A Method for Constructing a Non-Repudiable Protocol Based on a Three-Phase Trust Management Model and the New Protocol Constructed Thereof

By selecting peer-trusted neighbors as trusted third parties through a three-stage trust management model, and combining fair exchange and repudiation detection, the problem that non-repudiation protocols cannot rely on fully trusted third parties in the Internet environment is solved, realizing the practicality of the protocol and the ability to suppress repudiation in advance.

CN117914503BActive Publication Date: 2026-11-10NANJING AUDIT UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410130585.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-31
Publication Date
2026-11-10
Estimated Expiration
2044-01-31

AI Technical Summary

Technical Problem

Existing non-repudiation protocols cannot effectively rely on fully trusted third parties in the Internet environment, resulting in high deployment difficulty and poor practicality. Furthermore, existing weakening solutions require that all third-party security information be known in advance, which is difficult to implement in dynamic collaboration systems.

Method used

A non-repudiation protocol design method based on a three-stage trust management model is adopted. A trust selection mechanism is used to select a neighboring party with equal trust as a trusted third party. Combined with fair exchange and denial detection mechanisms, a non-repudiation protocol is constructed.

Benefits of technology

A reliable non-repudiation protocol was built in an environment without a trusted third party, which improved the applicability of the protocol in real network environments and provided the ability to suppress repudiation in advance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117914503B_ABST
    Figure CN117914503B_ABST
Patent Text Reader

Abstract

This invention discloses a method for constructing a non-repudiation protocol based on a three-stage trust management model and the new protocol constructed therefrom. Addressing the problem that the current internet lacks a centralized management authority and cannot obtain complete information about a particular entity, leading to the inability of existing non-repudiation protocols based on weakened trusted third-party solutions to function when third-party security information is incomplete, this invention utilizes the principle of trust management, which acknowledges the incompleteness of security information in the system and that security decisions require additional information provided by third parties. A method for constructing a non-repudiation protocol based on a three-stage trust management model is then developed, and a new protocol is constructed using this method.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to non-repudiation protocol technology, specifically to a method for constructing a non-repudiation protocol based on a three-phase trust management model and the new protocol constructed therefrom. Background Technology

[0002] In recent years, with the increasing number of online transactions, online fraud has also become rampant, seriously undermining the integrity and security of cyberspace. Non-repudiation protocols can bind tamper-proof data to online entities in online transactions, enabling the traceability of specific transaction activities, and are often used for evidence collection and prevention of online fraud. Due to the significant technical importance of non-repudiation protocols in preventing online fraud, academia and industry have conducted in-depth research on their design technology.

[0003] Traditional non-repudiation protocols are designed using a hypothetical, universally trusted third party to support the fair exchange logic within the protocol. However, this assumption is untenable in real-world network environments, making deployment difficult and impractical. To address this issue, non-repudiation protocol design schemes based on semi-trusted third parties, third-party consortia, third-party accountability auditing, and blockchain technologies have emerged. These schemes weaken the protocol's reliance on strong third-party trustworthiness but indirectly impose the requirement that all third-party security information be known in advance. This remains difficult to fully achieve in the current internet system: the internet has evolved into a dynamic collaborative system with no centralized management authority to rely on, making it impossible to obtain complete information about any single entity. To mitigate this problem...

[0004] (1) Based on the fact that security information in the system is incomplete and security decisions of the system need to rely on additional information provided by other parties, a non-repudiation protocol design method based on the three-stage model of trust management under the condition of missing security information is proposed. This method involves the selection of third parties based on credibility, the exchange of evidence based on fair exchange, and the credibility feedback based on denial detection. The new method is used to design the non-repudiation protocol.

[0005] (2) To evaluate the performance of the new protocol designed based on the new method, simulation experimental data was generated based on a complex network model. By calculating the existence probability of a peer trusted party, the feasibility of implementing third-party selection using the new method was demonstrated. Based on the malicious behavior injection model, and considering four cross-situations of collusion between the third party and the protocol entity and defamation, the interaction success rate and pre-repudiation inhibition of the non-repudiation protocol based on the new method and the traditional non-repudiation protocol based on a trusted third party were compared, proving the practicality of the non-repudiation protocol design method proposed in this invention.

[0006] Related work

[0007] Non-repudiation technology resolves disputes related to a specific action by exchanging evidence between interacting parties. A fair non-repudiation agreement guarantees that either each party obtains the evidence they want through the exchange, or none obtain evidence favorable to themselves. Given that parties may not be entirely honest, current practices introduce a trusted third party into the non-repudiation agreement to prevent selfish behavior during evidence exchange. This third party's intervention ensures fairness in the exchange. However, relying on a completely trustworthy third party is impractical in the open internet environment. To mitigate the reliance on third parties for trust in non-repudiation agreements, over the past 30 years, non-repudiation agreement designs based on semi-trusted third parties (STTPs), third-party consortia, third-party accountability, and blockchain have emerged.

[0008] Unlike Trust Third Party (TTP) schemes, which require complete trustworthiness, STTP schemes can engage in misconduct, but they cannot collude with any party to deceive others. These schemes typically employ verifiable cryptography and verifiable secret sharing techniques using one-way hash functions with specific algebraic properties, and their theoretical basis lies in intractable problems such as large number factorization, discrete logarithms, and graph isomorphisms. Qiu et al. proposed a secure and efficient PSI (SE-PSI) protocol on a scalable dataset using deterministic encryption and Bloom filters under a semi-trusted server. Yang Wang et al. proposed a collusion-resistant fair transaction model that can maintain the fairness of transactions even when STTP and signers collude (such as STTP providing the arbitration private key to the signer). Nedal Tahat proposed a proxy signature scheme using STTP based on the elliptic curve discrete logarithm problem, achieving the unforgeability, non-repudiation, and non-abuse-resistance of evidence. J. Cui et al. proposed a new message authentication scheme using edge servers as semi-trusted entities, group signature technology, and proxy re-encryption technology to ensure data integrity, confidentiality, and anonymity. Wu Jinxi et al. designed a new privacy-preserving fair contract signing protocol based on coin mixing technology, by introducing a semi-honest trusted third party to provide obfuscation services for participant identities, and combining it with a blind verifiable cryptographic signature scheme.

[0009] The third-party consortium-based approach aims to divide trust, breaking down the strong trust originally assigned to TTP into weak trust in multiple third parties. For example, Macia Mut-Puigserver et al. used a third-party consortium formed by voting from all members of the third party in the signed email system, ensuring that even if a small number of malicious third parties collude, they cannot undermine the fairness of the exchange. G. Ateniese et al. distributed TTP tasks to numerous low-trust online agents in the Tricert signed email system, with TTP only intervening when disputes arise. Alptekin Kupcu et al. proposed a general class of optimistic fair exchange protocols based on multiple arbitrators, whose fairness depends on the cooperation of multiple autonomous arbitrators.

[0010] The non-repudiation scheme based on third-party accountability ensures that the TTP must maintain the fairness of evidence exchange by holding TTPs accountable for violations. For example, Xinyi Huang et al. proposed an optimistic fair exchange protocol that combines third-party transparency and accountability, which can hold third parties accountable for malicious acts that undermine the fairness of the exchange.

[0011] Blockchain-based solutions aim to replace trusted third parties with the immutable data registry technology provided by blockchain. For example, M. Mut-Puigserver et al. proposed a new protocol that does not require a TTP (Trusted Transaction Processor) and uses blockchain to remove trusted third parties from confidential multi-party registration delivery protocols; Huang et al. proposed a fair three-party contract signing protocol based on blockchain primitives, which allows participants to sign contracts fairly without arbitrators and protects the privacy of contract content on a public blockchain; Yuan Boao et al. attempted to design a multi-party non-repudiation protocol using the decentralized network structure of blockchain, addressing the centralization problem of TTPs, and proposed a TTP-free multi-party non-repudiation protocol based on a public blockchain, analyzing whether it satisfies non-repudiation, fairness, and timeliness; Tian Haibo et al., based on the traditional ideas of verifiable cryptographic signatures and blind signatures, proposed a fair and confidential contract signing protocol by constructing a blind verifiable cryptographic signature system; Yu Lei et al. used blockchain technology to implement a multi-party, TTP-free protocol, providing fair contract addition, update, and deletion management functions, and establishing multi-directional trust in the blockchain network by changing the "one-way trust" model of blockchain.

[0012] The aforementioned approach reduces the protocol's reliance on strong third-party trustworthiness, but it indirectly imposes the requirement that all third-party security information be known in advance. This is still difficult to fully achieve in the current internet system: the internet has evolved into a dynamic collaborative system where there is no centralized management authority to rely on, and it is impossible to obtain all information about a particular entity. Therefore, it is undeniable that the protocol must seek a new design. Summary of the Invention

[0013] Purpose of the invention: The purpose of this invention is to address the shortcomings of existing technologies and provide a method for constructing a non-repudiation protocol based on a three-phase trust management model, as well as a new protocol constructed therefrom.

