of record to the trust value of node in the described grouping to the management node of each grouping, and asks described management node to be calculated as follows the place grouping to the reference trust value of selected node,

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{n_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m}$

g _MjEmpty body node D _mCorresponding grouping is to the reference trust value of selected node j, T _Ik ^mThat node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate P _Kj ^mThe trust value of node; P _Kj ^mEmpty body node D _mCorresponding grouping interior nodes is to k the evaluation of estimate of node j; N _mEmpty body node D in timing statistics _mCorresponding grouping interior nodes is to the number of node j evaluation of estimate;

Step 112, described node receive the reference trust value for selected node that each grouping is calculated, and form with reference to the trust value vector with reference to trust value, are expressed as [g _1jg _2jG _IjG _Wj], w is the number of dividing into groups in the network;

Step 113, described node obtains the evaluation of estimate number of the selected node that node provides each grouping from the management node of each grouping in timing statistics, be calculated as follows the trust value of described selected node,

Wherein, t _IjThat described node i is to the trust value of the selected node j in the grouping of non-place; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes; N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping; N ₃It is the number to the evaluation of estimate of node j that node outside dividing into groups in timing statistics interior nodes j place grouping and node i place provides; g _IjThat the grouping of node i place is to the reference trust value of node j; g _JjThat the grouping of node j place is to the reference trust value of node j; g _KjTo divide into groups to the reference trust value of node j except other of node i and node j place grouping; LIM is interaction times threshold value in the grouping of presetting.

Described step 3 further is,

Step 121, described management node is collected each node to the trust value of other nodes from the grouping of place;

Step 122, described management node is grouped into empty body node with each, and the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _iTo the evaluation of estimate of empty body node in the network,

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{t}_k{p}_k$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kBe empty body node D _iNode in the corresponding grouping report for empty body node D _jThe evaluation of estimate of the node in the corresponding grouping; t _kRepresent empty body node D _iNode in the corresponding grouping reports evaluation of estimate p _kThe time, other nodes are to the mean value of the trust value of the node that reports in the described grouping;

Step 123, described management node is with empty body node D _iEvaluation of estimate to Dummy node in the network is carried out normalization by following formula, and income value is empty body node D _iTo the degree of belief of Dummy node in the network,

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}}$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is empty body node D in the timing statistics _iProvide total number of the empty body node of evaluation of estimate.

The node that the invention also discloses in the P2P network of a kind of P4P of application is trusted selective system, comprising: node and management node, and use P4P the P2P nodes is divided into grouping;

Described node comprises evaluation module and selection module, and described management node comprises degree of belief computing module and strategy matrix adjusting module,

Described evaluation module is used for providing the evaluation of estimate to described another node after finishing alternately with another node, and described evaluation of estimate is reported the interior management node of place grouping;

Described degree of belief computing module is used for being grouped into empty body node with each, calculates empty body node corresponding to place grouping to the degree of belief of empty body node in the network according to the evaluation of estimate that the place grouping interior nodes that receives reports;

Described strategy matrix adjusting module, be used for by P4P acquisition strategy matrix, determine that empty body node corresponding to place grouping is to the degree of belief of all empty body nodes in the network, use described degree of belief and adjust described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger;

Described selection module, the selection percentage after being used for being adjusted from the management node of place grouping selects to carry out mutual node according to described selection percentage.

Beneficial effect of the present invention is, the strategy matrix that obtains according to P4P by the evaluation of estimate adjustment that provides between application node can be so that the degree of belief of the grouping of node be higher, and selected to carry out mutual probability larger for node in this grouping, thus the minimizing network risks; Differentiation is treated between the grouping interior nodes and trust value calculating between the grouping exterior node, when the grouping interior nodes is calculated trust value, and the evaluation information of outstanding grouping interior nodes, the interference of eliminating grouping exterior node; When the grouping intermediate node was calculated trust value, the evaluation information of the intermediate node of emphasizing to divide into groups was to reduce with the conspiracy deceptive practices in the group node.

Description of drawings

Fig. 1 is the node trust selection method flow chart in the P2P network of application of the present invention P4P;

Fig. 2 uses P4P to the schematic diagram of P2P network packet;

Fig. 3 is the structure chart that the node in the P2P network of application of the present invention P4P is trusted selective system.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail.

Node trust selection method flow process in the P2P network of application P4P of the present invention as shown in Figure 1.

Step S100 uses P4P the P2P nodes is divided into groups.

Territory according to P4P is divided, and the P2P network is divided into different groupings.Which grouping node specifically belongs to is divided by ISP (ISP).

Use P4P to the schematic diagram of P2P network packet as shown in Figure 2, each node is divided in the corresponding grouping in the network.Whole network is divided into two-layer, and ground floor is the one-level physical layer, is comprised of the node in the P2P network; The second layer is the empty body layer of secondary, is comprised of empty body node.Empty body node is a grouping of dividing in the P2P network, and each empty body node comprises a plurality of P2P network nodes, consists of dummy node 1 such as grouping among Fig. 21, and grouping 2 consists of dummy node 2, and grouping 3 consists of dummy node 3.

Step S200, in the P2P network, each node provides the evaluation of estimate to described another node after finishing alternately with another node, and described evaluation of estimate is reported the interior management node of place grouping.

Management node is node or the interior nodes node the highest to its degree of belief of grouping of the best performance of selection from grouping, and simultaneously for avoiding single point failure, selecting in grouping has most backup management nodes.

There is the multiple method that provides mutual Node evaluation value in the prior art, such as, as evaluation of estimate, claim that with the real resource that obtains from respective nodes or service and this node the ratio of the resource that can provide or service is as the evaluation of estimate method with the ratio of the difference of interaction success number of times and the frequency of failure between the node and interaction success number of times.

Be implemented as follows described among the present invention the method for the evaluation of estimate of egress.

After node is finished alternately, be success or failure alternately in the local node judgement, successful then increase that peer node provides service successful number of times for local node in the timing statistics, otherwise, the interior peer node of timing statistics provides serv-fail for local node number of times increased.For the result of failure, think that the peer node performance is bad, need to punish peer node, provide relatively relatively poor evaluation of estimate, with the behavior of restraint joint P2P application.

Be calculated as follows

S _Ij-μ F _Ij＜0 or S _Ij=0 o'clock, p _Ij=0

p _IjThat node i is to the evaluation of estimate of node j, S _IjThat timing statistics interior nodes i is successfully completed mutual number of times, F from node j _IjBe that timing statistics interior nodes i is not successfully completed mutual number of times from node j, μ is the penalty factor greater than 1.Node i is local node, and node j is peer node.

Applicating counter control timing statistics in the cycle of whenever finishing a timing statistics, then restarts to add up.

Local node is with evaluation of estimate p _kReport management node, p _k=p _Ij

Step S300, management node is grouped into empty body node with each, calculates empty body node corresponding to place grouping to the degree of belief of empty body node in the network according to the evaluation of estimate that the place grouping interior nodes that receives reports.

Wherein, be managed the empty body node that empty body node that node calculates degree of belief comprises the grouping correspondence at management node place, and the empty body node of the grouping correspondence of outer other of place grouping.

