Distributed identity trust evaluation system based on trusted network

Abstract

The invention relates to the technical field of distributed digital identity authentication, in particular to a distributed identity trust evaluation system based on a trusted network, which comprises three roles of a certificate holder, an issuer and a verifier, and a block chain storage module used for storing interaction behavior records among nodes in a non-tampering manner; the trust calculation module is used for periodically calculating trust values of the nodes through the smart contract; the real-time detection module is used for monitoring whether the nodes have malicious behaviors or not; the dynamic punishment module is used for punishing the trust value of the node generating the malicious behavior when the malicious behavior is detected; the access control module is used for monitoring whether the trust value of the node is lower than a preset security value or not, and if yes, implementing a network access isolation measure on the node to limit the node to continuously participate in network interaction; according to the invention, the problem of decentralizing malicious nodes in the trusted network is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of distributed digital identity authentication technology, and more specifically to a distributed identity trust evaluation system based on a trust network. Background Technology

[0002] With the rapid development of the Internet of Things (IoT) and Web3 ecosystems, distributed systems are finding increasingly widespread applications across various fields, from smart healthcare and supply chain management to education and financial services. Nodes in these systems are often deployed in a distributed manner and work collaboratively, enabling users to interact freely across multiple platforms. However, the openness and complexity of distributed systems also bring significant challenges in data security, privacy protection, and trust management. Particularly in decentralized identity (DID) networks, while user data autonomy and transparency are further enhanced, efficient assessment of identity trust becomes a critical issue affecting system reliability and security.

[0003] Most existing blockchain-based trust networks employ a single global reputation score or static evaluation model, failing to design differentiated trust evaluation dimensions for different roles in a distributed identity system (such as credential holders, issuers, and verifiers). This results in evaluation results that are difficult to comprehensively reflect the behavioral characteristics and risk sources of each role. Furthermore, these solutions typically lack real-time detection and dynamic punishment mechanisms for multi-category, hierarchical malicious behavior by nodes, and are insufficient in identifying and suppressing complex attack patterns (such as collusion and selective denial). In addition, most trust models do not adequately consider time decay and the cumulative effect of behavior, causing trust value updates to lag and failing to dynamically adapt to changes in network behavior, thus affecting the system's real-time performance and fairness.

[0004] Therefore, this invention aims to solve the following specific technical problems existing in the current blockchain-based distributed trust network: First, the trust assessment dimension is too singular and cannot reflect the differences in roles; second, there is a lack of fine-grained, classifiable and graded response mechanisms for identifying and punishing malicious behavior; and third, the trust model lacks dynamic adaptability and is difficult to accurately and timely reflect the trust status of nodes. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a distributed identity trust evaluation system based on a trust network, comprising multiple nodes, each node having the identity of a credential issuer, credential holder, or credential verifier; the credential holder is the party submitting the credential; the credential issuer issues a verifiable credential to the credential holder; the credential verifier verifies the verifiable credential of the credential holder; and further comprising:

[0006] A blockchain storage module is used to immutably store records of interactions between nodes.

[0007] The trust calculation module is used to periodically calculate the trust value of nodes through smart contracts, specifically including:

[0008] The holder trust calculation module is used to calculate the trust value of the certificate holder. The trust value of the certificate holder is obtained by weighted summation of the certificate holder's historical behavior trust score and transaction feedback trust score.

[0009] The issuer trust calculation module is used to calculate the trust value of the credential issuer. The trust value of the credential issuer is obtained by weighted summation of the issuer's historical issuance trust score and verification feedback trust score.

[0010] The verifier trust calculation module is used to calculate the trust value of the credential verifier. The trust value of the credential verifier is calculated based on the verification accuracy rate.

[0011] The real-time detection module is used to monitor whether nodes are engaging in malicious behavior;

[0012] The dynamic penalty module is used to penalize the trust value of the node that generates malicious behavior when malicious behavior is detected.

[0013] The access control module monitors whether a node's trust value is lower than a preset security value. If so, it implements network access isolation measures for the node, restricting it from continuing to participate in network interactions.

[0014] The beneficial effects of this invention are:

[0015] Decentralization and Security: By fully leveraging the transparency, immutability, and automatic execution of smart contracts in blockchain, the calculation, updating, and penalty mechanisms of trust values ​​are completely decentralized, eliminating the risk of single points of failure and significantly improving the overall security of the system.