[0014] Technical solution: The present invention provides a method for constructing a non-repudiation protocol based on a three-phase trust management model, comprising the following steps:

[0015] Step 1: Employ a trust-based third-party selection mechanism for trust selection: When entity A needs to communicate with entity B, based on the trust levels of other entities, entities A and B select a nearest neighbor δ that is equally trusted by both parties in the protocol interaction to act as a trusted third party. The nearest neighbor entity δ has the highest trust level. This includes the following process:

[0016] Step 1.1: Extract connected nearest neighbors. From the common neighbors of interacting entities A and B, select qualified common neighbors based on the connectivity density (ECTD) assessment as a candidate set of peer nearest neighbors. The specific method is as follows:

[0017] First, based on the entity communication network ECN, find the neighbor node sets NSA and NSB of entities A and B respectively. If ECNS = NSA ∩ NSB, then... Then entities A and B have no common neighbors; otherwise, for each entity δ in NSA∩NSB, calculate the estimated and revaluation values ​​of the connectivity between entities A and B and δ, and set all entities whose revaluation values ​​are not less than the connectivity thresholds TECTD(A) and TECTD(B) as the connected nearest neighbor set ECNS of A and B.

[0018] Step 1.2: Selecting Peer Neighbors. Communication neighbors that are mutually trusted by both the sender and receiver are selected as peer neighbors. The specific method is as follows:

[0019] First, iterate through the connected nearest neighbor set ECNS of entities A and B. For any entity v∈ECNS, if TVmin≥|TV(A,v)-TV(B,v)|, it indicates that the system is not in a cold start state of trust evaluation. Then, select the peer nearest neighbor CNTN(A,B) of entities A and B based on the trust level. If for each entity v in ECNS(A,B), TV(A,v)=0 or TV(B,v)=0, it indicates that the system is in a cold start state of trust evaluation. Then, select the peer nearest neighbor CNTN(A,B) of entities A and B based on the connectivity density. TVmin is the trust difference threshold; TV(A,v) is the trust level of A to v, and TV(B,v) is the trust level of B to v.

[0020] Step 2: Employ a fair exchange-based evidence collection mechanism for trust-based interaction;

[0021] The selected peer neighbor δ is regarded as a trusted third party. Entity A and entity B use an existing fair and non-repudiable protocol based on a trusted third party (such as the LL protocol) to transmit information M from A to B. During the transmission of information M, the collection of interaction evidence is completed simultaneously. That is, A holds non-repudiable evidence of B's ​​receipt of information M {B's digital signature NRR, δ's digital signature con_K}, while B holds non-repudiable evidence of A's transmission of information M {A's digital signature NRO, δ's digital signature con_K}.

[0022] Step 3: Use a reliability feedback mechanism based on denial detection to conduct trust assessment;

[0023] After the information transfer between entity A and entity B is completed, based on the evidence held by A and B, a signature verification method is used to detect whether A and B have repudiated the transfer of M. The credibility of A, B and δ is evaluated and updated according to the repudiation detection results. Here, the repudiation detection includes sending repudiation detection and receiving repudiation detection.

[0024] Step 4: Based on the reliability assessment results of Step 3, the third-party selection process of Step 1 is repeated to obtain new peer neighbors, and a new round of trust interaction and trust assessment is conducted. Here, Step 4 forms a closed loop with the previous three steps, enabling the newly constructed protocol to have the ability to pre-emptively suppress repudiation.

[0025] Furthermore, the detailed method for extracting nearest neighbors in step 1.1 is as follows:

[0026] Step 1: Set the initial values ​​of the neighbor set NSA of entity A, the neighbor set NSB of entity B, and the common neighbor set ECNS of A and B to empty;

[0027] Step 2: Target each node v in the ECN of the China Unicom network sequentially. x ∈V, if <A,v x If >∈E is an edge in ECN, then v x If added to A's neighbor set NSA, <B,v x If >∈E is an edge in ECN, then v x Add it to B's neighbor set NSB;

[0028] Step 3: If the intersection of neighbor set NSA and neighbor set NSB is empty, then set the entity connectivity nearest neighbor set ECNS of A and B to empty; otherwise, if the intersection of NSA and NSB is not empty, for each entity δ in the intersection of NSA and NSB, calculate their recalculated connectivity value ECTD(A,δ) based on the obtained initial connectivity value Pre_ECTD(A,δ) of A and δ; at the same time, calculate their recalculated connectivity value ECTD(B,δ) based on the obtained initial connectivity value Pre_ECTD(B,δ) of B and δ.

[0029] If both ECTD(A,δ) and ECTD(B,δ) are true, then δ is added to the entity connectivity nearest neighbor set ECNS of A and B.

[0030] Step 4: Return the connected nearest neighbor set ECNS(A,B) of entities A and B.

[0031] Furthermore, the communication tightness ECTD assessment includes, in sequence, an initial ECTD assessment and a reassessment of the communication tightness ECTD.

[0032] The initial estimate of communication tightness (ECTD) is calculated as follows:

[0033] Step 1: Set the initial values ​​of the maximum number of interactions between any entities (maxMailNum) and the maximum number of shared neighbors (maxNghShrNm) to 0;

[0034] Step 2: For the entity communication network ECN = (V, E), calculate the number of interactions between any two nodes MailNum, and assign the maximum value among all interaction counts to maxMailNum; similarly, calculate the number of neighbor shares between any two nodes NghShrNm, and assign the maximum value among all neighbor shares to maxNghShrNm.

[0035] Step 3: For any two nodes v i ,v j For ∈V, calculate their entity communication frequency ECF(v i ,v j ) = MailNum(v i ,v j ) / maxMailNum; simultaneously, calculate their neighbor sharing rate NSR(v i ,v j )=NghShrNm(v i ,v j ) / maxNghShrNm.

[0036] Step 4: Calculate v i and v j Estimated connectivity density of ECTD in ECF weighting factor;

[0037] Step 5: Return Pre_ECTD(v i ,v j );

[0038] The calculation method for the ECTD revalued communication tightness is as follows:

[0039]

[0040] Among them, 0<α,β,χ,δ<1; P=α*Pre_ECTD(v i ,v j ), D = χ * Pre_ECTD(v i ,v j )′、T=δ*Pre_ECTD(v i ,v j )′|Pre_ECTD(v i ,v j The units are, in order, proportional, integral, differential, and time-dependent.

[0041] In the above formula, ECTD t-m (v i ,v j ), ECTD t-m+1 (v i ,v j ), ..., ECTD t-1 (v i ,v j ),ECTD t (v i ,v j ) represent v in sequence i and v j The re-estimated communication tightness obtained after the first m rounds [round tm...round t-1] and round t; component Pre_ECTD(v i ,v j ) for v i With v j The initial estimated network density obtained after the t-th round of communication; component For v i and v j The historical average reassessed connection density obtained before the t-th round of communication; component Pre_ECTD(v i ,v j )′ is vi and v j The volatility of the initial estimated connectivity tightness in round t; component Pre_ECTD(v i ,v j )′|Pre_ECTD(v i ,v j )″| for v i With v j The correction amount for the tightness of communication in round t; α, β, χ, δ are the weighting factors of the above four components in turn;

[0042] χ,δ follow the following two formulas:

[0043]

[0044]

[0045] Furthermore, the detailed method for peer-to-peer neighbor selection in step 1.2 is as follows:

[0046] Step 1: Let TV(A,v) and TV(B,v) be the reliability between v and A, and v and B, respectively;

[0047] For each entity v in ECNS(A,B), calculate the confidence distance |TV(A,v)-TV(B,v)|; select the v that minimizes the value of |TV(A,v)-TV(B,v)| to enter the peer nearest neighbor set CNTN(A,B) of A and B;

[0048] Step 2: If the system is in the cold start state of reliability assessment, that is, for each entity v in ECNS(A,B), TV(A,v)=0 or TV(B,v)=0, then it is necessary to select peer neighbors based on connectivity density: For each entity v in ECNS(A,B), calculate the connectivity density distance |ECTD(A,v)-ECTD(B,v)|; select the v that minimizes |ECTD(A,v)-ECTD(B,v)| to enter the peer nearest neighbor set CNTN(A,B) of A and B;

[0049] Step 3: Return the set of equal nearest neighbors of A and B, CNTN(A,B);

[0050] Furthermore, after the interactive communication is completed in step 3, if entity A subsequently denies having sent information M to B for some private interest, B may submit NRO, NRR, and con_K to the arbitrator; the arbitrator will then trace entity A's previous actions using the following method:

[0051] The arbitrator verifies the correctness of the con_K signature to confirm whether δ has received the sub_K submitted by entity A; and verifies the NRR and NRO to determine whether B has indeed received the ciphertext C sent by A, and whether the ciphertext C is correct.