The embodiment one of described step S300

The empty body node of the grouping correspondence at management node place is expressed as D _i, empty body node D _iTo empty body node D _jEvaluation of estimate be calculated as follows.

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{p}_k$

DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate;

N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes, namely empty body node D in the timing statistics _iNode in the corresponding grouping is with empty body node D _jNode in the corresponding grouping carries out mutual total degree;

p _kBe empty body node D _iNode in the corresponding grouping report for empty body node D _jThe evaluation of estimate of the node in the corresponding grouping, k represents evaluation of estimate p _kBe empty body node D in timing statistics _iThe k that corresponding grouping interior nodes reports is individual for empty body node D _jThe evaluation of estimate of the node in the corresponding grouping.

With empty body node D _iTo the evaluation of estimate normalization of Dummy node in the network, income value is empty body node D _iDegree of belief to Dummy node in the network.

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}}$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is empty body node D _iProvide total number of the empty body node of evaluation of estimate.

Step S400, described management node is by P4P acquisition strategy matrix, determine that empty body node corresponding to place grouping is to the degree of belief of all empty body nodes in the network, use described degree of belief and adjust described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger.

Can obtain the following strategy matrix of form by P4P:

sp ₁₁ sp ₁₂ … sp _1n

sp ₂₁ sp ₂₂ … sp _2n

… … … …

sp _n1 sp _n2 … sp _nn

Every delegation in the strategy matrix has indicated node in grouping corresponding to the empty body node of correspondence of this row is selected nodes in each grouping selection percentage, empty body node D _iThe ratio vector be [sp _I1Sp _I2Sp _In], element is empty body node D in this ratio vector _iNode in the corresponding grouping is selected the ratio of node in each grouping.Sp for example _IjRepresent empty body node D _iNode in the corresponding grouping is at empty body node D _jSelect to carry out in all selections of this node for the quantity of mutual node the ratio of mutual node in the corresponding grouping; Sp _IiRepresent that then node selection in its place grouping is used for the ratio of all nodes that account for this node selection of mutual node.The number of dividing into groups in the network is expressed as n.

Use described empty body node the degree of belief of empty body node in the network is adjusted described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger.

Management node determines that the place divides into groups corresponding empty body node to the degree of belief of all empty body nodes in the network, and described degree of belief is formed the degree of belief vector of described empty body node, is expressed as [DT _I1DT _I2DT _In], element is empty body node D in the degree of belief vector _iDegree of belief to each empty body node.

Described management node judges whether to calculate to empty body node degree of belief, if so, and then take the degree of belief of the calculating degree of belief of corresponding empty body node to described empty body node of dividing into groups as the management node place; Otherwise, take the default default value degree of belief of corresponding empty body node to described empty body node of dividing into groups as the management node place.

For example, empty body node D _iBe [DT to the degree of belief vector _I1DT _I2DT _In], DT _IjRepresent empty body node D _iTo empty body node D _jDegree of belief, this degree of belief calculate to obtain by management node, if not to empty body node D _jCalculate, namely empty body node D _iNode is not also with empty body node D in the corresponding grouping _jNode is mutual in the corresponding grouping, empty body node D _iManagement node do not obtain for empty body node D _jThe evaluation of estimate of node, then degree of belief DT in the corresponding grouping _IjBe default default value.

Described strategy matrix method of adjustment one

Step S410 obtains one newly to heavy, [sp with the element of ratio vector and the corresponding element addition of degree of belief vector _I1+ DT _I1Sp _I2+ DT _I2Sp _In+ DT _In].

Step S420 obtains the element normalization of new vector the new ratio vector [sp after the adjustment of empty body node _I1' sp _I2' ... sp _In'], wherein the normalization of each element is undertaken by following formula,

${\mathrm{sp}}_{\mathrm{il}}^{\′}=\frac{{\mathrm{sp}}_{\mathrm{il}}+D{T}_{\mathrm{il}}}{\underset{k=1}{\overset{n}{\mathrm{\Σ}}}({\mathrm{sp}}_{\mathrm{ik}}+{\mathrm{DT}}_{\mathrm{ik}})}.$

Described strategy matrix method of adjustment two

Step S410 ', the corresponding element of the element of ratio vector and degree of belief vector multiplied each other obtains a new vector, [sp _I1DT _I1Sp _I2DT _I2Sp _InDT _In].

Step S420 ' obtains the element normalization of new vector the new ratio vector [sp after the adjustment of empty body node _I1' sp _I2' ... sp _In'], wherein the normalization of each element is undertaken by following formula,

${\mathrm{sp}}_{\mathrm{il}}^{\text{'}}=\frac{{\mathrm{sp}}_{\mathrm{il}}{\mathrm{DT}}_{\mathrm{il}}}{\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{sp}}_{\mathrm{ik}}{\mathrm{DT}}_{\mathrm{ik}}}.$

Step S500, the selection percentage after the management node that node divides into groups from the place is adjusted selects to carry out mutual node according to described selection percentage.

The execution mode one of described step S500

Selection percentage after the management node that node divides into groups from the place is adjusted; Node determines to select the number of node from each grouping according to described selection percentage; Node comprises respectively the tabulation of the node that service can be provided of respective number to the management node request of each grouping; Node carries out with the node in the described tabulation after obtaining tabulation alternately.

For example, node will select X node to carry out when mutual in network, obtains ratio vector after the adjustment from the interior management Nodes that divides into groups first, if a plurality of management nodes are arranged, then selects at random a management node to obtain the ratio vector.According to the definite number to the mutual node of each minute group selection of ratio vector after adjusting, [Xsp _I1Xsp _I2Xsp _In].Node is respectively to the tabulation of the node that service can be provided of the management node request respective number of each grouping.After node obtains tabulation, carry out alternately with the node in the tabulation.

The embodiment two of described step S500

Step S510, the selection percentage after the management node that node divides into groups from the place is adjusted.

Step S520, node determine to select the number of node from each grouping according to described selection percentage.

Step S530, node can provide respectively the tabulation of the node of service to the management node request of each grouping, select the node of corresponding number from described tabulation.

Step S540, node carries out with the node of selecting alternately.

Selection mode comprises multiple, for example selects at random, perhaps selects by the mode of step S600.

From described tabulation, select in the described step 83 to carry out mutual node further be,

Step S600, described node calculate described node to the trust value of selecteed node in the grouping of place when selecting node from the grouping at place; Described node calculates described node to the trust value of selecteed node in the grouping at described non-place when selecting node from the grouping at non-place; By described trust value order from high to low, from the tabulation of each grouping, select the node of corresponding number.

The non-place grouping of node refers to not comprise the grouping of described node.

Adopt step S600 system of selection, not only can guarantee node as much as possible in the corresponding grouping of the high empty body node of degree of belief, select mutual node, in each grouping, also need to select credible node to carry out alternately as far as possible.

It is following described that described node calculates the embodiment one of same grouping interior nodes trust value.

After described node has calculated trust value, record the trust value that calculates, and described trust value is reported management node.

Step 611, described node in the management node of each grouping obtains timing statistics each grouping node the evaluation of estimate of selecteed node is obtained described node to the trust value of the node that provides evaluation of estimate from record.

