Block chain-based computing power transaction platform and working method
By using blockchain decentralized storage and dynamic trust evaluation model on computing power trading platforms, combining direct trust and indirect trust, the problem of subjective bias in user feedback is solved, and the accuracy of trust evaluation and transaction reliability are improved. Achieves performance and cost balance and data security.
Patent Information
- Application Number
- CN202411101887.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-12-03
AI Technical Summary
Existing computing power trading platforms rely on user feedback to calculate trust, which has subjective bias. Malicious feedback may interfere with the trust evaluation process and affect transaction reliability.
Using blockchain to decentralize transaction records and node feedback, dynamically adjust the weight of direct and indirect trust, combine historical interactions and trusted node recommendations, improve the trust evaluation model, reduce the impact of malicious feedback, and select through reverse auctions Trusted node.
Improves the accuracy of trust evaluation, reduces subjectivity bias, ensures fair participation of new nodes, achieves a balance between performance and cost, and protects data privacy and security through data ciphertext hash.
Smart Images

Figure CN119067762A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computing power trading technology, and specifically to a computing power trading platform and working method based on blockchain. Background Art
[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.
[0003] As the demand for computing power resources increases, the local hardware equipment of users can no longer meet their own computing power needs. Therefore, users will publish computing power requirements to other nodes in the network, and use the computing power of other nodes to perform their own tasks by leasing or purchasing. The network formed in this way is a computing power trading platform.
[0004] When users select available nodes, in addition to considering the cost of renting or purchasing nodes, they also need to consider the behavior patterns and performance characteristics of the nodes to ensure the efficiency of their computing tasks and the security of their data.
[0005] In order to solve these problems, some computing power trading platforms control and maintain the trust relationship in the computing power network by combining blockchain and trust management mechanisms, helping users determine whether the nodes providing computing resources are trustworthy, and providing computing power trading platforms with a basis for rewarding and punishing transaction participants. However, existing solutions calculate trust based on user node feedback to determine whether a node is a trustworthy node, and user feedback is highly subjective. Untrustworthy nodes may cause deviations in evaluation results by submitting malicious feedback, interfering with the accuracy of the trust evaluation process, and thus affecting the reliability of transactions. Summary of the invention
[0006] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a computing power trading platform and working method based on blockchain, which uses blockchain to replace the traditional trading platform, and stores transaction records and service feedback of participating nodes in a decentralized manner. The trust evaluation model in the platform has been improved, and the weights of direct trust and indirect trust are dynamically adjusted to track the behavior changes of computing nodes; before each transaction settlement, the reliability of the feedback node can be measured to reduce the deviation of the evaluation result caused by malicious feedback submitted by untrustworthy nodes.
[0007] In order to achieve the above object, the present invention adopts the following technical solution:
[0008] The first aspect of the present invention provides a computing power trading platform based on blockchain, comprising:
[0009] The task node module is configured to: publish computing power demand tasks through the blockchain module;
[0010] The computing node module is configured to: send a response message according to the computing power requirement task issued by the blockchain module, receive the trust evaluation result sent by the smart contract module after becoming a candidate node for this transaction, and provide a quotation according to the computing power requirement task; after the transaction is reached, execute the computing power requirement task and send the result ciphertext to the task node module;
[0011] The blockchain module is configured to: match the corresponding computing nodes using the smart contract module according to the computing power demand tasks issued by the task node module;
[0012] The smart contract module is configured to: generate trust based on the trust evaluation model, select computing nodes with trust exceeding a set threshold as candidate nodes, and determine transaction nodes through quotations;
[0013] Trust includes direct trust and indirect trust. Based on the historical interactions and feedback with the computing nodes, the task node obtains the response ratio of the computing nodes to the task node's computing power requests within a set time interval, and the trust level of the task node to the computing node within the same set time period. The product of the two is the direct trust. Based on the recommendations of other trusted nodes, the indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated.
[0014] Furthermore, direct trust and indirect trust adaptively adjust their respective weights according to the volatility of the task node's direct trust in the computing node compared with the previous direct trust during this transaction, and determine the final total trust. The computing nodes whose total trust exceeds the set threshold are selected as candidate nodes.
[0015] Furthermore, the response message includes signing the digest of the respective identity and computing power information using the latest timestamp, and the computing node becomes a candidate computing node for this transaction by sending the response message.