[0016] Dynamic quantitative assessment: By quantitatively analyzing the trust value of three core roles through multiple dimensions (behavioral history, interactive feedback), and introducing time decay factor and behavior frequency influence factor, dynamic and accurate assessment of trust value is achieved, which can reflect the reliability changes of node behavior in real time.

[0017] Resisting Malicious Behavior: The designed dynamic penalty mechanism can effectively identify and severely punish various malicious behaviors (fake issuance, forged submission, improper rejection, etc.), and the penalty intensity dynamically increases with the severity and frequency of the behavior. Finally, the isolation mechanism excludes seriously malicious nodes from the network, greatly enhancing the network's resistance to attacks and robustness.

[0018] Adaptability and scalability: The model takes into account the time factor in the processing of behavioral history, feedback and penalties, which enables trust assessment to adapt to the dynamic changes in network environment and node behavior; the blockchain-based architecture has good scalability and is suitable for large-scale distributed identity management scenarios.

[0019] Efficient Trust Management: Automated smart contracts significantly improve the efficiency of trust assessment and management, solving the problems of inefficiency and lack of transparency in trust management in traditional centralized or simple distributed models, and providing core support for building a reliable and secure decentralized identity ecosystem. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of a distributed identity trust evaluation system based on a trust network, as shown in some embodiments of the present invention.

[0021] Figure 2 This is a framework diagram of a distributed identity trust evaluation system based on a trust network, as shown in some embodiments of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Figure 1 This is a schematic diagram of a distributed identity trust evaluation system based on a trust network, as shown in some embodiments of the present invention. Figure 2 This is a framework diagram of a distributed identity trust evaluation system based on a trust network, as shown in some embodiments of the present invention.

[0024] This invention specification provides a distributed identity trust evaluation system based on a trust network, such as... Figure 1-2 As shown, a distributed trust network is established based on blockchain technology, defining three core roles: credential issuer, credential holder, and credential verifier.

[0025] Credential holders are responsible for submitting and managing digital credentials.

[0026] The credential issuer is responsible for generating and issuing verifiable credentials.

[0027] A voucher verifier is used to verify the authenticity and validity of vouchers.

[0028] All role interactions (such as credential issuance, verification, and revocation) are recorded in the blockchain storage module via smart contracts, ensuring that the data is immutable and traceable.

[0029] Specifically, the system also includes:

[0030] The trust calculation module is used to periodically calculate the trust value of nodes through smart contracts, specifically including:

[0031] The holder trust calculation module is used to calculate the trust value of the certificate holder. The trust value of the certificate holder is obtained by weighted summation of the certificate holder's historical behavior trust score and transaction feedback trust score.

[0032] Furthermore, the trust value T of the credential holder H Represented as:

[0033]

[0034] In the formula, α H β H The trust score T represents the historical behavior. Hh And transaction feedback trust score T Hf The weights satisfy α H +β H =1.

[0035] The issuer trust calculation module is used to calculate the trust value of the credential issuer. The trust value of the credential issuer is obtained by weighted summation of the issuer's historical issuance trust score and verification feedback trust score.

[0036] Furthermore, the trust value T of the credential issuer. I Represented as:

[0037]

[0038] In the formula, α I β I The historical issuance trust score T is respectively Ih And verification feedback trust score T If The weights satisfy α I +β I =1.

[0039] The verifier trust calculation module is used to calculate the trust value of the credential verifier. The trust value of the credential verifier is calculated based on the verification accuracy rate.

[0040] The real-time detection module is used to monitor whether nodes are engaging in malicious behavior;

[0041] The dynamic penalty module is used to penalize the trust value of the node that generates malicious behavior when malicious behavior is detected.

[0042] The access control module monitors whether a node's trust value is lower than a preset security value. If so, it implements network access isolation measures for the node, restricting it from continuing to participate in network interactions.

[0043] In some embodiments, the specific processing steps of the holder trust calculation module include:

[0044] The historical behavior data of credential holders can be obtained from the blockchain distributed ledger through smart contracts. The historical behavior data includes the number of times the credential holder has successfully submitted and been verified, the number of times false and erroneous credentials have been identified, and the number of times invalid credentials have been actively revoked.