Step 612, described node are calculated as follows the trust value of node in the grouping of place from the management node of each grouping obtains that node provides each grouping in timing statistics the number to the evaluation of estimate of selected node,

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-N_1}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}$

Wherein, t _IjThat described node i is to the trust value of described selected node j;

P _KjK evaluation of estimate to node j in the grouping of timing statistics interior nodes i place;

T _IkBe to provide k to the trust value of the node of the evaluation of estimate of node j in timing statistics interior nodes i divides into groups to the place, namely node i is to providing evaluation of estimate P _KjThe trust value of node;

P _LjOuter l the evaluation of estimate to node j of timing statistics interior nodes i place grouping;

T _IlBe that timing statistics interior nodes i provides l to the trust value of the node of the evaluation of estimate of node j to the place grouping is outer, namely node i is to providing evaluation of estimate P _LjThe trust value of node;

N ₁It is the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes;

N ₂It is the number to node j evaluation of estimate that the node outside the grouping of timing statistics interior nodes i place provides;

LIM is interaction times threshold value in the grouping of presetting.

In this embodiment, the information that node will be selected node is divided into two groups, and one group is to divide into groups interior nodes to being selected the evaluation information of node, and another group is that other packet nodes are to the evaluation information of selected node.

It is as described below that described node calculates the embodiment two of same grouping interior nodes trust value.

After described node has calculated trust value, record the trust value that calculates, and described trust value is reported management node.

Step 611 ', described node sends the trust value to node in the grouping of management node place of record to the management node of each grouping, and asks described management node to be calculated as follows the place grouping to the reference trust value of selected node,

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m}$

g _MjEmpty body node D _mCorresponding grouping is to the reference trust value of selected node j, T _Ik ^mThat node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate P _Kj ^mThe trust value of node; P _Kj ^mBy empty body node D in the timing statistics _mK the evaluation of estimate to node j that corresponding grouping interior nodes provides; N _mEmpty body node D in timing statistics _mThe number to node j evaluation of estimate that corresponding grouping interior nodes provides.

Step 612 ', described node receives the reference trust value for selected node that each grouping is calculated, and forms with reference to the trust value vector with reference to trust value, is expressed as [g _1jg _2jG _IjG _Wj], w is the number of dividing into groups in the network.

Step 613 ', the management node that described node divides into groups from the place obtains the evaluation of estimate number of the selected node that node provides the grouping of place in timing statistics, be calculated as follows the trust value of described selected node,

Wherein, t _IjThat described node i is to the trust value of selected node j;

g _IjThat the grouping of node i place is to the reference trust value of node j;

g _KjTo divide into groups to the reference trust value of node j except other of node i place grouping;

N ₁It is the number to node j evaluation of estimate that provides in timing statistics interior nodes i place grouping interior nodes;

LIM is interaction times threshold value in the grouping of presetting.

For the calculating of trust value between the grouping interior nodes, specific implementation method one node obtains all evaluation informations, directly calculates the trust value of selected node by these evaluation informations; Specific implementation method two is that the management node of each grouping of node request calculates this grouping to the reference trust value of selected node according to the evaluation information of this grouping, selects node to obtain these with reference to the trust value that calculates again selected node after the trust value.Specific implementation method one alleviates the pressure of management node, and specific implementation method two is for selecting node, and the trust value that calculates selected node is more convenient.

It also is to divide into groups in selecteed node place that specific implementation method one and specific implementation method two have all considered to utilize node place grouping, this grouping, the number of evaluation information give prominence to the evaluation information of this grouping.When mutual less of node and selecteed node in this grouping, the evaluation of estimate to selecteed node that node provided during other divided into groups is considered as that node provides the evaluation of estimate of selecteed node of equal importance in the grouping with selecteed node place.And when mutual abundant of node in the selecteed node place grouping and selecteed node, node provides in other groupings to the importance reduction of the evaluation of estimate of selecteed node.Thereby, in above-mentioned formula, use predetermined threshold value LIM, when the evaluation of estimate number to selecteed node that provides in selecteed node place grouping interior nodes was less than this threshold value, the significance level to the evaluation of estimate of selecteed node that node provides in other groupings divides into groups to the evaluation of estimate of selecteed node and selecteed node place that interior nodes provides was suitable.Selecteed node place grouping interior nodes provide to the evaluation of estimate number of selecteed node during much larger than this threshold value, the node of other groupings provides to be compared with the evaluation of estimate to selecteed node that selecteed node place grouping interior nodes provides the evaluation of estimate of selecteed node, importance lightens, and almost can ignore.

Use above-mentioned computing formula, along with grouping inner evaluation value number N ₁Increase, the importance of the evaluation of estimate that the grouping interior nodes provides improves, the importance of the evaluation of estimate that the grouping exterior node provides descends, so more can estimate out really and be evaluated node to the service ability in dividing into groups, the interference that exterior node estimates of also can better avoiding simultaneously dividing into groups, because node is subjected to the impact of flow control between the P4P territory poor to the ability of grouping exterior node service, the exterior node that causes dividing into groups is estimated bad to it.Some that simultaneously, also can play opposing grouping exterior node are slandered behavior.

When node is selected node in other groupings, also to then node sorted according to trust value in the grouping separately from high to low first for the node that service can be provided calculates a trust value, in each grouping, selecting the high node of trust value to communicate at last.

When node calculated the place outer selected node trust value of grouping, what value was selected node provides service to the grouping exterior node ability.Therefore reduce the evaluation of selected node place grouping because be subjected to the control of flow between the P4P territory, the evaluation of selected node place grouping can not fine embodiment its ability of service is provided for the grouping exterior node.

The embodiment one of the node trust value that described node calculating place grouping is outer is following described.

The trust value that described nodes records is calculated.

Step 621, described node in the management node of each grouping obtains timing statistics each grouping node the evaluation of estimate of selecteed node is obtained described node to the trust value of the node that provides evaluation of estimate from record.

Step 622, described node are calculated as follows the trust value of selected node,

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-(N_1+N_3)}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}+e^{\frac{-N_1}{\mathrm{LIM}}}(1-e^{\frac{-N_3}{\mathrm{LIM}}})\frac{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}{P}_{\mathrm{rj}}}{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}}$

Wherein, t _IjThat described node i is to the trust value of selected node j;

N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes;

N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping;

N ₃It is the number to the evaluation of estimate of node j that node during other divide into groups except node j place grouping and the grouping of node i place in timing statistics provides;

P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i provides evaluation of estimate P _KjThe trust value of node;

P _LjL the evaluation of estimate to node j that is provided by node j place grouping interior nodes in the timing statistics, T _IlThat node i is to providing evaluation of estimate P _LjThe trust value of node;

P _RjR the evaluation of estimate to node j that is provided by node in other groupings except the grouping of node i place and the grouping of node j place in the timing statistics, T _IrThat node i is to providing evaluation of estimate P _RjThe trust value of node;

LIM is interaction times threshold value in the grouping of presetting.