[0052] If all the above verifications are correct, it is determined that entity A did indeed send information M to B, thus confirming that entity A denied the information exchange with entity B.

[0053] The denial detection results for the message M sent by entity A to entity B include the following four categories: A honestly claims to have sent message M; A honestly claims not to have sent message M; A dishonestly claims to have sent message M; A dishonestly claims not to have sent message M; and then a reliability assessment is performed based on the denial detection results.

[0054] Furthermore, after the interactive communication is completed in step 3, if B subsequently denies having received M sent by A for some private interest, A may submit NRO, NRR, and con_K to the arbitrator. The arbitrator will then trace B's previous actions. The tracing method is as follows:

[0055] The arbitrator verifies the correctness of the con_K signature to determine whether δ has received the NRR submitted by B, whether δ has generated evidence and published it in the public directory, and thus whether B can always obtain the key K between time point T and time point T+t; based on the verification of the NRR, the arbitrator determines whether B has received the ciphertext C sent by A and whether B is willing to obtain the key K from δ; based on the verification of the NRO, the arbitrator determines whether the ciphertext C received by B is correct.

[0056] If all the above verifications are correct, it is determined that entity B did indeed receive information M between time point T and T+t, thus confirming that entity B committed fraud in the information transmission with entity A.

[0057] The reception denial detection results of information M received by entity B from entity A include four cases: B honestly claims to have received information M; B honestly claims not to have received information M; B does not honestly claim to have received information M; B does not honestly claim not to have received information M; and then a reliability assessment is performed based on the denial detection results.

[0058] Furthermore, in the reliability assessment, A and B are protocol interaction entities, and δ is the selected peer nearest neighbor of A and B. The specific reliability assessment method is as follows:

[0059] Step 1) First, estimate the original reliability according to the following four scenarios;

[0060] Scenario 1: If entity A is arbitrated as honest, then entity A assigns a normal confidence level to δ and entity B.

[0061] If entity A is found guilty of denial in arbitration, and if entity A has indeed denied the allegations, entity A will retaliate against δ and entity B. Therefore, entity A will defame δ and entity B, assigning δ and entity B a confidence level lower than their actual performance. If entity A has not actually denied the allegations, then entity A will consider δ to have either neglected its duties, assigning δ a low confidence level; or δ to have conspired with entity B, naturally assigning δ and entity B a low confidence level.

[0062] Scenario 2: If entity B is arbitrated as honest, then entity B will assign a normal confidence level to δ and entity A.

[0063] If entity B is found guilty of denial in arbitration, and if entity B has indeed denied the allegations, entity B will retaliate against δ and entity A. Therefore, entity B will defame δ and entity A, assigning δ and entity A a confidence level lower than their actual performance. If entity B has not actually denied the allegations, then entity B will consider δ to have either neglected its duties and assigned δ a low confidence level, or δ to have conspired with entity A, naturally assigning δ and entity A a low confidence level.

[0064] Scenario 3: If entity B colludes with δ, then entity B will assign a high confidence level to δ in order to inflate the confidence level of δ.

[0065] Scenario 4: If entity A colludes with δ, then entity A will assign a high confidence level to δ in order to inflate the confidence level of δ.

[0066] Step 2) Calculate the overall reliability of δ based on the reliability evaluations of A and B for δ;

[0067] Let TV(A,δ) be the reliability rating of A for δ in step 1), TV(B,δ) be the reliability rating of B for δ, and RV(δ) be the historical comprehensive reliability of δ. Then the comprehensive reliability of δ is updated as follows:

[0068] First, update TV(A,δ) using TV(A,δ); then update TV(B,δ) using TV(B,δ); finally, update RV(δ) using RV(δ).

[0069] TV(A,δ)=κ*TV(A,δ)+λ*RV(δ)+μ*TV(A,δ)'+ξ*TV(A,δ)'*|TV(A,δ)″|;

[0070] TV(B,δ)=κ*TV(A,δ)+λ*RV(δ)+μ*TV(B,δ)'+ξ*TV(B,δ)'*|TV(B,δ)″|;

[0071]

[0072] This invention also discloses a new protocol for constructing a non-repudiation protocol based on a three-phase trust management model. The operation of the new protocol is divided into three phases:

[0073] In the first stage, entities A and B select a trustworthy entity δ based on their credibility.

[0074] In the second phase, entities A and B will select δ as a third party to interact with the protocol; the interaction consists of five processes:

[0075] A→B:f NRO ,B,L,C,NRO, that is, NRO=sS A (f NRO (,B,L,C);

[0076] A→δ:f SUB ,B,L,t A ,K,NRO,sub_K, that is, sub_K=sS A (f SUB ,B,L,t A ,K,NRO);

[0077] B→δ:f NRR ,A,L,t B ,NRO,NRR, that is, NRR=sS B (f NRR ,A,L,t B (,NRO);

[0078] B←δ:f CON ,A,B,L,T,t A ,t B ,K,NRR,con_K,that is,con_K=sS δ (f CON ,A,B,L,T,t A ,t B ,K,NRO,NRR);

[0079] A←δ:f CON ,A,B,L,T,t A ,t B ,K,NRR,con_K;

[0080] The message M to be sent is divided into two parts: the encryption key K and the ciphertext C; A It is a time period defined by A, indicating that δ has saved sub_K in its own private directory for t. A If the NRR submitted by B is not received after a certain number of time units, then sub_K will be deleted from its private directory. BIt is a time period defined by B, indicating that δ has saved NRR in its own private directory for t. B If sub_K submitted by A is not received after a certain number of time units, NRR will be removed from its private directory. T represents a point in time when con_K is published in the public directory;

[0081] The third stage involves reliability evaluation based on the denial detection results, including reliability assessment of entity A against entity B and delta reliability assessment, and reliability assessment of entity B against entity A and delta reliability assessment.

[0082] After the three phases of the new agreement are completed, entity A sends message M to entity B. Simultaneously, entities A and B collect three identical digital signatures as immutable evidence to prove the other party's message transmission behavior, namely: NRO = sS A (f NRO ,B,L,C),NRR=sS B (f NRR ,A,L,t B ,NRO) and con_K=sS δ (f CON ,A,B,L,T,t A ,t B ,K,NRO,NRR).

[0083] Furthermore, in the third stage, when conducting reliability assessment based on the repudiation detection results, the specific process for determining whether the sender has repudulated is as follows:

[0084] CESB A The actions claimed by sender A; digital signatures NRO, NRR, and con_K represent immutable evidence held by receiver B; L is the protocol round tag;

[0085] Step 1: Initialize the real behavior variable RESB A Both the detection result variable SR and SR are 0;

[0086] Step 2: Determine if con_K is equal to the signature sS of δ. δ (f CON ,A,B,L,T,t A ,t B (K, NRO, NRR), is the signature sS of B equal to NRR? B (f NRR ,A,L,t B (NRO), is NRO equal to the signature sS of A? A (f NRO Given the hash values ​​of M, B, L, and C, determine if L is equal to the hash value h(M). If all four conditions are met, then set RESB to 0.A Set to 1; if not all four conditions are met, then set RESB. A Set it to 0; if not all four conditions are met, then set RESB. A It is still set to 0 (meaning that A must not have sent message M).

[0087] Step 3: If RESB A =1, judge RESB A Is it equal to CESB? A If they are equal, SR is set to 1 (i.e., A honestly claims to have sent message M); otherwise, SR is set to 2 (i.e., A did not honestly claim to have sent message M).

[0088] If RESB A =0, judge RESB A Is it equal to CESB? A If they are equal, SR is set to 3 (i.e., A honestly claims that it did not send message M); otherwise, SR is set to 4 (i.e., A did not honestly claim that it did not send message M).

[0089] Step 4: Return the sender's denial detection result (SR).