[0045] Based on the historical behavior data of credential holders, a Bayesian probability model is used to quantify the overall reliability of historical behavior, resulting in a historical behavior trust score. The quantification logic of the Bayesian probability model is as follows: the number of successfully submitted and verified credentials is used as a positive factor, the number of times false and erroneous credentials are identified is used as a negative factor, and the number of times invalid credentials are actively withdrawn partially offsets the failure impact by a preset coefficient; a decay coefficient is introduced to adjust the influence of the total number of historical behaviors on the current trust value.

[0046] Specifically, the historical behavior trust score, obtained by quantifying using a Bayesian probability model, can be expressed as:

[0047]

[0048] In the formula, T Hh S represents the trust score of the credential holder's historical behavior. H F represents the number of times a credential holder has successfully submitted and had it verified. H C represents the number of times a credential holder's submission of false or erroneous credentials has been identified. H λ represents the number of times the certificate holder actively revokes an invalid certificate. H The contribution factor controlling the active revocation behavior is represented by N, which represents the number of counterparties of the certificate holder; μ represents the influence coefficient of the number of behaviors.

[0049] μ determines the degree to which the total number of actions affects the trust value. A larger μ value is suitable for scenarios that require a rapid response to the impact of the number of actions.

[0050] In some embodiments, the contribution factor λ that controls the active revocation behavior H =0.5.

[0051] The system calculates a transaction feedback trust score by tracking the ratings given to the credential holder by counterparties who interact with the holder. These counterparties include credential validators and other dependent parties.

[0052] Specifically, a time decay mechanism is introduced into the calculation of the feedback trust score, so that the older the rating is, the lower its weight in influencing the current trust value, as expressed as:

[0053]

[0054]

[0055] In the formula, T Hf R represents the transaction feedback trust score of the certificate holder. H,i T represents the rating provided by the i-th counterparty in the interaction with the credential holder. H,i D represents the trust value of the i-th counterparty interacting with the credential holder. H,i Let λ represent the time decay factor, and Δt represent the decay factor. H,i This indicates the time elapsed since the occurrence of the i-th transaction.

[0056] In some embodiments, the specific processing steps of the issuer trust calculation module include:

[0057] The historical signing behavior data of the voucher issuer includes the number of times the voucher issuer successfully issued and verified as valid vouchers, the number of times false and erroneous vouchers were identified, and the number of times the issuer voluntarily revoked invalid vouchers that had already been issued.

[0058] Based on the historical issuance behavior data of the credential issuer, a Bayesian probability model is used to quantify the historical issuance trust score, which is expressed as:

[0059]

[0060] In the formula, T Ih S represents the historical issuance trust score of the credential issuer. I F represents the number of times a credential issuer has successfully issued and verified a credential as valid. I C represents the number of times false and erroneous credentials have been identified. I λ indicates the number of times the issuer of a certificate has voluntarily revoked an invalid certificate that has already been issued. I The contribution factor for controlling the active revocation behavior is represented by μ', the influence coefficient of the number of behaviors is represented by N', and the number of credential validators that interact with the credential validator is represented by N.

[0061] The credential verifiers who interact with the credential issuer rate the credential issuer, and calculate the verification feedback trust score, expressed as follows:

[0062]

[0063] In the formula, T IfR represents the transaction feedback trust score of the credential issuer. I,j T represents the score given by the i-th credential validator who interacts with the credential issuer. I,j D represents the trust value of the i-th credential verifier that interacts with the credential issuer. I,j This represents the time decay factor.

[0064] In some embodiments, the specific processing steps of the verifier trust calculation module include:

[0065] The number of times a voucher verifier correctly completes the verification operation, the number of times erroneous verification occurs, and the number of times biased behavior is exhibited are quantified using an accuracy model to obtain a trust value, which is expressed as:

[0066]

[0067] Among them, T V S represents the trust value of the credential verifier. V F represents the number of times the credential verifier correctly completed the verification operation. V μ'' represents the number of times the credential validator makes incorrect validations and exhibits biased behavior, μ'' represents the influence coefficient of the number of behaviors, and N'' represents the number of counterparties of the credential validator.

[0068] In some embodiments, the trust value of each node is constantly updated. The calculated trust value does not exist in isolation, but is immediately put into the actual operation of the system as the core basis for subsequent access control module interaction decisions, forming a closed-loop control of "behavior-trust-permission".

[0069] The specific process of the access control module includes:

[0070] A smart contract is automatically triggered when a credential holder attempts to submit credentials to a service (its validator) to gain access.

[0071] The smart contract first queries the current trust value of the credential holder and compares it with the preset security value of the requested credential verifier. Different credential verifiers can set different preset security values ​​according to the security level requirements of their services.