In embodiment one, select node that information is divided into three groups, one group is that grouping interior nodes in node place is to the evaluation information of selected node, one group be selected node place packet node to the evaluation information of selected node, the 3rd group for other each packet nodes of residue to the evaluation information conduct of selected node.The grouping of corresponding normalized node i place is to the trust evaluation value of node j respectively for equal sign the right three parts in the formula, and normalized node j place group is to the trust evaluation value of node j, and other groupings of normalized residue are to the trust evaluation value of node j.

The embodiment two of the node trust value that described node calculating place grouping is outer is as described below.

The trust value that described nodes records is calculated.

Step 621 ', described node sends the described node of record to the trust value of node in the described grouping to the management node of each grouping, and asks described management node to be calculated as follows the place grouping to the reference trust value of selected node,

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m}$

g _MjEmpty body node D _mCorresponding grouping is to the reference trust value of selected node j, T _Ik ^mThat node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate P _Kj ^mThe trust value of node; P _Kj ^mEmpty body node D _mCorresponding grouping interior nodes is to k the evaluation of estimate of node j; N _mEmpty body node D in timing statistics _mCorresponding grouping interior nodes is to the number of node j evaluation of estimate.

Step 622 ', described node receives the reference trust value for selected node that each grouping is calculated, and forms with reference to the trust value vector with reference to trust value, is expressed as [g _1jg _2jG _IjG _Wj], w is the number of dividing into groups in the network.

Step 623 ', described node obtains the evaluation of estimate number of the selected node that node provides each grouping from the management node of each grouping in timing statistics, be calculated as follows the trust value of described selected node,

Wherein, t _IjThat described node i is to the trust value of the selected node j in the grouping of non-place;

N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes;

N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping;

N ₃It is the number to the evaluation of estimate of node j that node outside dividing into groups in timing statistics interior nodes j place grouping and node i place provides;

g _IjThat the grouping of node i place is to the reference trust value of node j;

g _JjThat the grouping of node j place is to the reference trust value of node j;

g _KjTo divide into groups to the reference trust value of node j except other of node i and node j place grouping;

LIM is interaction times threshold value in the grouping of presetting.

Calculating to trust value between the grouping exterior node, specific implementation method one and specific implementation method two are all so that the evaluation information importance of the node place of selecting grouping is the highest, the evaluation information importance of selected node place grouping is minimum, and the evaluation information importance of other each groupings falls between.

At N ₁Smaller, when the evaluation information of the node place grouping of selecting will be not enough to the trust value of the selected node of authentic assessment, the importance of all the other evaluation informations that respectively divide into groups of third part uprised.Only at N ₁And N ₃In the time of all smaller, the evaluation information of the selected node of second portion place grouping just is taken seriously.The trust value of selected node could be estimated out more really like this, the interference of the conspiracy deceptive practices of selected node place grouping can be reduced simultaneously.

Select between the grouping interior nodes together to select between mutual or the different grouping interior nodes alternately no matter be, can both determine more accurately the trust value of selected node between the node, after selecting node that selected node is just sorted according to trust value, selection comes the node of front, it is mutual to obtain the high node of more reliabilities, reduces network risks.

The embodiment two of described step S300

Described management node is collected each node to the trust value of other nodes from the grouping of place.

Described management node is grouped into empty body node with each, and the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _iTo the evaluation of estimate of empty body node in the network,

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{t}_k{p}_k$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate;

N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes;

p _kBe empty body node D _iNode in the corresponding grouping report for empty body node D _jThe evaluation of estimate of the node in the corresponding grouping;

t _kRepresent empty body node D _iNode in the corresponding grouping reports evaluation of estimate p _kThe time, other nodes are to the mean value of the trust value of the node that reports in the described grouping.

Described management node is with empty body node D _iEvaluation of estimate to Dummy node in the network is carried out normalization by following formula, and income value is empty body node D _iTo the degree of belief of Dummy node in the network,

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}}$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is empty body node D in the timing statistics _iProvide total number of the empty body node of evaluation of estimate.

Evaluation information after each empty body node utilizes node in the grouping corresponding with dummy node in the network of the grouping interior nodes of its correspondence mutual in the embodiment two of described step S300 can calculate this empty body node to network in the evaluation of estimate of empty body node, must consider simultaneously in the computational process that node provides when estimating the trust value of self.Consider the trust value of evaluation node self, can give prominence to that evaluation information that the high node of trust value provides is more worth believes that the confidence level of the evaluation information that the node that trust value is low provides is lower, and then reflect more accurately the trust value of empty body node.

In better method of the present invention, select as follows the management node of grouping.

In each grouping, choose trust value in the group surpass the pre-set threshold value node as management node or with node by the group of correspondence between from big to small arranged sequentially of trust value, select front K node as management node, K is default value.When the grouping interior nodes provides a certain Node evaluation value, these evaluations of estimate need to be reported respectively these management nodes.

Can choose the relatively high node of the interior trust value of group as management node.In order to guarantee not because management node leaves and causes evaluation information to report nowhere maybe can't to calculate trust value, at least choose 10 management nodes in general each group, on give the correct time evaluation information reported each management node as backup, 10 management nodes are the not online very little 1-0.5 of probability ¹⁰≈ 0.999.When a management node rolls off the production line, can choose a new management node and substitute.

Compare with the faith mechanism in the existing P2P network, the present invention has following advantage.The present invention takes full advantage of the P4P technology P2P network is divided into not same area and the characteristics that strategy matrix is provided, utilize the mode of grouping that the nodes trust value is carried out management of computing, put all nodes in the network under different grouping easily, the packet mode in other patents is more effectively simple.Because the division of network domains is that ISP by awareness network state carries out among the P4P, division more reasonable.And on node is selected, under the guidance of strategy matrix, accomplished that the territory interior nodes selects the node in the territory mutual more.On this basis, when the present invention calculates at the network node trust value, the evaluation of outstanding this group interior nodes of calculating of the mutual trust value of group interior nodes, emphasize the quality of node performance in group, the evaluation of the outstanding group of the calculating intermediate node of the mutual trust value of group intermediate node emphasizes that node is to the quality of group exterior node performance.Like this, mutual for the group interior nodes, slandering of effectively minimizing group exterior node, mutual for the group intermediate node, can effectively avoid organizing interior nodes and conspire to organizing the deceptive practices of exterior node.The present invention also treats as an empty body node to each group, and calculated trust value for it, and then the trust value of utilization group goes to adjust the pro rate that node in the strategy matrix is selected, and improves the ratio of selecting in the high group interior nodes of trust value, reduces the ratio that the low group interior nodes of trust value is selected.So that can choosing the high node of more trust values, node carries out reducing network risks alternately.

The structure of the node trust selective system in the P2P network of application P4P of the present invention is shown in Figure 3, comprising: node 200 and management node 100, and use P4P the P2P nodes is divided into grouping.

Node 200 comprises evaluation module and selects module that management node 100 comprises degree of belief computing module and strategy matrix adjusting module.

Evaluation module is used for providing the evaluation of estimate to described another node after finishing alternately with another node, and described evaluation of estimate is reported the interior management node of place grouping.

The degree of belief computing module is used for being grouped into empty body node with each, calculates empty body node corresponding to place grouping to the degree of belief of empty body node in the network according to the evaluation of estimate that the place grouping interior nodes that receives reports.