[0090] Furthermore, in the third stage, when conducting reliability assessment based on the denial detection results, the specific process for determining whether the recipient has denied the claim is as follows:

[0091] CERB B For the claimed actions of receiver B; digital signatures NRO, NRR, and con_K represent immutable evidence held by sender A; L is the protocol theory label;

[0092] Step 1: Initialize the real behavior variable RERB B Both the detection result variable RR and the detection result variable are 0;

[0093] Step 2: Determine if con_K equals signature sS δ (f CON ,A,B,L,T,t A ,t B (K, NRO, NRR), is NRR equal to signature sS? B (f NRO ,A,L,t B (NRO), is NRO equal to signature sS? A (f NRO Let L be the hash value of RB, B, L, C, and C. If L is equal to the hash value h(M), then RERB is set to 1. B Set to 1 (meaning B has definitely received information M); if not all four conditions are met, then RERB is set to 1. B It is still set to 0 (meaning that B has definitely not received information M);

[0094] Step 3: If RERB B =1, judge RERB B Is it equal to CERB? B If they are equal, RR is set to 1 (meaning B honestly claims to have received information M); otherwise, RR is set to 2 (meaning B did not honestly claim to have received information M).

[0095] If RERB B =0, judge RERB B Is it equal to CERB? B If they are equal, then RR is set to 3 (i.e., B honestly claims that it did not receive message M); otherwise, RR is set to 4 (i.e., B did not honestly claim that it did not receive message M).

[0096] Step 4: Return the receiver's denial detection result (RR).

[0097] Beneficial effects: This invention addresses the problem that the current Internet lacks a centralized management authority to rely on, making it impossible to obtain all information about a particular entity. This leads to the inability of existing non-repudiation protocols built on the weakening of trusted third-party solutions to function when third-party security information is incomplete. Based on the idea of ​​trust management, which acknowledges the incompleteness of security information in the system and that the system's security decisions require additional information provided by other parties, this invention constructs a non-repudiation protocol construction method based on a three-stage trust management model and uses this method to build a new protocol.

[0098] The advantages of this invention are as follows:

[0099] (1) The non-repudiation protocol construction method based on the three-stage trust management model can be used to construct non-repudiation protocols in untrusted scenarios where there is no trusted third party and all information of a subject cannot be obtained. It can be used to adapt existing online or offline non-repudiation protocols that use trusted third parties and various schemes that weaken the use of trusted third parties to environmental conditions, so as to improve the practicality of these protocols in real network environments.

[0100] (2) The new non-repudiation protocol based on the three-stage trust management model constructed by the method proposed in this invention is close to the non-repudiation protocol based on a trusted third party in terms of the success rate of ensuring non-repudiation of information transmission.

[0101] (3) The closed-loop self-feedback operation of the three stages of trust management enables the non-repudiation protocol based on the three-stage trust management model constructed in this invention to provide the information transmission denial pre-suppression capability that the non-repudiation protocol based on a trusted third party cannot provide. Attached Figure Description

[0102] Figure 1This is the three-stage model of the present invention;

[0103] Figure 2 This is a schematic diagram of the nearest neighbor extraction process in the embodiment;

[0104] Figure 3 This is a schematic diagram of the 512-node topology in the embodiment;

[0105] Figure 4 This is a schematic diagram of the common nearest neighbor probability in the embodiment;

[0106] Figure 5 This is a schematic diagram of the probability distribution of some common neighbors in the example;

[0107] Figure 6 This is a schematic diagram illustrating the connectivity density of some nodes in the embodiment;

[0108] Figure 7 This is a schematic diagram of the confidence value at a certain time point in the embodiment.

[0109] Figure 8 This is a schematic diagram of the entity malicious behavior matrix injection model in the embodiment;

[0110] Figure 9 This is a schematic diagram illustrating the successful interaction of the new protocol constructed in a scenario without collusion or defamation, as shown in the example.

[0111] Figure 10 This is a schematic diagram illustrating the successful interaction of a new protocol constructed in a scenario with defamation but no collusion, as described in the example.

[0112] Figure 11 This is a schematic diagram illustrating the successful interaction of a new protocol constructed in a scenario of collusion without defamation, as shown in the example.

[0113] Figure 12 This is a schematic diagram illustrating the successful interaction of a new protocol constructed under a scenario involving collusion and defamation, as described in the embodiment.

[0114] Figure 13 This is a schematic diagram illustrating the message repudiation suppression capability (rejection recall, precision, and harmonic mean) in a scenario without collusion or defamation as described in the example.

[0115] Figure 14 This is a schematic diagram illustrating the message repudiation suppression capability (rejection recall, precision, and harmonic mean) in a scenario with defamation but no collusion, as illustrated in the example.

[0116] Figure 15 This is a schematic diagram illustrating the message repudiation suppression capability (rejection recall, precision, and harmonic mean) in a scenario with collusion but no defamation, as illustrated in the example.

[0117] Figure 16This is a schematic diagram illustrating the message repudiation suppression capability (recall rate, precision rate, and harmonic mean) in a scenario involving collusion and defamation, as shown in the example. Detailed Implementation

[0118] The technical solution of the present invention will be described in detail below, but the scope of protection of the present invention is not limited to the embodiments described.

[0119] like Figure 1 As shown, the present invention provides a method for constructing a non-repudiation protocol based on a three-phase trust management model, comprising the following steps:

[0120] Step 1: Employ a trust-based third-party selection mechanism for trust selection: When entity A needs to communicate with entity B, based on the trust levels of other entities, entities A and B select a nearest neighbor δ that is equally trusted by both parties in the protocol interaction to act as a trusted third party. The nearest neighbor entity δ has the highest trust level. This includes the following process:

[0121] Step 1.1: Extract connected nearest neighbors. From the common neighbors of interacting entities A and B, select qualified common neighbors based on the connectivity density (ECTD) assessment as a candidate set of peer nearest neighbors. The specific method is as follows:

[0122] First, based on the entity communication network ECN, find the neighbor node sets NSA and NSB of entities A and B respectively. If ECNS = NSA ∩ NSB, then... Then A and B have no common neighbors; otherwise, for each entity δ in NSA∩NSB, calculate the estimated and revaluation values ​​of the connectivity between entities A and B and δ, and take all entities whose revaluation values ​​are not less than the connectivity thresholds TECTD(A) and TECTD(B) as the connected nearest neighbor set ECNS of A and B.

[0123] Step 1.2: Selecting Peer Neighbors. Communication neighbors that are mutually trusted by both the sender and receiver are selected as peer neighbors. The specific method is as follows:

[0124] First, iterate through the connected nearest neighbor set ECNS of entities A and B. For any entity v∈ECNS, if TVmin≥|TV(A,v)-TV(B,v)|, it indicates that the system is not in a cold start state of trust evaluation. Then, select the peer nearest neighbor CNTN(A,B) of entities A and B based on the trust level. If for each entity v in ECNS(A,B), TV(A,v)=0 or TV(B,v)=0, it indicates that the system is in a cold start state of trust evaluation. Then, select the peer nearest neighbor CNTN(A,B) of entities A and B based on the connectivity density. TVmin is the trust difference threshold; TV(A,v) is the trust level of A to v, and TV(B,v) is the trust level of B to v.

[0125] Step 2: Employ a fair exchange-based evidence collection mechanism for trust-based interaction;

[0126] The selected peer neighbor δ is regarded as a trusted third party. Entity A and entity B use the existing fair and non-repudiable protocol based on trusted third parties to transmit information M from A to B. During the transmission of information M, the collection of interaction evidence is completed simultaneously. That is, A will hold non-repudiable evidence of B's ​​receipt of information M {B's digital signature NRR, δ's digital signature con_K}, and B will hold non-repudiable evidence of A's transmission of information M {A's digital signature NRO, δ's digital signature con_K}.

[0127] Step 3: Use a reliability feedback mechanism based on denial detection to conduct trust assessment;

[0128] After the information transfer between entity A and entity B is completed, based on the evidence held by A and B, a signature verification method is used to detect whether A and B have repudiated the transfer of M. The credibility of A, B and δ is evaluated and updated according to the repudiation detection results. Here, the repudiation detection includes sending repudiation detection and receiving repudiation detection.

[0129] Step 4: Based on the reliability assessment results of Step 3, repeat the third-party selection process of Step 1 to obtain new peer neighbors, and conduct a new round of trust interaction and trust assessment.

[0130] The detailed method for extracting nearest neighbors in step 1.1 of this embodiment is as follows:

[0131] Step 1: Set the initial values ​​of the neighbor set NSA of entity A, the neighbor set NSB of entity B, and the common neighbor set ECNS of A and B to empty;

[0132] Step 2: Target each node v in the ECN of the China Unicom network sequentially. x ∈V, if <A,v x If >∈E is an edge in ECN, then vx If added to A's neighbor set NSA, <B,v x If >∈E is an edge in ECN, then v x Add it to B's neighbor set NSB;

[0133] Step 3: If the intersection of neighbor set NSA and neighbor set NSB is empty, then set the entity connectivity nearest neighbor set ECNS of A and B to empty; otherwise, if the intersection of NSA and NSB is not empty, for each entity δ in the intersection of NSA and NSB, calculate the recalculated connectivity value ECTD(A,δ) of A and δ based on the obtained initial connectivity value Pre_ECTD(A,δ); at the same time, calculate the recalculated connectivity value ECTD(B,δ) of B and δ based on the obtained initial connectivity value Pre_ECTD(B,δ).