[0072] If the holder's trust value is greater than or equal to a preset security value, the smart contract allows the verification process to continue. The smart contract then proceeds to verify the authenticity of the digital signature of the credential and other subsequent processes.

[0073] If the trust value of the credential holder is lower than the preset security value, the smart contract will automatically reject the access request and record this "rejection due to insufficient trust value" event on the blockchain. This effectively prevents nodes with low trust values ​​from abusing service resources.

[0074] Similarly, when assessing whether to trust a particular credential issuer, the credential verifier can also set a trust value requirement for the issuer. The smart contract will automatically check and mark credentials issued by issuers with too low a trust value as high-risk or reject them outright, thereby controlling risk at the source.

[0075] In some embodiments, the real-time detection module is configured with a first type of malicious behavior and a second type of malicious behavior, each of which includes three levels: minor, moderate, and severe.

[0076] For credential holders, malicious behavior is primarily identified through two dimensions: the number of times false or erroneous credential submissions are detected, and the rate of decline in trust value within adjacent calculation periods. The specific judgment logic is as follows:

[0077] Firstly, regarding the determination of the number of times false and incorrect credentials are identified, the system presets a first-time threshold to distinguish between occasional errors and malicious behavior. When the number of times false and incorrect credentials are identified exceeds the first-time threshold, a first-category malicious behavior is triggered. The severity of the first-category malicious behavior is further subdivided based on the proportion of the number of times false and incorrect credentials are identified exceeding the first-time threshold: if the proportion is less than 20%, it is classified as minor; if the proportion is between 20% and 50%, it is classified as moderate; and if the proportion exceeds 50%, it is classified as severe. The excess proportion is calculated as: (Number of times false and incorrect credentials are identified - First-time threshold) / First-time threshold × 100%.

[0078] Specifically, the initial threshold should be set based on statistical analysis of historical normal network behavior data. In the early stages of system operation, a reasonable initial threshold can be determined by analyzing internal testing data or referring to operational experience with similar systems. It is recommended that the initial threshold be set between 5 and 20 times. For example, in a moderately active network, the initial threshold can be set to 10 times; that is, when the same credential holder has submitted false and erroneous credentials 10 times, the system will automatically trigger the first type of malicious behavior detection mechanism.

[0079] Secondly, regarding the determination of the rate of decline of trust value within adjacent computation cycles, the system presets a first difference threshold to capture abnormally sharp drops in trust value, indicating a sudden deterioration in node behavior or an attack. When the rate of decline of trust value within adjacent computation cycles exceeds the first difference threshold, a second type of malicious behavior is triggered. The level of the second type of malicious behavior is further subdivided based on the proportion by which the rate of decline of trust value within adjacent computation cycles exceeds the first difference threshold: if the proportion is less than 20%, it is judged as a minor level; if the proportion is between 20% and 50%, it is judged as a moderate level; and if the proportion is greater than 50%, it is judged as a severe level.

[0080] Specifically, the first difference threshold should be set with reference to the normal fluctuation range of the trust value. It is recommended that the first difference threshold be set to 10%-20% of the maximum range of the trust value. For example, if the trust value range is 0-1, then the threshold can be set between 0.1 and 0.2. When the rate of decrease of the trust value exceeds the first difference threshold within a single calculation cycle, the detection of the second type of malicious behavior is triggered.

[0081] For credential issuers, malicious behavior is primarily determined by the number of times they are identified as issuing false or erroneous credentials, and the proportion of negative feedback they receive from credential verifiers. Specific conditions are as follows:

[0082] First, regarding the determination of the number of times false and erroneous vouchers are identified, the system presets a second threshold. When the number of times false and erroneous vouchers are identified exceeds the second threshold, a first-class malicious behavior is triggered. The level of the first-class malicious behavior is further subdivided according to the proportion of the number of times false and erroneous vouchers are identified exceeding the second threshold: if the proportion is less than 20%, it is judged as a minor level; if the proportion is between 20% and 50%, it is judged as a moderate level; if the proportion is greater than 50%, it is judged as a serious level.