Wherein, the empty body node that the degree of belief computing module that is managed node calculates degree of belief comprises the empty body node of the grouping correspondence at management node place, and the empty body node of the grouping correspondence of outer other of place grouping.

The strategy matrix adjusting module, be used for by P4P acquisition strategy matrix, determine that empty body node corresponding to place grouping is to the degree of belief of all empty body nodes in the network, use described degree of belief and adjust described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger.

Select module, the selection percentage after being adjusted for the management node from the place grouping selects to carry out mutual node according to described selection percentage.

Wherein, management node be the high server of performance or from grouping the node as management node selected.

In the better execution mode, described evaluation module is further used at the place node after finishing alternately with another node, judge alternately whether success, according to more corresponding being successfully completed mutual number of times or not being successfully completed mutual number of times of peer node in the timing statistics of new record of described judgement; Be calculated as follows evaluation of estimate corresponding to peer node,

S _Ij-μ F _Ij＜0 or S _Ij=0 o'clock, p _Ij=0

p _IjThat described node i is to the evaluation of estimate of peer node j, S _IjThat described node i is successfully completed mutual number of times, F from peer node j in the timing statistics _IjBe that described node i is not successfully completed mutual number of times from peer node j in the timing statistics, μ is the size preset greater than 1 penalty factor; Evaluation of estimate is reported described management node.

In the better execution mode, described degree of belief computing module further is grouped into empty body node with each, and the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _iTo the evaluation of estimate of empty body node in the network,

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{p}_k$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kFor described management node in timing statistics, receive for empty body node D _jK evaluation of estimate of the node in the corresponding grouping; With empty body node D _iEvaluation of estimate to Dummy node in the network is carried out normalization by following formula, and income value is empty body node D _iTo the degree of belief of Dummy node in the network,

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}}$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is by empty body node D _iProvide total number of the empty body node of evaluation of estimate.

In the better execution mode, described strategy matrix adjusting module is further used for by P4P acquisition strategy matrix, and described strategy matrix is expressed as,

sp ₁₁ sp ₁₂ … sp _1n

sp ₂₁ sp ₂₂ … sp _2n

… … … …

sp _n1 sp _n2 … sp _nn

Wherein, empty body node D _iThe ratio vector be [sp _I1Sp _I2Sp _In], element is empty body node D in the ratio vector _iNode in the corresponding grouping is selected the selection percentage of node in each grouping; Determine that the place divides into groups corresponding empty body node to the degree of belief of all empty body nodes in the network, described degree of belief is formed the degree of belief vector of described empty body node, be expressed as [DT _I1DT _I2DT _In], element is empty body node D in the degree of belief vector _iDegree of belief to each empty body node; With the ratio vector sum degree of belief addition of vectors of same empty body node, obtain the new vector of described empty body node; Each element in the new vector of described empty body node is carried out the new ratio vector that normalization obtains described empty body node.

In the better execution mode, described strategy matrix adjusting module is further used for by P4P acquisition strategy matrix, and described strategy matrix is expressed as,

sp ₁₁ sp ₁₂ … sp _1n

sp ₂₁ sp ₂₂ … sp _2n

… … … …

sp _n1 sp _n2 … sp _nn

Wherein, empty body node D _iThe ratio vector be [sp _I1Sp _I2Sp _In], element is empty body node D in the ratio vector _iNode in the corresponding grouping is selected the selection percentage of node in each grouping; Determine that the place divides into groups corresponding empty body node to the degree of belief of all empty body nodes in the network, described degree of belief is formed the degree of belief vector, be expressed as [DT _I1DT _I2DT _In], element is empty body node D in the degree of belief vector _iDegree of belief to each empty body node; Corresponding element in the ratio vector sum degree of belief vector of same empty body node is multiplied each other, obtain the new vector of described empty body node; Each element in the new vector of described empty body node is carried out the new ratio vector that normalization obtains described empty body node.

Further in the better execution mode, described strategy matrix adjusting module is determining that the place corresponding empty body node that divides into groups is further used for judging whether having calculated to empty body node degree of belief during the degree of belief of all empty body nodes in to network, if so, then take the degree of belief of the calculating degree of belief of corresponding empty body node to described empty body node of dividing into groups as the management node place; Otherwise, take the default default value degree of belief of corresponding empty body node to described empty body node of dividing into groups as the management node place.

In the better execution mode, the selection percentage after described selection module is further used for being adjusted from the management node of affiliated node place grouping; Determine from each grouping, to select the number of node according to described selection percentage; Comprise respectively the tabulation of the node that service can be provided of respective number to the management node request of each grouping; After obtaining tabulation, carry out alternately with the node in the described tabulation.

In the better execution mode, the selection percentage after described selection module is further used for being adjusted from the management node of affiliated node place grouping; Determine from each grouping, to select the number of node according to described selection percentage; The tabulation of the node of service can be provided to the management node request of each grouping respectively, from described tabulation, select the node of corresponding number; Carry out alternately with the node of selecting.

In the better execution mode, described selection module is further used for calculating described node to the trust value of the selected node in the grouping at described place when selecting the node of corresponding number from described tabulation when selecting node from the grouping at place; When from the grouping at non-place, selecting node, calculate described node to the trust value of the selected node in the grouping at described non-place; By described trust value order from high to low, from the tabulation of each grouping, select the node of corresponding number, carry out alternately with the node of selecting.

Further in the better execution mode, described selection module also is used for the trust value that record calculates;

Described selection module is further used for obtaining in the timing statistics each grouping node to the evaluation of estimate of selecteed node from the management node of each grouping when calculating described node to the trust value of the selected node in the grouping at described place, obtains described node to the trust value of the node that provides evaluation of estimate from record; From the management node of each grouping obtains that node provides each grouping in timing statistics the number to the evaluation of estimate of selected node, be calculated as follows the trust value of node in the grouping of place,

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-N_1}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}$

Wherein, t _IjThat described node i is to the trust value of described selected node j; P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i is to providing evaluation of estimate P to node j in the grouping of place _KjThe trust value of node; P _LjBy l the evaluation of estimate of grouping exterior node in node i place to node j in the timing statistics; T _IlNode i to the place grouping outer node j is provided evaluation of estimate P _LjThe trust value of node; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes, N ₂The number to node j evaluation of estimate that provides for the node outside timing statistics interior nodes i place grouping; LIM is interaction times threshold value in the grouping of presetting.

Further in the better execution mode, described selection module also is used for the trust value that record calculates;

Described selection module is further used for sending to the management node of each grouping record when calculating described node to the trust value of the selected node in the grouping of described place affiliated node to the grouping of management node place in the trust value of node, and ask described management node to be calculated as follows the place grouping to the reference trust value of selected node

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m}$

g _MjEmpty body node D _mCorresponding grouping is to the reference trust value of selected node j, T _Ik ^mThat node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate P _Kj ^mThe trust value of node; P _Kj ^mBy empty body node D in the timing statistics _mK the evaluation of estimate to node j that corresponding grouping interior nodes provides; N _mEmpty body node D in timing statistics _mThe number to node j evaluation of estimate that corresponding grouping interior nodes provides; Receive the reference trust value for selected node that each grouping management node calculates, form with reference to the trust value vector with reference to trust value, be expressed as [g _1jg _2jG _IjG _Wj], w is the number of dividing into groups in the network; The management node that divides into groups from the place obtains the evaluation of estimate number of the selected node that node provides the grouping of place in timing statistics, be calculated as follows the trust value of described selected node,

Wherein, t _IjThat described node i is to the trust value of selected node j; g _IjThat the grouping of node i place is to the reference trust value of node j; g _KjTo divide into groups to the reference trust value of node j except other of node i place grouping; N ₁Be the number to node j evaluation of estimate that provides in timing statistics interior nodes i place grouping interior nodes; LIM is interaction times threshold value in the grouping of presetting.