[0016] Furthermore, based on the recommendations of other trusted nodes, indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated, including:
[0017] According to the recommendation node r, the direct trust of the computing node s at time t is used as the recommendation opinion, and the indirect trust of the task node u on the computing node s is calculated, as shown in the following formula:
[0018]
[0019] Among them, DT r,s represents the direct trust of the recommending node r on the computing node s within Δt, t u,r Represents the recommendation credibility of task node u to recommend node r.
[0020] Furthermore, based on the recommendations of other trusted nodes, indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated, which also includes:
[0021] The inverse deviation between individual feedback and global feedback is set to reflect the credibility of the recommendation, as shown in the following formula: u,r =f u,s Where, f u,s For feedback credibility;
[0022] Recommended credibility u,r The trust level of the task node u in the recommendation provided by the recommendation node r, which is calculated by the feedback credibility of the recommendation node To verify.
[0023] Furthermore, a quotation is provided according to the computing power demand task, specifically: the candidate node is the computing node that meets the demand of this transaction and is selected to participate;
[0024] The set of candidate nodes that are higher than the set threshold is the candidate node set;
[0025] According to the auction rules, bidder V i Interaction continues to give lower quotes P i , until no new bids appear, and the winning bidder V is selected * Provide computing power services for task nodes and return bidding results to task nodes.
[0026] Furthermore, after the transaction is concluded, the task node module uses the public key of the selected computing node to encrypt the data of the computing power demand task, and sends the data ciphertext hash value of the computing power demand task to the blockchain module, and sends the data ciphertext of the computing power demand task to the computing node module;
[0027] The computing node module obtains the data ciphertext hash value of the computing power requirement task through the blockchain module. After verification, it uses the private key of the computing node to decrypt the data ciphertext of the computing power requirement task and executes the computing power requirement task.
[0028] Furthermore, after the task is executed, the computing node module encrypts the calculation result of the computing power requirement task through the public key of the task node, and stores the encrypted result ciphertext hash value in the blockchain;
[0029] The task node module obtains the ciphertext hash value of the result of the computing power requirement task through the blockchain module. After verification, it uses the private key of the task node to decrypt the ciphertext of the result of the computing power requirement task and provides feedback based on the calculation results.
[0030] Furthermore, the blockchain module monitors the changing trend of the trust of the computing node. If the trust changes by more than the set value within the set time period, the computing node is a potential malicious service node. Based on the node feedback score, the abnormality of the autonomous dynamic service behavior of the computing node is determined.
[0031] The second aspect of the present invention provides a working method of a computing power trading platform based on blockchain, comprising the following steps:
[0032] The task node module publishes computing power demand tasks to candidate computing nodes through the blockchain module;
[0033] The computing node module sends a response message according to the computing power demand task forwarded by the blockchain module, receives the trust evaluation result sent by the smart contract module, and provides a quotation according to the computing power demand task; after the transaction is reached, it executes the computing power demand task and sends the result ciphertext to the task node module;
[0034] The blockchain module uses the smart contract module to match the corresponding computing nodes according to the computing power demand tasks issued by the task node module;
[0035] The smart contract module generates trust based on the trust evaluation model, selects computing nodes with trust exceeding the set threshold as candidate nodes, and determines the transaction node through quotation;
[0036] Trust includes direct trust and indirect trust. Based on the historical interactions and feedback with the computing nodes, the task node obtains the response ratio of the computing nodes to the task node's computing power requests within a set time interval, and the trust level of the task node to the computing node within the same set time period. The product of the two is the direct trust. Based on the recommendations of other trusted nodes, the indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated.
[0037] Compared with the prior art, one or more of the above technical solutions have the following beneficial effects:
[0038] 1. After the computing power demand task is released, the candidate computing nodes that can bear the computing power demand are screened out through a combination of direct trust and indirect trust. Direct trust is determined based on the historical interaction and feedback between the task node and the computing node, and indirect trust is generated based on the recommendation list obtained from the computing node's recommendation. The combination of direct trust and indirect trust can reduce subjectivity as much as possible, minimize the deviation of the evaluation results caused by malicious feedback, improve the accuracy of the trust evaluation process, and also allow computing nodes newly added to the platform to participate in transactions in a relatively fair manner without historical transaction data.