[0083] Specifically, since credential issuers are responsible for issuing verifiable credentials, their malicious actions can have widespread cascading negative consequences. Therefore, the second count threshold should generally be more stringent than the first count threshold for credential holders. An initial range of 3 to 10 counts is recommended. For example, in a medium-sized network with high data authenticity requirements, a threshold of 5 counts could be set. That is, malicious behavior detection is triggered when an issuer has issued and been identified as issuing 5 counterfeit or erroneous credentials. This value can be set with reference to the average issuance success rate and anomaly detection rate of all issuers within the network during the observation period.

[0084] Secondly, regarding the determination of the proportion of negative reviews, the system presets a threshold, which reflects the degree of collective distrust among the validators towards a particular issuer. When the proportion of negative reviews in the total scores provided by all validators who have interacted with the issuer exceeds the threshold, a second type of malicious behavior is triggered. The level of the second type of malicious behavior is further subdivided based on the percentage of negative reviews exceeding the threshold: if the percentage exceeds 20%, it is judged as a minor level; if the percentage exceeds 20% to 50%, it is judged as a moderate level; and if the percentage exceeds 50%, it is judged as a severe level.

[0085] Specifically, it is recommended that the initial range for the percentage threshold be set to 5% to 20% of the total scores for negative ratings. For example, in networks that emphasize reputation accumulation, it could be set to 10%. That is, when negative ratings account for more than 10% of all ratings a single issuer receives from validators, malicious behavior detection based on reputation evaluation is triggered. This threshold can be determined based on the distribution of validator ratings in the early stages of network operation (such as the median or a specific quantile of negative ratings).

[0086] To identify malicious behavior by credential verifiers, the primary criterion is whether their verification rejection rate is abnormally higher than the network average, combined with a comprehensive assessment of the proportion of high-trust counterparties among the rejected requests.

[0087] Calculate the current credential verifier's rejection rate. The rejection rate is the proportion of the number of verification requests rejected by the credential verifier to the total number of verification requests processed within a preset calculation period.

[0088] When a current credential verifier's verification rejection rate exceeds the average verification rejection rate of all credential verifiers, and the proportion of high-trust counterparties among the rejected verification requests is below a certain threshold, a Type I malicious behavior is triggered. The excess percentage between the Type I malicious behavior verification rejection rate and the average verification rejection rate is further subdivided: if the excess percentage is less than 20%, it is classified as minor; if the excess percentage is between 20% and 50%, it is classified as moderate; and if the excess percentage is greater than 50%, it is classified as severe.

[0089] The threshold for determining malicious behavior by credential validators measures whether the proportion of requests rejected by a validator from high-trust counterparties is abnormal. This threshold directly affects the system's accuracy and response speed in identifying different malicious patterns such as "targeted rejection" or "indiscriminate rejection." This threshold aims to distinguish between normal validation strictness and malicious targeted behavior. A validator may occasionally reject requests from high-trust counterparties under reasonable questioning. An initial value of 15% to 35% is recommended. This type of behavior may manifest as indiscriminate rejection of low-reputation nodes.

[0090] When a current credential validator's rejection rate exceeds the average rejection rate of all credential validators, and the proportion of high-trust counterparties in rejected verification requests is not less than a certain threshold, a second type of malicious behavior is triggered. This type of behavior may manifest as targeted exclusion of high-reputation nodes. The severity of the second type of malicious behavior is further subdivided based on the percentage of high-trust counterparties in the current credential validator's rejected verification requests exceeding the threshold: if the percentage exceeds 20%, it is classified as minor; if the percentage exceeds 20% to 100%, it is classified as moderate; and if the percentage exceeds 100%, it is classified as severe.

[0091] When a node is detected as exhibiting malicious behavior by the real-time detection module, the dynamic penalty module penalizes the node's trust value, including:

[0092]

[0093] In the formula, T old T represents the trust value of a node before it is penalized. new This represents the trust value of a node after it has been penalized. L represents the basic penalty factor for the m-th type of malicious behavior. m Indicates the level of the m-th type of malicious behavior, n m The cumulative number of malicious acts of type m, where β represents the weighting coefficient for the number of acts. The time interval from the occurrence of the m-th type of malicious behavior to the present, where λ1 represents the time decay coefficient.

[0094] This invention proposes a distributed identity trust evaluation method based on a trust network, aiming to solve the technical challenges of identity trust management in decentralized environments. Traditional centralized trust models suffer from single points of failure and poor scalability, while existing distributed systems lack a quantitative evaluation mechanism for dynamic trust. This invention constructs a trust network based on blockchain to dynamically calculate and update the trust values ​​of the behaviors of credential holders, issuers, and verifiers, ensuring the secure and reliable operation of decentralized identity (DID) systems.