Further in the better execution mode, described selection module also is used for the trust value that record calculates;

Described selection module is further used for obtaining in the timing statistics each grouping node to the evaluation of estimate of selecteed node from the management node of each grouping when calculating described node to the trust value of the selected node in the grouping at described non-place, obtains described node to the trust value of the node that provides evaluation of estimate from record; Be calculated as follows the trust value of selected node,

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-(N_1+N_3)}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}+e^{\frac{-N_1}{\mathrm{LIM}}}(1-e^{\frac{-N_3}{\mathrm{LIM}}})\frac{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}{P}_{\mathrm{rj}}}{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}}$

Wherein, t _IjThat described node i is to the trust value of selected node j; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes; N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping; N ₃It is the number to the evaluation of estimate of node j that node during other divide into groups except node j place grouping and the grouping of node i place in timing statistics provides; P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i provides evaluation of estimate P _KjThe trust value of node; P _LjL the evaluation of estimate to node j that is provided by node j place grouping interior nodes in the timing statistics, T _IlThat node i is to providing evaluation of estimate P _LjThe trust value of node; P _RjR the evaluation of estimate to node j that is provided by node in other groupings except the grouping of node i place and the grouping of node j place in the timing statistics; T _IrThat node i is to providing evaluation of estimate P _RjThe trust value of node; LIM is interaction times threshold value in the grouping of presetting.

Further in the better execution mode, described selection module also is used for the trust value that record calculates;

Described selection module is further used for sending to the management node of each grouping record when calculating described node to the trust value of the selected node in the grouping at described non-place described node to described grouping in the trust value of node, and ask described management node to be calculated as follows the place grouping to the reference trust value of selected node

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m}$

g _MjEmpty body node D _mCorresponding grouping is to the reference trust value of selected node j, T _Ik ^mThat node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate P _Kj ^mThe trust value of node; P _Kj ^mEmpty body node D _mCorresponding grouping interior nodes is to k the evaluation of estimate of node j; N _mEmpty body node D in timing statistics _mCorresponding grouping interior nodes is to the number of node j evaluation of estimate; Receive the reference trust value for selected node of the management node calculating of each grouping, form with reference to the trust value vector with reference to trust value, be expressed as [g _1jg _2jG _IjG _Wj], w is the number of dividing into groups in the network; Obtain the evaluation of estimate number of the selected node that node provides each grouping from the management node of each grouping in timing statistics, be calculated as follows the trust value of described selected node,

Wherein, t _IjThat described node i is to the trust value of the selected node j in the grouping of non-place; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes; N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping; N ₃It is the number to the evaluation of estimate of node j that node outside dividing into groups in timing statistics interior nodes j place grouping and node i place provides; g _IjThat the grouping of node i place is to the reference trust value of node j; g _JjThat the grouping of node j place is to the reference trust value of node j; g _KjTo divide into groups to the reference trust value of node j except other of node i and node j place grouping; LIM is interaction times threshold value in the grouping of presetting.

Further in the better execution mode, described degree of belief technology modules is further used for collecting each node to the trust value of other nodes from the grouping of affiliated management node place; Be grouped into empty body node with each, the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _jTo the evaluation of estimate of empty body node in the network,

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{t}_k{p}_k$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kBe empty body node D _iNode in the corresponding grouping report for empty body node D _jThe evaluation of estimate of the node in the corresponding grouping; t _kRepresent empty body node D _iNode in the corresponding grouping reports evaluation of estimate p _kThe time, other nodes are to the mean value of the trust value of the node that reports in the described grouping; With empty body node D _iEvaluation of estimate to Dummy node in the network is carried out normalization by following formula, and income value is empty body node D _iTo the degree of belief of Dummy node in the network,

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}}$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is empty body node D in the timing statistics _iProvide total number of the empty body node of evaluation of estimate.

Those skilled in the art can also carry out various modifications to above content under the condition that does not break away from the definite the spirit and scope of the present invention of claims.Therefore scope of the present invention is not limited in above explanation, but determined by the scope of claims.

Claims (11)

1. the node trust selection method in the P2P network of using P4P is characterized in that, comprising:

Step 1 is used P4P the P2P nodes is divided into groups;

Step 2, in the P2P network, each node provides the evaluation of estimate to described another node after finishing alternately with another node, and described evaluation of estimate is reported the interior management node of place grouping;

Step 3, management node is grouped into empty body node with each, calculates empty body node corresponding to place grouping to the degree of belief of empty body node in the network according to the evaluation of estimate that the place grouping interior nodes that receives reports;

Step 4, described management node is by P4P acquisition strategy matrix, after the degree of belief of determining all empty body nodes during empty body node corresponding to place grouping is to network, use described degree of belief and adjust described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger;

Step 5, the selection percentage after the management node that node divides into groups from the place is adjusted selects the high node of degree of belief to carry out alternately according to described selection percentage in every group;

Described step 3 further is:

Step 31, described management node is grouped into empty body node with each, and the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _iEvaluation of estimate to empty body node in the network:

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{p}_k,$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kFor described management node in timing statistics, receive for empty body node D _jK evaluation of estimate of the node in the corresponding grouping;

Step 32, described management node is with empty body node D _iEvaluation of estimate to empty body node in the network is carried out normalization by following formula, and income value is empty body node D _iDegree of belief to empty body node in the network:

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}},$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is by empty body node D _iProvide total number of the empty body node of evaluation of estimate;

Described step 4 further is:

Step 41, described management node is by P4P acquisition strategy matrix, and described strategy matrix is expressed as:

$\begin{array}{cccc}{\mathrm{sp}}_{11}& {\mathrm{sp}}_{12}& \·\·\·& {\mathrm{sp}}_{1n}\\ {\mathrm{sp}}_{21}& {\mathrm{sp}}_{22}& \·\·\·& {\mathrm{sp}}_{2n}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{sp}}_{n1}& {\mathrm{sp}}_{n2}& \·\·\·& {\mathrm{sp}}_{\mathrm{nn}}\end{array},$

Wherein, empty body node D _iThe ratio vector be [sp _I1Sp _I2... sp _In], element is empty body node D in the ratio vector _iNode in the corresponding grouping is selected the selection percentage of node in each grouping;

Step 42, described management node determine the place divide into groups corresponding empty body node to network in after the degree of belief of all empty body nodes, described degree of belief is formed the degree of belief vector of described empty body node, be expressed as [DT _I1DT _I2... DT _In], element is empty body node D in the degree of belief vector _iDegree of belief to each empty body node;

Step 43 with the corresponding element addition in the ratio vector sum degree of belief vector of same empty body node or multiply each other, obtains the new vector of described empty body node;

Step 44 is carried out the new ratio vector that normalization obtains described empty body node with each element in the new vector of described empty body node.