[0134] If both ECTD(A,δ) and ECTD(B,δ) are true, then δ is added to the entity connectivity nearest neighbor set ECNS of A and B.

[0135] Step 4: Return the connected nearest neighbor set ECNS(A,B) of entities A and B.

[0136] The communication tightness ECTD assessment in this embodiment includes, in sequence, the initial ECTD assessment and the ECTD reassessment.

[0137] The initial estimate of communication tightness (ECTD) is calculated as follows:

[0138] Step 1: Set the initial values ​​of the maximum number of interactions between any entities (maxMailNum) and the maximum number of shared neighbors (maxNghShrNm) to 0;

[0139] Step 2: For the entity communication network ECN = (V, E), calculate the number of interactions between any two nodes MailNum, and assign the maximum value among all interaction counts to maxMailNum; similarly, calculate the number of neighbor shares between any two nodes NghShrNm, and assign the maximum value among all neighbor shares to maxNghShrNm.

[0140] Step 3: Any two nodes v i ,v j The frequency of entity connections between ∈V is:

[0141] ECF(v i ,v j ) = MailNum(v i ,v j) / maxMailNum;

[0142] Two nodes v i ,v j The neighbor sharing rate between them is:

[0143] NSR(v i ,v j )=NghShrNm(v i ,v j ) / maxNghShrNm;

[0144] Step 4: Calculate v i and v j Estimated ECTD connectivity density:

[0145]

[0146] in ECF weighting factor;

[0147] Step 5: Return Pre_ECTD(v i ,v j );

[0148] The calculation method for the ECTD revalued communication tightness is as follows:

[0149]

[0150] Among them, 0<α,β,χ,δ<1; P=α*Pre_ECTD(v i ,v j ), D = χ * Pre_ECTD(v i ,v j c. T = δ * Pre_ECTD(v i ,v j )′|Pre_ECTD(v i ,v j The units are, in order, proportional, integral, differential, and time-dependent.

[0151] In the above formula, ECTD t-m (v i ,v j ), ECTD t-m+1 (v i ,v j ), ..., ECTD t-1 (v i ,v j ),ECTD t (v i ,v j ) represents vi and v j The re-estimated communication tightness obtained after the first m rounds [round tm...round t-1] and round t; component Pre_ECTD(v i ,v j ) for v i With v j The initial estimated network density obtained after the t-th round of communication; component For v i and v j The historical average reassessed connection density obtained before the t-th round of communication; component Pre_ECTD(v i ,v j )′ is v i and v j The volatility of the initial estimated connectivity tightness in round t; component Pre_ECTD(v i ,v j )′|Pre_ECTD(v i ,v j )″| for v i With v j The correction amount for the tightness of communication in round t; α, β, χ, δ are the weighting factors of the above four components in turn;

[0152] χ,δ follow the following two formulas:

[0153]

[0154]

[0155] The detailed method for peer-to-peer neighbor selection in step 1.2 of this embodiment is as follows:

[0156] Step 1: Let TV(A,v) and TV(B,v) be the reliability between v and A, and v and B, respectively;

[0157] For each entity v in ECNS(A,B), calculate the confidence distance |TV(A,v)-TV(B,v)|; select the v that minimizes the value of |TV(A,v)-TV(B,v)| to enter the peer nearest neighbor set CNTN(A,B) of A and B;

[0158] Step 2: If the system is in the cold start state of reliability assessment, that is, for each entity v in ECNS(A,B), TV(A,v)=0 or TV(B,v)=0, then it is necessary to select peer neighbors based on connectivity density: For each entity v in ECNS(A,B), calculate the connectivity density distance |ECTD(A,v)-ECTD(B,v)|; select the v that minimizes |ECTD(A,v)-ECTD(B,v)| to enter the peer nearest neighbor set CNTN(A,B) of A and B;

[0159] Step 3: Return the set of equal nearest neighbors of A and B, CNTN(A,B).

[0160] After step 3 of this embodiment completes the interactive communication, if entity A subsequently denies having sent information M to B for some private interest, B may submit NRO, NRR, and con_K to the arbitrator; the arbitrator will then trace entity A's previous actions using the following method:

[0161] The arbitrator verifies the correctness of the con_K signature to confirm whether δ has received the sub_K submitted by entity A; and verifies the NRR and NRO to determine whether B has indeed received the ciphertext C sent by A, and whether the ciphertext C is correct.

[0162] If all the above verifications are correct, it is determined that entity A did indeed send information M to B, thus confirming that entity A denied the information exchange with entity B.

[0163] The denial detection results for the message M sent by entity A to entity B include the following four categories: A honestly claims to have sent message M; A honestly claims not to have sent message M; A dishonestly claims to have sent message M; A dishonestly claims not to have sent message M; and then a reliability assessment is performed based on the denial detection results.

[0164] After step 3 of this embodiment completes the interactive communication, if B subsequently denies having received M sent by A for some private interest, A can submit NRO, NRR, and con_K to the arbitrator. The arbitrator will then trace B's previous actions. The tracing method is as follows:

[0165] The arbitrator verifies the correctness of the con_K signature to determine whether δ has received the NRR submitted by B, whether δ has generated evidence and published it in the public directory, and thus whether B can always obtain the key K between time point T and time point T+t; based on the verification of the NRR, the arbitrator determines whether B has received the ciphertext C sent by A and whether B is willing to obtain the key K from δ; based on the verification of the NRO, the arbitrator determines whether the ciphertext C received by B is correct.

[0166] If all the above verifications are correct, it is determined that entity B did receive information M between time point T and time point T+t, thus confirming that entity B committed fraud in the information transmission with entity A.

[0167] The reception denial detection results of information M received by entity B from entity A include four cases: B honestly claims to have received information M; B honestly claims not to have received information M; B does not honestly claim to have received information M; B does not honestly claim not to have received information M; and then a reliability assessment is performed based on the denial detection results.

[0168] In this embodiment, the reliability assessment assumes that A and B are protocol interaction entities, and δ is the selected peer nearest neighbor of A and B. The specific reliability assessment method is as follows:

[0169] Step 1) First, estimate the original reliability according to the following four scenarios;

[0170] Scenario 1: If entity A is arbitrated as honest, then entity A assigns a normal confidence level to δ and entity B.

[0171] If entity A is found guilty of denial in arbitration, and if entity A has indeed denied the allegations, entity A will retaliate against δ and entity B. Therefore, entity A will defame δ and entity B, assigning δ and entity B a confidence level lower than their actual performance. If entity A has not actually denied the allegations, then entity A will consider δ to have either neglected its duties, assigning δ a low confidence level; or δ to have conspired with entity B, naturally assigning δ and entity B a low confidence level.

[0172] Scenario 2: If entity B is arbitrated as honest, then entity B will assign a normal confidence level to δ and entity A.

[0173] If entity B is found guilty of denial in arbitration, and if entity B has indeed denied the allegations, entity B will retaliate against δ and entity A. Therefore, entity B will defame δ and entity A, assigning δ and entity A a confidence level lower than their actual performance. If entity B has not actually denied the allegations, then entity B will consider δ to have either neglected its duties and assigned δ a low confidence level, or δ to have conspired with entity A, naturally assigning δ and entity A a low confidence level.

[0174] Scenario 3: If entity B colludes with δ, then entity B will assign a high confidence level to δ in order to inflate the confidence level of δ.

[0175] Scenario 4: If entity A colludes with δ, then entity A will assign a high confidence level to δ in order to inflate the confidence level of δ.

[0176] Step 2) Calculate the overall reliability of δ based on the reliability evaluations of A and B for δ;

[0177] Let TV(A,δ) be the reliability rating of A for δ in step 1), TV(B,δ) be the reliability rating of B for δ, and RV(δ) be the historical comprehensive reliability of δ. Then the comprehensive reliability of δ is updated as follows:

[0178] First, update TV(A,δ) using TV(A,δ); then update TV(B,δ) using TV(B,δ); finally, update RV(δ) using RV(δ).

[0179] TV(A,δ)=κ*TV(A,δ)+λ*RV(δ)+μ*TV(A,δ)'+ξ*TV(A,δ)'*|TV(A,δ)″|;

[0180] TV(B,δ)=κ*TV(A,δ)+λ*RV(δ)+μ*TV(B,δ)'+ξ*TV(B,δ)'*|TV(B,δ)″|;

[0181]

[0182] The new protocol, based on the three-phase trust management model, comprises three phases:

[0183] In the first stage, entities A and B select a trustworthy entity δ based on their credibility.