[0039] 2. Before the transaction, the computing nodes are screened by trust and untrustworthy nodes are removed, effectively reducing the communication and computing pressure. Based on the reverse auction strategy, the transaction parties are selected from the nodes that meet the computing power requirements and are trustworthy, achieving a balance between performance and cost.
[0040] 3. During the execution of tasks that require computing power, the blockchain module only saves the hash value of the data ciphertext and completes the integrity check of the data ciphertext, but does not save the data ciphertext itself. The data ciphertext is transmitted interactively by the computing node and the task node that have reached the transaction. This can not only protect the privacy and security of the data, but also use blockchain technology to ensure that the data cannot be tampered with and is consistent, and can also reduce the storage occupancy of the blockchain.
[0041] 4. There is anomaly detection during the feedback reliability evaluation after the computing power demand task is completed. When a sudden change in trust is monitored, potential malicious service nodes can be identified; the overall reduction in node feedback scores can also indicate the abnormality of the autonomous dynamic service behavior of the computing node, providing a basis for identifying malicious service nodes. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.
[0043] Figure 1 It is a schematic diagram of the architecture of a computing power trading platform based on blockchain provided by one or more embodiments of the present invention;
[0044] Figure 2 is a schematic diagram of a trust evaluation model in a computing power trading platform provided by one or more embodiments of the present invention;
[0045] Figure 3 It is a schematic diagram of a transaction party selection process in a computing power transaction platform provided by one or more embodiments of the present invention;
[0046] Figure 4 It is a schematic diagram of the working process of the computing power trading platform provided by one or more embodiments of the present invention. DETAILED DESCRIPTION
[0047] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0048] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.
[0049] As introduced in the background technology, when users select available nodes, in addition to considering the cost issues brought by leasing or purchasing nodes, they also need to consider the behavior patterns and performance characteristics of the nodes to ensure the efficiency of their own computing tasks and the security of data. In order to solve these problems, existing solutions calculate the trust based on the feedback of user nodes to determine whether the node is a trusted node. However, user feedback is highly subjective, and untrustworthy nodes may cause deviations in the evaluation results by submitting malicious feedback, interfering with the accuracy of the trust evaluation process, and thus affecting the reliability of transactions.
[0050] In addition, in terms of the selection of transaction parties, if trust is used as one of the criteria for selecting transaction parties, users will usually choose nodes with high trust, which easily ignores the balance between performance and cost. It also makes it almost impossible for newly added nodes to have a high trust level, and thus they cannot enjoy the opportunity to participate in transactions fairly.
[0051] Therefore, the following embodiments provide a computing power trading platform and working method based on blockchain, using blockchain to replace the traditional trading platform, storing transaction records and service feedback of participating nodes in a decentralized manner, and improving the trust evaluation model in the platform. The weights of direct trust and indirect trust are adjusted dynamically based on volatility, and the trust of nodes is accurately and effectively evaluated in combination with historical interactions between nodes and recommendations of trusted nodes; the computing nodes are screened in combination with the trust evaluation results and reverse auctions to select the final trading parties. Before the transaction, the computing nodes are accurately and effectively evaluated for trust based on historical feedback, and the weights of direct trust and indirect trust are dynamically adjusted using adaptive functions to track the behavioral changes of computing nodes; before each transaction is settled, the reliability of the feedback nodes can be measured to reduce the deviation of the evaluation results caused by malicious feedback submitted by untrusted nodes, and the feedback and its reliability are securely stored on the chain; in conjunction with the reverse auction algorithm, the trading parties are selected from the nodes that meet the computing power requirements and are trustworthy, achieving a balance between performance and cost.
[0052] Terminology explanation:
[0053] Blockchain, a decentralized distributed ledger technology, was originally designed for cryptocurrencies such as Bitcoin. It ensures the security and consistency of data through encryption algorithms and consensus mechanisms. Due to its open, transparent, decentralized, traceable, and tamper-proof characteristics, it has quickly become a hot topic in academic research. Blockchain technology is widely used, including but not limited to finance, supply chain management, the Internet of Things, healthcare, digital identity authentication and other fields. The emergence and development of blockchain is expected to change the traditional way of trust and transaction and bring more innovation and opportunities.