2. the node trust selection method in the P2P network of application as claimed in claim 1 P4P is characterized in that, described step 2 further is,

Step 21, node are judged alternately whether success after finishing alternately with another node, according to more corresponding being successfully completed mutual number of times or not being successfully completed mutual number of times of peer node in the timing statistics of new record of described judgement;

Step 22 is calculated as follows evaluation of estimate corresponding to peer node:

$p_{\mathrm{ij}}=\frac{S_{\mathrm{ij}}-{\mathrm{\μF}}_{\mathrm{ij}}}{S_{\mathrm{ij}}},$ S _Ij-μ F _Ij＜0 or S _Ij=0 o'clock, p _Ij=0,

p _IjThat described node i is to the evaluation of estimate of peer node j, S _IjThat described node i is successfully completed mutual number of times, F from peer node j in the timing statistics _IjBe that described node i is not successfully completed mutual number of times from peer node j in the timing statistics, μ is the size preset greater than 1 penalty factor;

Step 23, described node reports described management node with evaluation of estimate.

3. the node trust selection method in the P2P network of application P4P as claimed in claim 1 is characterized in that,

Described step 5 further is,

Step 51, the selection percentage after the management node that node divides into groups from the place is adjusted;

Step 52, node determine to select the number of node from each grouping according to described selection percentage;

Step 53, node comprise respectively the tabulation of the node that service can be provided of respective number to the management node request of each grouping;

Step 54, node carries out with the node in the described tabulation after obtaining tabulation alternately.

4. the node trust selection method in the P2P network of application P4P as claimed in claim 1 is characterized in that,

Described step 5 further is,

Step 61, the selection percentage after the management node that node divides into groups from the place is adjusted;

Step 62, described node determine to select the number of node from each grouping according to described selection percentage;

Step 63, described node can provide respectively the tabulation of the node of service to the management node request of each grouping, select the node of corresponding number from described tabulation;

Step 64, described node carries out with the node of selecting alternately.

5. the node trust selection method in the P2P network of application P4P as claimed in claim 4 is characterized in that, selects the node of corresponding number further to be in the described step 63 from described tabulation,

Step 71, described node calculates described node to the trust value of the selected node in the grouping at described place when selecting node from the grouping at place; Perhaps, described node calculates described node to the trust value of the selected node in the grouping at described non-place when selecting node from the grouping at non-place;

Step 72 by described trust value order from high to low, is selected the node of corresponding number from the tabulation of each grouping.

6. the node trust selection method in the P2P network of application P4P as claimed in claim 5 is characterized in that,

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping at described place further is in the described step 71,

Step 81, described node in the management node of each grouping obtains timing statistics each grouping node the evaluation of estimate of selecteed node is obtained described node to the trust value of the node that provides evaluation of estimate from record;

Step 82, described node are calculated as follows the trust value of node in the grouping of place from the management node of each grouping obtains that node provides each grouping in timing statistics the number to the evaluation of estimate of selected node:

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-N_1}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}},$

Wherein, t _IjThat described node i is to the trust value of described selected node j; P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i is to providing evaluation of estimate P to node j in the grouping of place _KjThe trust value of node; P _IjBy l the evaluation of estimate of grouping exterior node in node i place to node j in the timing statistics; T _IlNode i to the place grouping outer node j is provided evaluation of estimate P _LjThe trust value of node; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes, N ₂The number to node j evaluation of estimate that provides for the node outside timing statistics interior nodes i place grouping; LIM is interaction times threshold value in the grouping of presetting.

7. the node trust selection method in the P2P network of application P4P as claimed in claim 5 is characterized in that,

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping of described place further is in the described step 71,

Step 91, described node sends the trust value to node in the grouping of management node place of record to the management node of each grouping, and asks described management node to be calculated as follows the place grouping to being selected the reference trust value of node:

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m},$

g _MjEmpty body node D _mThe corresponding grouping to the reference trust value of selected node j, That node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate

The trust value of node; By empty body node D in the timing statistics _mK the evaluation of estimate to node j that corresponding grouping interior nodes provides; N _mEmpty body node D in timing statistics _mThe number to node j evaluation of estimate that corresponding grouping interior nodes provides;

Step 92, described node receive the reference trust value for selected node that each grouping is calculated, and form with reference to the trust value vector with reference to trust value, are expressed as [g _1jg _2j... g _Ij... g _Wj], w is the number of dividing into groups in the network;

Step 93, the management node that described node divides into groups from the place obtains the evaluation of estimate number of the selected node that node provides the grouping of place in timing statistics, be calculated as follows the trust value of described selected node:

Wherein, t _IjThat described node i is to the trust value of selected node j; g _IjThat the grouping of node i place is to the reference trust value of node j; g _KjTo divide into groups to the reference trust value of node j except other of node i place grouping; N ₁Be the number to node j evaluation of estimate that provides in timing statistics interior nodes i place grouping interior nodes; LIM is interaction times threshold value in the grouping of presetting.

8. the node trust selection method in the P2P network of application P4P as claimed in claim 5 is characterized in that,

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping at described non-place further is in the described step 72,

Step 101, described node in the management node of each grouping obtains timing statistics each grouping node the evaluation of estimate of selecteed node is obtained described node to the trust value of the node that provides evaluation of estimate from record;

Step 102, described node are calculated as follows the trust value of selected node:

$t_{\mathrm{ij}}=(1-e^{\frac{-N_1}{\mathrm{LIM}}})\frac{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}{P}_{\mathrm{kj}}}{\underset{k=1}{\overset{N_1}{\mathrm{\Σ}}}{T}_{\mathrm{ik}}}+e^{\frac{-(N_1+N_3)}{\mathrm{LIM}}}\frac{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}{P}_{\mathrm{lj}}}{\underset{l=1}{\overset{N_2}{\mathrm{\Σ}}}{T}_{\mathrm{il}}}{e}^{\frac{-N_1}{\mathrm{LIM}}}(1-e^{\frac{-N_3}{\mathrm{LIM}}})\frac{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}{P}_{\mathrm{rj}}}{\underset{r=1}{\overset{N_3}{\mathrm{\Σ}}}{T}_{\mathrm{ir}}},$

Wherein, t _IjThat described node i is to the trust value of selected node j; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes; N ₂The number to node j evaluation of estimate that provides for the node in timing statistics interior nodes j place grouping; N ₃It is the number to the evaluation of estimate of node j that node during other divide into groups except node j place grouping and the grouping of node i place in timing statistics provides; P _KjK the evaluation of estimate to node j that is provided by node i place grouping interior nodes in the timing statistics, T _IkThat node i provides evaluation of estimate P _KjThe trust value of node; P _LjL the evaluation of estimate to node j that is provided by node j place grouping interior nodes in the timing statistics, T _IlThat node i is to providing evaluation of estimate P _LjThe trust value of node; P _RjR the evaluation of estimate to node j that is provided by node in other groupings except the grouping of node i place and the grouping of node j place in the timing statistics; T _IrThat node i is to providing evaluation of estimate P _RjThe trust value of node; LIM is interaction times threshold value in the grouping of presetting.