[0184] In the second phase, entities A and B will select δ as a third party to interact with the protocol; the interaction consists of five processes:

[0185] A→B:f NRO ,B,L,C,NRO, that is, NRO=sS A (f NRO (,B,L,C);

[0186] A→δ:f SUB ,B,L,t A ,K,NRO,sub_K, that is, sub_K=sS A (f SUB ,B,L,t A ,K,NRO);

[0187] B→δ:f NRR ,A,L,t B ,NRO,NRR, that is, NRR=sS B (f NRR ,A,L,t B (,NRO);

[0188] B←δ:f CON ,A,B,L,T,t A ,t B ,K,NRR,con_K,that is,con_K=sS δ (f CON ,A,B,L,T,t A ,t B ,K,NRO,NRR);

[0189] A←δ:f CON ,A,B,L,T,t A ,t B ,K,NRR,con_K;

[0190] The message M to be sent is divided into two parts: the encryption key K and the ciphertext C;A It is a time period defined by A, indicating that δ has saved sub_K in its own private directory for t. A If the NRR submitted by B is not received after a certain number of time units, then sub_K will be deleted from its private directory. B It is a time period defined by B, indicating that δ has saved NRR in its own private directory for t. B If sub_K submitted by A is not received after a certain number of time units, NRR will be removed from its private directory. T represents a point in time when con_K is published in the public directory;

[0191] The third stage involves reliability evaluation based on the denial detection results, including reliability assessments of entity A against entity B and delta reliability assessments, and reliability assessments of entity B against entity A and delta reliability assessments.

[0192] After the three phases of the new agreement are completed, entity A sends message M to entity B. Simultaneously, entities A and B collect three identical digital signatures as immutable evidence to prove the other party's message transmission behavior, namely: NRO = sS A (f NRO ,B,L,C),NRR=sS B (f NRR ,A,L,t B ,NRO) and con_K=sS δ (f CON ,A,B,L,T,t A ,t B ,K,NRO,NRR).

[0193] In the third stage, when conducting reliability assessment based on the repudiation detection results, the specific process for determining whether the sender has repudulated is as follows:

[0194] CESB A The actions claimed by sender A; digital signatures NRO, NRR, and con_K represent immutable evidence held by receiver B; L is the protocol round tag;

[0195] Step 1: Initialize the real behavior variable RESB A Both the detection result variable SR and SR are 0;

[0196] Step 2: Determine if con_K is equal to the signature sS of δ. δ (f CON ,A,B,L,T,t A ,t B (K, NRO, NRR), is the signature sS of B equal to NRR? B (f NRR,A,L,t B (NRO), is NRO equal to the signature sS of A? A (f NRO Let L be the hash value of information M. If all four conditions are met, then set RESB to 0.5. A Set to 1; if not all four conditions are met, then set RESB. A Set to 0;

[0197] Step 3: If RESB A =1, judge RESB A Is it equal to CESB? A If they are equal, set SR to 1 (i.e., A honestly claimed to have sent message M); otherwise, set SR to 2 (i.e., A did not honestly claim to have sent message M); if RESB A =0, judge RESB A Is it equal to CESB? A If they are equal, set SR to 3 (i.e., A honestly claims not to have sent message M); otherwise, set SR to 4 (i.e., A did not honestly claim not to have sent message M).

[0198] Step 4: Return the sender's denial detection result (SR).

[0199] In the third stage, when conducting reliability assessment based on the denial detection results, the specific process for determining whether the recipient has denied the claim is as follows:

[0200] CERB B For the claimed actions of receiver B; digital signatures NRO, NRR, and con_K represent immutable evidence held by sender A; L is the protocol theory label;

[0201] Step 1: Initialize the real behavior variable RERB B Both the detection result variable RR and the detection result variable are 0;

[0202] Step 2: Determine if con_K is equal to the signature sS of δ. δ (f CON ,A,B,L,T,t A ,t B (K, NRO, NRR), is the signature sS of B equal to NRR? B (f NRO ,A,L,t B (NRO), is NRO equal to the signature sS of A? A (f NRO Let L be the hash value of information M. If all four conditions are met, then RERB is set to 0. B Set to 1; if not all four conditions are met, then RERB will be set to 1. BSet to 0;

[0203] Step 3: If RERB B =1, judge RERB B Is it equal to CERB? B If they are equal, then RR is set to 1 (i.e., B honestly claims to have received information M); otherwise, RR is set to 2 (i.e., B did not honestly claim to have received information M); if RERB B =0, judge RERB B Is it equal to CERB? B If they are equal, then RR is set to 3 (i.e., B honestly claims that it did not receive information M); otherwise, RR is set to 4 (i.e., B did not honestly claim that it did not receive information M).

[0204] Step 4: Return the receiver's denial detection result (RR).

[0205] Example

[0206] To verify the performance of this invention, simulation experimental data was first generated based on a complex network model. By calculating the existence probability of a peer trusted party, the feasibility of implementing third-party selection in this invention was illustrated. Based on a malicious behavior injection model, and considering four cross-situations of collusion / non-collusion and defamation between the third party and the protocol entity, the interaction success rate and pre-repudiation deterrence of the technical solution of this invention were compared with those of traditional non-repudiation protocols based on trusted third parties.

[0207] 1. Experimental Preparation

[0208] 1.1 Communication Topology

[0209] This embodiment uses the BA model to create a scale-free complex network communication topology model.

[0210] The network initially has m0 nodes with arbitrary connections, each node having at least one edge. Each time a new node is added to the network, m edges are created connecting it. The new node chooses which node to connect to with a certain probability, and the degree of choice is k. i The probability of connecting to a node is After t time steps, the number of network nodes in the model becomes N = t + m0, and the number of connections becomes m0 + m*t.

[0211] Given that the communication topology constructed by the BA model closely resembles the characteristics of real networks, the parameters m0 = 10, m = 3, and N = 512 were set in the experiment. The communication topology generated using the BA model is as follows: Figure 3 .

[0212] 1.2 Analysis of the rationality of the common nearest neighbor hypothesis

[0213] Two entity nodes are randomly selected from 512 entity nodes. As the total number of network nodes increases, the probability that any two nodes have exactly one common neighbor decreases. When the total number of network nodes reaches 100, the probability of having exactly one common neighbor approaches 0. Figure 4 As shown, the probability that any two nodes have exactly two common neighbors reaches its maximum value of 0.4 when the total number of nodes is around 20, and then decreases continuously. Furthermore, the probability that any two nodes have common neighbors reaches its maximum value of 1 when the total number of nodes is around 20, meaning that almost every pair of nodes has common neighbors.

[0214] The probabilities of any two randomly selected nodes having only one, two, or even more common neighbor nodes are shown in Table 1.

[0215] Table 1 shows the distribution of the number of common neighbors between any two nodes.

[0216] Number of neighbors 1 2 3 4 5 6 … 510 511 512 11 nodes 0.127 0.381 0.273 0.164 0.018 0.000 … 0.000 0.000 0.000 12 nodes 0.106 0.409 0.227 0.167 0.076 0.000 … 0.000 0.000 0.000 13 nodes 0.103 0.282 0.205 0.244 0.128 0.026 … 0.000 0.000 0.000 14 nodes 0.077 0.209 0.220 0.198 0.198 0.077 … 0.000 0.000 0.000 15 nodes 0.057 0.114 0.219 0.190 0.248 0.086 … 0.000 0.000 0.000 16 nodes 0.042 0.100 0.092 0.150 0.200 0.167 … 0.000 0.000 0.000 17 nodes 0.059 0.081 0.074 0.118 0.206 0.170 … 0.000 0.000 0.000 18 nodes 0.059 0.052 0.092 0.078 0.098 0.131 … 0.000 0.000 0.000 19 nodes 0.018 0.082 0.088 0.047 0.047 0.117 … 0.000 0.000 0.000 20 nodes 0.042 0.037 0.063 0.079 0.047 0.089 … 0.000 0.000 0.000 21 nodes 0.029 0.047 0.076 0.052 0.038 0.105 … 0.000 0.000 0.000 22 nodes 0.026 0.052 0.035 0.039 0.056 0.074 … 0.000 0.000 0.000 23 nodes 0.028 0.059 0.051 0.032 0.043 0.032 … 0.000 0.000 0.000 24 nodes 0.022 0.054 0.033 0.036 0.025 0.051 … 0.000 0.000 0.000 25 nodes 0.017 0.037 0.053 0.027 0.070 0.037 … 0.000 0.000 0.000 26 nodes 0.012 0.031 0.046 0.037 0.034 0.034 … 0.000 0.000 0.000 27 nodes 0.020 0.028 0.034 0.043 0.051 0.026 … 0.000 0.000 0.000 … … … … … … … … … … … 510 nodes 0.001 0.001 0.002 0.003 0.002 0.002 … 0.000 0.000 0.000 511 nodes 0.002 0.002 0.001 0.001 0.001 0.001 … 0.000 0.000 0.000 512 nodes 0.001 0.004 0.002 0.002 0.001 0.001 … 0.000 0.000 0.000

[0217] Figure 5 This example illustrates the probability trends of having exactly one, exactly two, exactly three, exactly four, and exactly five common neighbor nodes. In other words, when the number of common neighbor nodes remains constant, the probability of having only the specified number of common neighbor nodes decreases as the total number of nodes increases; conversely, when the total number of nodes remains constant, the probability increases as the number of common neighbor nodes increases.