[0054] Trust management, a certain degree of subjective belief in the behavior of a particular entity, is a multidimensional concept that spans multiple domains. Formulating strategies to evaluate and maintain trust among different entities or systems is called Trust Management (TM). For service-oriented environments such as IoT, TM acts as an intermediary between service requesters and providers to facilitate trustworthy interactions, providing multiple security services such as identity management, reliable service composition, access control, resource management, etc.
[0055] The hybrid trust management model considers two sources of trust: direct trust and indirect trust to describe and measure the trust relationship between one party and another. It is shown in the following formula:
[0056]
[0057] Direct Trust The indirect trust that party i has in party j is generated by historical interactions, which include both positive and negative interactions. It is composed of the recommendations of other trusted recommendation nodes k and their recommendation credibility. The recommendation is the direct trust of recommendation node k in j within Δt. The recommendation credibility represents the degree of trust that i has in the recommendation given by recommendation node k, which is measured by the similarity between the recommendation and individual feedback.
[0058] Reverse Auction, also known as reverse bidding, reverse auction, reverse auction, etc., is led by the buyer. The purchaser (buyer) provides a detailed description of the goods to be purchased, and multiple suppliers continue to give quotations. Finally, the lowest bidder wins the bid. The process is as follows:
[0059] 1) Buyer publishes demand D i , including demand quantity Q, quality standard S, delivery time T and other conditions.
[0060] D i ={Q,S,T}
[0061] 2)Various suppliers i According to demandD i Provide a quote i , and may include other service conditions C i .
[0062] P i =f(D i ,C i )
[0063] 3) The buyer selects the best offer P based on the set evaluation function g * and the corresponding supplier V * .
[0064] (P * ,V* )=argmin(g(P i ,V i ))
[0065] 4) Buyer and successful bidder V * Reach a formal contract and agree on specific service or product delivery details.
[0066] Embodiment 1:
[0067] The computing power trading platform based on blockchain includes:
[0068] The task node module is used to publish computing power demand tasks to the blockchain module, and send the ciphertext hash value of the computing task data to the blockchain module after the transaction is completed. In the computing power network, nodes need to request computing resources from other nodes to complete computing tasks due to their own limited resources.
[0069] The computing node module sends a response message according to the computing power demand task forwarded by the blockchain module; receives the trust assessment result sent by the smart contract module; provides a quotation according to the computing power demand task; receives the screening result sent by the smart contract module; takes the hash of the computing power demand task data ciphertext sent by the task node module, and compares and verifies it with the computing power demand task data ciphertext hash value provided by the blockchain module. After the verification is passed, the computing power demand task is executed, the result ciphertext is sent to the task node module, and the result ciphertext hash value is sent to the blockchain module; and the transaction settlement is carried out through the smart contract module and the task node module.
[0070] The blockchain module uses the smart contract module to match the corresponding computing nodes according to the computing power demand tasks issued by the task node module; stores the trust data obtained from the smart contract module, the response message sent from the computing node module, and the service feedback from the task node to the computing node; after the task node and the computing node reach a transaction, stores the ciphertext hash value of the computing power demand task data sent by the task node; after the computing task is executed, receives the ciphertext hash value of the computing power demand task result sent by the computing node; and saves transaction records.
[0071] In this embodiment, the blockchain module registers the identity and computing power of the task nodes and computing nodes within the system. The computing power transactions are managed and executed through distributed ledgers, smart contract modules, consensus mechanisms and decentralized storage. Task nodes and computing nodes interact directly through the blockchain platform to ensure the transparency and automation of transactions. Task nodes can browse resources on the trading platform and submit corresponding computing power requests. Transaction records are stored in the blockchain to ensure the transparency and immutability of transactions.
[0072] The smart contract module matches computing nodes according to computing power requirements and generates trust indicators to measure the credibility of response nodes during transactions based on the trust evaluation model. This is used to evaluate the services of response nodes in order to provide a reference for other task nodes to select transaction parties.
[0073] Trust includes direct trust and indirect trust. During this transaction, the task node's direct trust in the computing node is adjusted by the volatility of the previous direct trust to determine the final total trust. The smart contract module selects computing nodes whose total trust exceeds the set threshold and obtains bids through the reverse auction stage.