9. the node trust selection method in the P2P network of application P4P as claimed in claim 5 is characterized in that,

Described step 63 also comprises, the trust value that described nodes records is calculated;

The described node of calculating to the trust value of the selected node in the grouping at described non-place further is in the described step 72,

Step 111, described node sends the described node of record to the trust value of node in the described grouping to the management node of each grouping, and asks described management node to be calculated as follows the place grouping to being selected the reference trust value of node:

$g_{\mathrm{mj}}=\frac{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m{P_{\mathrm{kj}}}^m}{\underset{k=1}{\overset{N_m}{\mathrm{\Σ}}}{T_{\mathrm{ik}}}^m},$

g _MjEmpty body node D _mThe corresponding grouping to the reference trust value of selected node j,

That node i is to empty body node D _mIn the corresponding grouping node j is provided evaluation of estimate

The trust value of node;

Empty body node D _mCorresponding grouping interior nodes is to k the evaluation of estimate of node j; N _mEmpty body node D in timing statistics _mCorresponding grouping interior nodes is to the number of node j evaluation of estimate;

Step 112, described node receive the reference trust value for selected node that each grouping is calculated, and form with reference to the trust value vector with reference to trust value, are expressed as [g _1jg _2j... g _Ij... g _Wj], w is the number of dividing into groups in the network;

Step 113, described node obtains the evaluation of estimate number of the selected node that node provides each grouping from the management node of each grouping in timing statistics, be calculated as follows the trust value of described selected node:

Wherein, t _IjThat described node i is to the trust value of the selected node j in the grouping of non-place; N ₁Be the number to the evaluation of estimate of node j that provides in timing statistics interior nodes i place grouping interior nodes; N ₃It is the number to the evaluation of estimate of node j that node outside dividing into groups in timing statistics interior nodes j place grouping and node i place provides; g _IjThat the grouping of node i place is to the reference trust value of node j; g _JjThat the grouping of node j place is to the reference trust value of node j; g _KjTo divide into groups to the reference trust value of node j except other of node i and node j place grouping; LIM is interaction times threshold value in the grouping of presetting.

10. the node trust selection method in the P2P network of application P4P as claimed in claim 5 is characterized in that,

Described step 3 further is,

Step 121, described management node is collected each node to the trust value of other nodes from the grouping of place;

Step 122, described management node is grouped into empty body node with each, and the empty body node of the grouping correspondence at described management node place is expressed as D _i, be calculated as follows empty body node D _iEvaluation of estimate to empty body node in the network:

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{t}_k{p}_k,$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be described management node in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kBe empty body node D _iNode in the corresponding grouping report for empty body node D _jThe evaluation of estimate of the node in the corresponding grouping; t _kRepresent empty body node D _iNode in the corresponding grouping reports evaluation of estimate p _kThe time, other nodes are to the mean value of the trust value of the node that reports in the described grouping;

Step 123, described management node is with empty body node D _iEvaluation of estimate to empty body node in the network is carried out normalization by following formula, and income value is empty body node D _iDegree of belief to empty body node in the network:

${\mathrm{DT}}_{\mathrm{ij}}=\frac{{\mathrm{DP}}_{\mathrm{ij}}}{\underset{k=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{DP}}_{\mathrm{ik}}},$

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is empty body node D in the timing statistics _iProvide total number of the empty body node of evaluation of estimate.

11. the node in the P2P network of using P4P is trusted selective system, it is characterized in that, comprising: node and management node, use P4P the P2P nodes is divided into grouping;

Described node comprises evaluation module and selects module that described management node comprises degree of belief computing module and strategy matrix adjusting module;

Described evaluation module is used for providing the evaluation of estimate to described another node after finishing alternately with another node, and described evaluation of estimate is reported the interior management node of place grouping;

Described degree of belief computing module is used for being grouped into empty body node with each, calculates empty body node corresponding to place grouping to the degree of belief of empty body node in the network according to the evaluation of estimate that the place grouping interior nodes that receives reports;

Described strategy matrix adjusting module, be used for by P4P acquisition strategy matrix, after the degree of belief of determining all empty body nodes during empty body node corresponding to place grouping is to network, use described degree of belief and adjust described strategy matrix, make the higher then selection percentage that empty body node is corresponding described in the described strategy matrix of degree of belief of empty body node larger;

Described selection module, the selection percentage after being used for being adjusted from the management node of place grouping selects to carry out mutual node according to described selection percentage;

Described degree of belief computing module further is embodied as:

Step 31, described degree of belief computing module is grouped into empty body node with each, and the empty body node of the grouping correspondence at the described management node place at described degree of belief computing module place is expressed as D _i, described degree of belief computing module is calculated as follows empty body node D _iEvaluation of estimate to empty body node in the network:

${\mathrm{DP}}_{\mathrm{ij}}=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{p}_k,$

Wherein, DP _IjEmpty body node D _iTo empty body node D _jEvaluation of estimate; N be the described management node at described degree of belief computing module place in timing statistics, receive to empty body node D _jThe number of the evaluation of estimate of corresponding grouping interior nodes; p _kFor the described management node at described degree of belief computing module place in timing statistics, receive for empty body node D _jK evaluation of estimate of the node in the corresponding grouping;

Step 32, described degree of belief computing module is with empty body node D _iEvaluation of estimate to empty body node in the network is carried out normalization by following formula, and income value is empty body node D _iDegree of belief to empty body node in the network:

DT _IjEmpty body node D _iTo empty body node D _jDegree of belief, M is by empty body node D _iProvide total number of the empty body node of evaluation of estimate;

Described strategy matrix adjusting module further is embodied as:

Step 41, described strategy matrix adjusting module is by P4P acquisition strategy matrix, and described strategy matrix is expressed as:

$\begin{array}{cccc}{\mathrm{sp}}_{11}& {\mathrm{sp}}_{12}& \·\·\·& {\mathrm{sp}}_{1n}\\ {\mathrm{sp}}_{21}& {\mathrm{sp}}_{22}& \·\·\·& {\mathrm{sp}}_{2n}\\ \·\·\·& \·\·\·& \·\·\·& \·\·\·\\ {\mathrm{sp}}_{n1}& {\mathrm{sp}}_{n2}& \·\·\·& {\mathrm{sp}}_{\mathrm{nn}}\end{array},$

Wherein, empty body node D _iThe ratio vector be [sp _I1Sp _I2... sp _In], element is empty body node D in the ratio vector _iNode in the corresponding grouping is selected the selection percentage of node in each grouping;

Step 42, described degree of belief computing module determine the place divide into groups corresponding empty body node to network in after the degree of belief of all empty body nodes, described degree of belief is formed the degree of belief vector of described empty body node, be expressed as [DT _I1DT _I2... DT _In], element is empty body node D in the degree of belief vector _iDegree of belief to each empty body node;

Step 43 with the corresponding element addition in the ratio vector sum degree of belief vector of same empty body node or multiply each other, obtains the new vector of described empty body node;

Step 44 is carried out the new ratio vector that normalization obtains described empty body node with each element in the new vector of described empty body node.

Images