[0095] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "setting," "connection," "fixing," "rotation," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Unless otherwise explicitly limited, those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0096] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A distributed identity trust evaluation system based on a trust network, characterized in that, It includes multiple nodes, each of which is one of the following: a credential issuer, a credential holder, or a credential verifier; the credential holder is the party submitting the credential; the credential issuer issues a verifiable credential to the credential holder; the credential verifier verifies the verifiable credential of the credential holder; it also includes: A blockchain storage module is used to immutably store records of interactions between nodes. The trust calculation module is used to periodically calculate the trust value of nodes through smart contracts, specifically including: The holder trust calculation module is used to calculate the trust value of the certificate holder. The trust value of the certificate holder is obtained by weighted summation of the certificate holder's historical behavior trust score and transaction feedback trust score. The issuer trust calculation module is used to calculate the trust value of the credential issuer. The trust value of the credential issuer is obtained by weighted summation of the issuer's historical issuance trust score and verification feedback trust score. The verifier trust calculation module is used to calculate the trust value of the credential verifier. The trust value of the credential verifier is calculated based on the verification accuracy rate. The real-time detection module is used to monitor whether nodes are engaging in malicious behavior; The dynamic penalty module is used to penalize the trust value of the node that generates malicious behavior when malicious behavior is detected. The access control module monitors whether a node's trust value is lower than a preset security value. If so, it implements network access isolation measures for the node, restricting it from continuing to participate in network interactions.

2. The distributed identity trust evaluation system based on a trust network according to claim 1, characterized in that, The specific processing steps of the holder trust calculation module include: The historical behavior data of the credential holder includes the number of times the credential holder successfully submitted and was verified, the number of times false and erroneous credentials were identified, and the number of times the holder actively revoked invalid credentials. Based on the historical behavior data of credential holders, a Bayesian probability model is used to quantify the historical behavior trust score, which is expressed as: ， In the formula, T Hh S represents the trust score of the credential holder's historical behavior. H F represents the number of times a credential holder has successfully submitted and had it verified. H C represents the number of times a credential holder's submission of false or erroneous credentials has been identified. H λ represents the number of times the certificate holder actively revokes an invalid certificate. H The contribution factor for controlling the active revocation behavior is represented by μ, which represents the influence coefficient of the number of behaviors, and N represents the number of counterparties of the certificate holder. The transaction feedback trust score is calculated by tracking the ratings of counterparties who interact with the certificate holder, and is represented as follows: ， ， In the formula, T Hf R represents the transaction feedback trust score of the certificate holder. H,i T represents the rating provided by the i-th counterparty in the interaction with the credential holder. H,i D represents the trust value of the i-th counterparty interacting with the credential holder. H,i Let λ represent the time decay factor, and Δt represent the decay factor. H,i This indicates the time elapsed since the occurrence of the i-th transaction.

3. The distributed identity trust evaluation system based on a trust network according to claim 1, characterized in that, The specific processing steps of the issuer trust calculation module include: The historical signing behavior data of the voucher issuer includes the number of times the voucher issuer successfully issued and verified as valid vouchers, the number of times false and erroneous vouchers were identified, and the number of times the issuer voluntarily revoked invalid vouchers that had been issued. Based on the historical issuance behavior data of the credential issuer, a Bayesian probability model is used to quantify the historical issuance trust score, which is expressed as: ， In the formula, T Ih S represents the historical issuance trust score of the credential issuer. I F represents the number of times a credential issuer has successfully issued and verified a credential as valid. I C represents the number of times false and erroneous credentials have been identified. I λ indicates the number of times the issuer of a certificate has voluntarily revoked an invalid certificate that has already been issued. I The contribution factor for controlling the active revocation behavior is represented by μ', the influence coefficient of the number of behaviors is represented by N', and the number of credential validators that interact with the credential validator is represented by N. The credential verifiers who interact with the credential issuer rate the credential issuer, and calculate the verification feedback trust score, expressed as follows: ， In the formula, T If R represents the transaction feedback trust score of the credential issuer. I,j T represents the score given by the i-th credential validator who interacts with the credential issuer. I,j D represents the trust value of the i-th credential verifier that interacts with the credential issuer. I,j This represents the time decay factor.