[0074] Based on the historical interactions and feedback with the computing nodes, the task node obtains the response ratio of the computing nodes to the task node's computing power requests within the set time interval, and the trust level of the task node to the computing node within the same set time period. The product of the two is the direct trust;
[0075] Based on the recommendations of other trusted nodes, indirect trust for each candidate task node is generated, and the trusted recommendation list is updated.
[0076] The above computing power trading platform adopts the following Figure 1 and Figure 4 The method shown is as follows:
[0077] 1. Node registration stage: The blockchain completes the identity registration of all nodes in the system. Among them, the computing nodes that serve as resource providers must also complete the computing power registration and update the computing power that can be provided in real time. The hash value of the registration information is stored on the blockchain.
[0078] 2. Task publishing stage: As the task publisher, the task node set is defined as U = {u1,u2,...,u max}, the computing node set is defined as S = {s1,s2,...,s max}, task node u∈U can browse computing resources on the trading platform and publish corresponding computing resource requirements D i ,After the smart contract module and the task node module interact and complete the order confirmation, all computing nodes that have registered computing power and meet the computing power resource requirements are notified to respond to the request.
[0079] 3. Response phase: The computing nodes that want to participate in this transaction send a response message to the smart contract module of the trading platform. The response message includes a signature of the summary of their identity and computing power information using the latest timestamp, becoming candidate nodes for this transaction.
[0080] 4. Trust evaluation stage: The smart contract module of the trading platform receives the response message from the candidate node, and calculates the direct trust of each candidate node based on the trust evaluation model and the historical interaction data with the candidate node downloaded from the blockchain; at the same time, the smart contract module of the trading platform generates indirect trust for each candidate node based on the recommendation of the trusted node, and continuously updates the trusted recommendation list; the smart contract module combines the direct trust and indirect trust of the candidate node to obtain the final trust. The specific process is as follows: Figure 2 and as shown in the following steps.
[0081] In this embodiment, the total trust is represented by TT u,s =α·DT u,s +β·IT u,s Indicates that, α and β represent direct trust weight and indirect trust weight respectively:
[0082]
[0083] In this formula, κ is the volatility adjustment coefficient, and σ is the volatility of u's direct trust in s this time compared with the previous direct trust.
[0084] 4.1. Assume that task node u performs trust evaluation on computing node s;
[0085] 4.2. Based on the historical interactions and feedback between task node u and computing node s, we can calculate the response ratio ρ of s to u’s computing power request within Δt. s,u and u's trust level in s within Δt To save computational overhead, only interactions within a fixed time interval Δt are considered here, and direct trust is calculated based on the two
[0086] 4.2.1. Computing node s returns a response message if it is willing to share computing power according to its own capabilities. Computing node s may not always respond to the computing power request issued by task node u. Therefore, the response ratio of computing node s to the computing power request issued by task node u is:
[0087]
[0088] In this formula, k refers to the number of times s responds to requests from task node u, and q refers to the total number of times task node u issues requests.
[0089] 4.2.2. Based on the feedback given by task node u in the historical interaction with computing node s, the trust level of task node u to computing node s is calculated according to the level of each feedback and the time decay of the feedback:
[0090]
[0091] In this formula, r u,s is the service feedback of task node u to s within Δt, e -λt represents the attenuation of feedback, and λ represents the attenuation coefficient.
[0092] 4.2.3. Calculate the direct trust of task node u in computing node s based on the response ratio of computing node s to the computing power request issued by task node u and the trust level of task node u in computing node s: DT u,s .
[0093] 4.3. Based on the recommendation of other trusted nodes r, task node u calculates the indirect trust of task node u on computing node s;
[0094] 4.3.1. Based on the recommendation node r, the direct trust of the computing node s at time t is used as the recommendation opinion, and the indirect trust of the task node u on the computing node s is calculated:
[0095]
[0096] Among them, DT r,s represents the direct trust of the recommending node r on the computing node s within Δt, t u,r Represents the credibility of the recommendation of task node u to the recommendation node r. Considering that the credibility of recommendation is not a subjective concept, the credibility of the recommendation cannot be measured only by the similarity between the recommendation and the individual feedback. Therefore, this trust evaluation model improves the traditional hybrid evaluation model and sets the inverse deviation of individual feedback and global feedback to reflect the credibility of the recommendation.
[0097] t u,r =f u,s
[0098] In this formula, f u,s For feedback credibility.