[0218] 1.3 Third-party screening

[0219] (1) Calculation of communication tightness

[0220] against Figure 3 The communication topology shown here assigns the number of communications between any two nodes to a random integer within the range [0, 10]. Table 2 shows the communication density between any two nodes.

[0221] Table 2. Connectivity density between any two nodes

[0222]

[0223]

[0224] The connectivity trends of nodes 100, 200, 300, 400, and 500 with all other nodes are as follows: Figure 6 As shown.

[0225] (2) Reliability calculation and third-party screening

[0226] The confidence values ​​of all nodes at a certain time are shown in Table 3 and Figure 7 As shown.

[0227] Table 3 Reliability values ​​at a certain moment

[0228]

[0229] The data above shows that, for any two entities, it is always possible to select an entity with a certain degree of confidence difference as a third party.

[0230] 2. Performance Analysis

[0231] 2.1 Behavior Injection Model

[0232] The malicious behavior injection model of the entity selected in this embodiment is as follows: Figure 8 The diagram shows a matrix injection model; the honest and trustworthy behavior of an entity is set to a value of 1, while the untrustworthy behavior is set to a value of 0.1, and the behavior strategy time slot is 10 time slices. This behavior injection model is used to represent the behavior of malicious entity nodes.

[0233] 2.2 Interaction Success Rate

[0234] use Figure 8 The malicious behavior injection model shown is as follows: Figures 9 to 12 The paper presents a comparison of the success rate of information exchange between the new protocol based on the three-stage trust management model constructed in this invention and the traditional non-repudiation protocol based on a trusted third party (TTP) in four cross-scenarios: no collusion and no defamation, no collusion but defamation, collusion but no defamation, and collusion but defamation.

[0235] As can be seen from the above embodiments, the new non-repudiation protocol based on the three-stage trust management model constructed in this invention is close to the non-repudiation protocol based on a trusted third party in terms of the success rate of ensuring non-repudiation of information transmission.

[0236] 2.3. Excuses and the ability to suppress denial

[0237] To assess the pre-repudiation deterrence capability of non-repudiation protocols based on a three-stage trust management model, three common metrics will be used: recall, precision, and harmonic mean. Among these, recall... Precision Harmonic Mean Where TP represents the number of samples correctly identified as message repudiation beforehand, FN represents the number of samples misidentified as message non-repudiation beforehand, and FP represents the number of samples misidentified as message repudiation beforehand.

[0238] Still using here Figure 8 Malicious behavior injection model, Figures 13 to 16The results demonstrate the recall, precision, and harmonic mean of message repudiation under four cross-scenarios: no collusion and no defamation, no collusion but defamation, collusion but no defamation, and collusion but defamation, as the number of messages gradually increases to 1000. The experimental results show that the non-repudiation protocol construction method based on the three-stage trust management model proposed in this invention has a certain ability to suppress repudiation beforehand.

Claims

1. A method for constructing a non-repudiation protocol based on a three-phase trust management model, characterized in that, Includes the following steps: Step 1: Use a trust-based third-party selection mechanism to select a trusted third party: When entity A needs to communicate with entity B, based on the trust of other entities, entity A and entity B select a neighbor δ that is equally trusted by both parties in the protocol interaction to act as a trusted third party. The neighbor entity δ has the highest trust. The process includes the following: Step 1.1: Extract connected nearest neighbors. From the common neighbors of interacting entities A and B, select qualified common neighbors based on the connectivity density (ECTD) assessment as a candidate set of peer nearest neighbors. The specific method is as follows: First, based on the entity communication network (ECN), find the neighbor node sets NSA and NSB of entities A and B respectively. = for If entity A and entity B have no common neighbors, then entity A and entity B have no common neighbors; otherwise, for Each entity in Find the relationships between entity A and entity B. Initial estimate of connectivity tightness and revaluation And all entities whose revalued values ​​are not less than the compactness thresholds TECTD(A) and TECTD(B) are designated as the connected nearest neighbor set ECNS of A and B; Step 1.2: Selecting Peer Neighbors. Communication neighbors that are mutually trusted by both the sender and receiver are selected as peer neighbors. The specific method is as follows: First, iterate through the connected nearest neighbor set ECNS of entity A and entity B. For any entity... ,like This indicates that the system is not in a cold start state for trust assessment, and that peer neighbors of entity A and entity B are selected based on trust levels. If for each entity in ECNS(A, B) or This indicates that the system is in a cold start state of trust assessment. Therefore, peer neighbors of entity A and entity B are selected based on the closeness of entity connectivity. ; This is the threshold for reliability differences; For pair A Trust level, right Trust level; Step 2: Employ a fair exchange-based evidence collection mechanism for trust-based interaction; The selected peer neighbor δ is considered a trusted third party. Entities A and B use an existing fair and non-repudiable protocol based on a trusted third party to transmit information M from A to B. During the transmission of information M, the collection of interaction evidence is completed simultaneously. That is, A will hold non-repudiable evidence of B's ​​receipt of information M {B's digital signature NRR, δ's digital signature}. Meanwhile, B will possess irrefutable evidence that A has sent the message M {A's digital signature NRO, δ's digital signature}. }; Step 3: Use a reliability feedback mechanism based on denial detection to conduct trust assessment; After the information transfer between entity A and entity B is completed, based on the evidence held by A and B, a signature verification method is used to detect whether A and B have repudiated the transfer of M. The credibility of A, B and δ is evaluated and updated according to the repudiation detection results. Here, the repudiation detection includes sending repudiation detection and receiving repudiation detection. Step 4: Based on the reliability assessment results of Step 3, repeat the third-party selection process of Step 1 to obtain new peer neighbors, and conduct a new round of trust interaction and trust assessment.

2. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1, characterized in that, The detailed method for extracting nearest neighbors in step 1.1 is as follows: Step 1: Set up the neighbor set for entity A The neighbor set of entity B , and the common neighborhood set of A and B The initial value is empty; Step 2: Target each node in the physical communication network sequentially. ,like If it is an edge in an ECN, then... If added to A's neighbor set NSA, If it is an edge in an ECN, then... Add it to B's neighbor set NSB; Step 3: If the neighbor set and Neighbors If the intersection of A and B is empty, then set the nearest neighbor set ECNS of entities connected to A and B to empty; otherwise, if the intersection of NSA and NSB is not empty, then for each entity in the intersection of NSA and NSB... Based on the obtained A and Initial estimate of connectivity tightness Calculate the revalued connectivity tightness of A and δ. Meanwhile, based on the obtained entity B and Initial estimate of connectivity tightness Calculate B and Revaluation of connectivity ; like The communication tightness threshold TECTD(A) is not less than A and If the connectivity tightness threshold TECTD(B) of not less than B is simultaneously true, then δ is added to the entity connectivity nearest neighbor set ECNS of A and B. Step 4: Return the connected nearest neighbor set ECNS(A,B) of entities A and B.

3. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1 or 2, characterized in that, The communication tightness ECTD assessment includes, in sequence, an initial ECTD assessment and a reassessment of the communication tightness ECTD. The initial estimate of communication tightness (ECTD) is calculated as follows: Step 1: Set the initial values ​​of the maximum number of interactions between any entities (maxMailNum) and the maximum number of shared neighbors (maxNghShrNm) to 0; Step 2: For the entity communication network ECN=(V, E), calculate the number of interactions between any two nodes MailNum, and assign the maximum value among all interaction counts to maxMailNum; Similarly, calculate the neighbor sharing number NghShrNm between any two nodes, and assign the maximum value among all neighbor sharing numbers to maxNghShrNm; Step 3: Any two nodes The frequency of communication between entities is: Two nodes The neighbor sharing rate between them is: ; Step 4: Calculation Initial estimate of ECTD connectivity density: ; in ECF weighting factor; ; Step 5: Return ; The calculation method for the ECTD revalued communication tightness is as follows: ; ; in ; I=β , , The units are, in order, proportional units, integral units, differential units, and time-dependent units; In the above formula, , , They represent and The re-estimated communication tightness obtained after the first m rounds [rounds tm to t-1] and round t; components for and The initial estimated network density obtained after the t-th round of communication; component for and Historical average reassessed network tightness obtained before round t of communication; component for and Volatility of the initial estimated connectivity tightness in round t; component for and The correction amount for the tightness of communication in round t; The weighting factors for the four components mentioned above are, in order. Follow these two formulas: ; 。 4. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1, characterized in that, The detailed method for peer-to-peer neighbor selection in step 1.2 is as follows: Step 1: Let TV(A,v) and TV(B,v) be the reliability between v and A, and v and B, respectively; For each entity v in ECNS(A, B), calculate the confidence distance |TV(A,v)-TV(B,v)|; select the v that minimizes the value of |TV(A,v)-TV(B,v)| to enter the peer nearest neighbor set CNTN(A, B) of A and B; Step 2: If the system is in a cold start state for reliability assessment, that is, for each entity in ECNS(A, B) Then, it is necessary to select peer neighbors based on the degree of connectivity: for For each entity v in the dataset, calculate the connectivity density distance. ; Selecting makes The smallest v is added to the equivalent nearest neighbor set CNTN(A, B) of A and B; Step 3: Return the set of equal nearest neighbors of A and B, CNTN(A, B).

5. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1, characterized in that, After the interactive communication is completed in step 3, if entity A subsequently denies sending information M to B for some private interest, B may submit NRO, NRR, and con_K to the arbitrator; the arbitrator will then trace entity A's previous actions using the following method: The arbitrator verifies the correctness of the con_K signature to confirm whether δ has received the sub_K submitted by entity A; and verifies the NRR and NRO to determine whether B has indeed received the ciphertext C sent by A, and whether the ciphertext C is correct. If all the above verifications are correct, it is determined that entity A did indeed send information M to B, thus confirming that entity A denied the information exchange with entity B. The denial detection results for the message M sent by entity A to entity B include the following four categories: A honestly claims to have sent message M; A honestly claims not to have sent message M; A dishonestly claims to have sent message M; A dishonestly claims not to have sent message M; and then a reliability assessment is performed based on the denial detection results.

6. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1, characterized in that, After the interactive communication is completed in step 3, if B subsequently denies having received M from A for some private interest, A may submit NRO, NRR, and con_K to the arbitrator. The arbitrator will then trace B's previous actions. The tracing method is as follows: The arbitrator verifies the correctness of the con_K signature to determine whether δ has received the NRR submitted by B, whether δ has generated evidence and published it in the public directory, and thus whether B can always obtain the key K between time point T and time point T+t; based on the verification of the NRR, the arbitrator determines whether B has received the ciphertext C sent by A and whether B is willing to obtain the key K from δ; based on the verification of the NRO, the arbitrator determines whether the ciphertext C received by B is correct. If all the above verifications are correct, it is determined that entity B did indeed receive information M between time point T and T+t, thus confirming that entity B committed fraud in the information transmission with entity A. The reception denial detection result of information M received by entity B from entity A includes four cases: B honestly claims to have received information M; B honestly claims that he did not receive message M; B did not honestly claim to have received message M; B did not honestly claim that he did not receive message M; Then, reliability is assessed based on the denial detection results.

7. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1, 5, or 6, characterized in that, In the reliability assessment, A and B are protocol interaction entities, and δ is the selected peer nearest neighbor of A and B. The specific reliability assessment method is as follows: Step 1) First, estimate the original reliability according to the following four scenarios; situation If entity A is arbitrated as honest, then entity A assigns a normal confidence level to both δ and entity B. If entity A is found guilty of denial in arbitration, and if entity A has indeed denied the allegations, entity A will retaliate against δ and entity B. Therefore, entity A will defame δ and entity B, assigning δ and entity B a confidence level lower than their actual performance. If entity A has not actually denied the allegations, then entity A will consider δ to have either neglected its duties, assigning δ a low confidence level; or δ to have conspired with entity B, naturally assigning δ and entity B a low confidence level. situation If entity B is arbitrated as honest, then entity B will assign a normal confidence level to δ and entity A. If entity B is found guilty of denial in arbitration, and if entity B has indeed denied the allegations, entity B will retaliate against δ and entity A. Therefore, entity B will defame δ and entity A, assigning δ and entity A a confidence level lower than their actual performance. If entity B has not actually denied the allegations, then entity B will consider δ to have either neglected its duties, assigning δ a low confidence level; or δ to have conspired with entity A, naturally assigning δ and entity A a low confidence level. situation If entity B colludes with δ, then entity B will assign a high confidence level to δ in order to inflate the confidence level of δ. situation If entity A colludes with δ, then entity A will assign a high confidence level to δ in order to inflate the confidence level of δ. Step 2) Calculate the overall reliability of δ based on the reliability evaluations of A and B for δ; Let the reliability rating of A for δ in step 1) be: B's reliability rating for δ is , Let δ be the historical comprehensive reliability. Then, the updated comprehensive reliability of δ is as follows: First use right Update; and utilize right Update and finally utilize right Update; ; ; 。 8. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 1, characterized in that, The operation of the non-repudiation protocol consists of three phases: In the first stage, entities A and B select a trustworthy entity δ based on their credibility. In the second phase, entities A and B will select δ as a third party to interact with the protocol; the interaction consists of five processes: ,Right now ; ,Right now ; ,Right now ; ; ; The message M to be sent is divided into two parts: the encryption key K and the ciphertext C; It is a time period defined by A, indicating that δ has saved sub_K in its own private directory for a period of time. If the NRR submitted by B is not received after a certain number of time units, sub_K will be deleted from its own private directory. It is a time period defined by B, indicating that δ has saved NRR in its own private directory. If sub_K submitted by A is not received after a certain number of time units, NRR will be deleted from its private directory; T represents a point in time when con_K is published in the public directory; The third stage involves reliability evaluation based on the denial detection results, including reliability assessments of entity A against entity B and delta reliability assessments, and reliability assessments of entity B against entity A and delta reliability assessments. , ; After the three phases of the new agreement are completed, entity A sends message M to entity B. Simultaneously, entities A and B collect three identical digital signatures as immutable evidence to prove the other party's message transmission behavior. These signatures are: , and .

9. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 8, characterized in that... , characterized in that, In the third stage, when conducting reliability assessment based on the repudiation detection results, the specific process for determining whether the sender has repudulated is as follows: make This is a claim made by sender A; Digital Signature con_K represents the immutable evidence held by receiver B; L represents the protocol round tag. Step 1: Initialize the actual behavior variables and detection result variables All are 0; Step 2: Determine if con_K is equal to the signature of δ. , Is it equal to B's signature? , Is it equal to A's signature? Is L equal to the hash value of M? If all four conditions are met, then... Set to 1; if not all four conditions are met, then... Set to 0; Step 3: If ,judge If they are equal, set SR to 1, meaning A honestly claimed to have sent message M; otherwise, set SR to 2, meaning A did not honestly claim to have sent message M. ,judge ; If they are equal, set SR to 3, meaning A honestly claims that it did not send message M; otherwise, set SR to 4, meaning A did not honestly claim that it did not send message M. Step 4: Return the sender's denial detection result (SR).

10. The method for constructing a non-repudiation protocol based on a three-phase trust management model according to claim 8, characterized in that, In the third stage, when conducting reliability assessment based on the denial detection results, the specific process for determining whether the recipient has denied the claim is as follows: make The claim made by recipient B; Digital signatures NRO, NRR, and con_K represent immutable evidence held by sender A; L is the protocol round tag. Step 1: Initialize the actual behavior variables and detection result variables All are 0; Step 2: Determine if con_K is equal to the signature of δ. , Is it equal to B's signature? , Is it equal to A's signature? Is L equal to the hash value of M? If all four conditions are met, then... Set to 1; if not all four conditions are met, then... Set to 0; Step 3: If ,judge Is it equal to ; If they are equal, then RR is set to 1, meaning B honestly claims to have received information M; otherwise, RR is set to 2, meaning B did not honestly claim to have received information M. ,judge Is it equal to ; If they are equal, RR is set to 3, meaning B honestly claims that it did not receive message M; otherwise, RR is set to 4, meaning B did not honestly claim that it did not receive message M. Step 4: Return the receiver's denial detection result (RR).

Citation Information

Patent Citations

  • Trust management system based trusted reconstructing method of IP routing protocol

    CN101888328A

  • GPSR secure routing protocol implementation method based on block chain trust model

    CN112911549A