4. The distributed identity trust evaluation system based on a trust network according to claim 1, characterized in that, The specific processing steps of the verifier trust calculation module include: The number of times a voucher verifier correctly completes the verification operation, the number of times erroneous verification occurs, and the number of times biased behavior is exhibited are quantified using an accuracy model to obtain a trust value, which is expressed as: ， Among them, T V S represents the trust value of the credential verifier. V F represents the number of times the credential verifier correctly completed the verification operation. V μ'' represents the number of times the credential validator makes incorrect validations and exhibits biased behavior, μ'' represents the influence coefficient of the number of behaviors, and N'' represents the number of counterparties of the credential validator.

5. A distributed identity trust evaluation system based on a trust network according to claim 1, characterized in that, The real-time detection module is configured with two types of malicious behavior, each of which includes three levels: minor, moderate, and severe. The detection process specifically includes: For certificate holders: When the number of times false and incorrect credentials are identified exceeds the first threshold, the first type of malicious behavior is triggered. The level of the first type of malicious behavior is further subdivided according to the proportion of the number of times false and incorrect credentials are identified to the first threshold: if the proportion is less than 20%, it is judged as a minor level; if the proportion is between 20% and 50%, it is judged as a moderate level; if the proportion is greater than 50%, it is judged as a serious level. When the rate of decline of the trust value in adjacent calculation cycles exceeds the first difference threshold, the second type of malicious behavior is triggered. The level of the second type of malicious behavior is further subdivided according to the proportion of the rate of decline of the trust value in adjacent calculation cycles exceeding the first difference threshold: if the proportion of ... For the certificate issuer: When the number of times false and erroneous vouchers are identified exceeds the second threshold, the first type of malicious behavior is triggered. The level of the first type of malicious behavior is further subdivided according to the proportion of the number of times false and erroneous vouchers are identified to the second threshold: if the proportion is less than 20%, it is judged as a minor level; if the proportion is between 20% and 50%, it is judged as a moderate level; if the proportion is greater than 50%, it is judged as a serious level. When the proportion of negative ratings in the total ratings provided by all credential verifiers who have interacted with the credential issuer exceeds a certain threshold, a second type of malicious behavior is triggered. The level of the second type of malicious behavior is further subdivided according to the proportion of negative ratings in the total ratings that exceeds the threshold: if the proportion exceeds the threshold by less than 20%, it is judged as a minor level; if the proportion exceeds the threshold by 20% to 50%, it is judged as a moderate level; if the proportion exceeds the threshold by more than 50%, it is judged as a severe level. For credential verifiers: Calculate the current credential verifier's rejection rate. The rejection rate is the proportion of the number of verification requests rejected by the credential verifier to the total number of verification requests processed within a preset calculation period. When the current credential verifier's verification rejection rate exceeds the average verification rejection rate of all credential verifiers, and the proportion of high-trust counterparties in the rejected verification requests is lower than the proportion threshold, a Type I malicious behavior is triggered. The excess ratio between the verification rejection rate and the average verification rejection rate for Type I malicious behavior is further subdivided: if the excess ratio is less than 20%, it is judged as a minor level; if the excess ratio is between 20% and 100%, it is judged as a moderate level; if the excess ratio is greater than 100%, it is judged as a severe level. When the current credential verifier's verification rejection rate exceeds the average verification rejection rate of all credential verifiers, and the proportion of high-trust counterparties in the rejected verification requests is not less than the percentage threshold, a second type of malicious behavior is triggered. The level of the second type of malicious behavior is further subdivided according to the excess ratio between the proportion of high-trust counterparties in the current credential verifier's rejected verification requests and the percentage threshold: if the excess ratio is less than 20%, it is judged as a minor level; if the excess ratio is between 20% and 100%, it is judged as a moderate level; if the excess ratio is greater than 100%, it is judged as a severe level.

6. A distributed identity trust evaluation system based on a trust network according to claim 1, characterized in that, When a node is detected as exhibiting malicious behavior by the real-time detection module, the dynamic penalty module penalizes the node's trust value, including: ， In the formula, T old T represents the trust value of a node before it is penalized. new This represents the trust value of a node after it has been penalized. L represents the basic penalty factor for the m-th type of malicious behavior. m Indicates the level of the m-th type of malicious behavior, n m The cumulative number of malicious acts of type m, where β represents the weighting coefficient for the number of acts. The time interval from the occurrence of the m-th type of malicious behavior to the present, where λ1 represents the time decay coefficient.

Images