[0099] 4.3.2. Recommendation credibility indicates the degree of trust that task node u has in the recommendation provided by recommender node r. To verify.
[0100] 4.4 Calculation of comprehensive trust
[0101] The total trust represents the comprehensive trust degree TT of the task node u to the computing node s u,s =α·DT u,s +β·IT u,s .
[0102] In order to dynamically adjust the trust evaluation model, a volatility-based adaptability function is designed to adaptively adjust the weights α and β:
[0103]
[0104] Volatility is a measure of the degree of change. Among them, DT i is the ith value in u’s direct trust set for s, is the mean of all direct trusts, and n is the total number of direct interactions between u and s in the time interval.
[0105] 5. Threshold screening: Based on the total trust, the trading platform selects candidate nodes above a certain threshold to enter the reverse auction stage, such as Figure 3 As shown in the figure, in the reverse auction stage, the candidate nodes are computing nodes that meet the transaction requirements and choose to participate; the set of candidate nodes above a certain threshold is the candidate node set; according to the auction rules, the bidder V i Interaction continues to give lower quotes P i , until no new bids appear, and the winning bidder V is selected * Provide computing power services for task nodes and return bidding results to task nodes (i.e. Figure 1 “6 Upload Bidding Results” shown).
[0106] 7. Signing a contract: During the transaction phase, the trading platform automatically generates a transaction order and an electronic contract based on the information of both parties, and provides them to both parties for signing. After signing the contract, both parties sign it with their own private keys. The electronic contract is stored in an offline database, and only the Hash value of the contract is stored on the chain to alleviate the storage pressure of the blockchain. The trading platform notifies the computing node to provide computing resources to the task node. This not only ensures the security and immutability of the contract, but also effectively reduces the storage burden of the blockchain.
[0107] 8. Encryption phase: The task node uses the public key of the final selected computing node to encrypt the calculation data and send it to the node. At the same time, the ciphertext hash is stored on the blockchain. The computing node completes the integrity check of the data ciphertext by comparing the hash value. The computing node encrypts the calculation result in the same encryption method and returns it to the task node. The task node also completes the integrity check (corresponding to Figure 1 In this way, the privacy and security of data can be protected, and the blockchain technology can be used to ensure that the data cannot be tampered with and is consistent.
[0108] Feedback stage: After completing the computing task, the task node releases resources and provides feedback on the transaction based on indicators such as results, time consumption, and price (i.e., 10 transaction feedback and 11 feedback reliability evaluation), and stores it on the chain (12 storage feedback and reliability). The feedback will be used for the selection of trading parties when the task node releases the task next time.
[0109] There is anomaly detection during the feedback reliability assessment: in each round of interaction between the trading platform and the participants, the trading platform will perform anomaly detection on the trading parties and task nodes, and record the behavior information of both parties; the abnormal nodes detected by anomaly detection will exit the transaction. At the same time, the trading platform will record the behavior information of both parties in each round of transactions; specifically including:
[0110] Abnormal detection of transaction parties. Through trust, on the one hand, blockchain can be used to monitor the trust change trend of computing nodes. If a sudden change in trust is detected, potential malicious service nodes can be identified through trust. On the other hand, the overall reduction in node feedback scores can also indicate the abnormality of the autonomous dynamic service behavior of computing nodes, providing a basis for identifying malicious service nodes.
[0111] b. Abnormal detection of task nodes. Feedback reliability indicates the credibility of the service feedback given by the node. Since individual feedback is subjective, if there is no effective method to evaluate its reliability, it is easy to cause distortion of node trust.
[0112] Therefore, a method for evaluating the reliability of feedback is designed to prevent abnormal feedback from interfering with the accuracy of the trust assessment process. It includes the following steps:
[0113] Assume that within Δt, task node u provides a service feedback to s, denoted as r u,s The global feedback is calculated by aggregating the feedback from all other nodes to s:
[0114]
[0115] in, It represents the average trust value of task node u within the time interval Δt, and uses the average trust value as the weight of service feedback to participate in the calculation of global feedback.
[0116] Calculating the credibility of feedback Finally, the reliability of feedback is measured by aggregating the credibility of feedback within a fixed time interval Δt Task nodes whose feedback reliability is lower than a certain threshold are considered abnormal nodes and lose their trading rights.
[0117] Finally, it is the settlement stage: the trading platform completes the verification by comparing the contract hash value to prevent transaction disputes caused by contract tampering. The trading platform automatically completes the settlement and the computing power transaction ends.
[0118] When the task node issues a new transaction request, the above process is repeated. The trading platform is responsible for maintaining the historical interaction information table and trusted recommendation list of all participants; the historical interaction information table stores historical transaction records; the trusted recommendation list stores the recommendation credibility of each recommendation node, and the recommendation credibility of each recommendation node constitutes a recommendation credibility queue, which is sorted in chronological order and stores a maximum of k items; if the queue exceeds the maximum storage limit, it is considered full. When a new recommendation credibility is received, the element at the head of the current queue is deleted, and then the latest recommendation credibility is inserted to update the recommendation list.
[0119] After the computing power demand task is released, the above-mentioned platform screens out candidate computing nodes that can bear the computing power demand through a combination of direct trust and indirect trust. Direct trust is determined based on the historical interaction and feedback between the task node and the computing node, and indirect trust is generated based on the recommendation list obtained from the computing node's recommendation opinions. The combination of direct trust and indirect trust can reduce subjectivity as much as possible, minimize the deviation of evaluation results caused by malicious feedback, improve the accuracy of the trust evaluation process, and also allow computing nodes newly joining the platform to participate in transactions in a relatively fair manner without historical transaction data.
[0120] Before a transaction, computing nodes are screened by trust and untrustworthy nodes are removed, effectively reducing communication and computing pressure. Based on a reverse auction strategy, transaction parties are selected from trustworthy nodes that meet computing power requirements, achieving a balance between performance and cost.
[0121] During the execution of tasks that require computing power, the blockchain module only saves the hash value of the data ciphertext and completes the integrity check of the data ciphertext, but does not save the data ciphertext itself. The data ciphertext is transmitted interactively by the computing node and the task node that complete the transaction. This can not only protect the privacy and security of the data, but also use blockchain technology to ensure that the data cannot be tampered with and is consistent, and can also reduce the storage occupancy of the blockchain.
[0122] There is anomaly detection during the feedback reliability evaluation after the computing power demand task is completed. When a sudden change in trust is monitored, potential malicious service nodes can be identified; the overall reduction in node feedback scores can also indicate the abnormality of the autonomous dynamic service behavior of the computing node, providing a basis for identifying malicious service nodes.
[0123] Blockchain is used instead of traditional platforms to realize the on-chain release of computing power requests, computing power usage requests, and feedback result storage. At the same time, when nodes participate in computing power transactions, the legitimacy of the nodes is verified to effectively prevent untrusted nodes from impersonating legitimate nodes to provide false computing power services and obtain improper rewards. Ultimately, a safe and reliable computing power transaction process is achieved through on-chain and off-chain collaboration.
[0124] During the trust evaluation process, the evaluation weights are dynamically adjusted through adaptive functions to effectively track changes in computing node behavior. The difference between global feedback and individual feedback is used to generate an inverse deviation function to reflect feedback credibility, and the feedback credibility is aggregated to calculate feedback reliability. This avoids the disadvantage of low accuracy of trust evaluation results due to feedback subjectivity, effectively identifies malicious feedback nodes, prevents collusion attacks, and ultimately achieves accurate and effective trust evaluation of nodes.
[0125] Embodiment 2:
[0126] The working method of the computing power trading platform based on blockchain includes the following steps:
[0127] The task node module publishes computing power demand tasks to candidate computing nodes through the blockchain module;
[0128] The computing node module sends a response message based on the computing power demand task forwarded by the blockchain module, receives the trust evaluation result sent by the smart contract module, and provides a quotation based on the computing power demand task; after the transaction is reached, it executes the computing power demand task and sends the result ciphertext to the task node module;
[0129] The blockchain module uses the smart contract module to match the corresponding computing nodes according to the computing power demand tasks issued by the task node module;
[0130] The smart contract module generates trust based on the trust evaluation model, selects computing nodes with trust exceeding the set threshold as candidate nodes, and determines the transaction node through quotation;
[0131] Trust includes direct trust and indirect trust. Based on the historical interactions and feedback with the computing nodes, the task node obtains the response ratio of the computing nodes to the task node's computing power requests within a set time interval, and the trust level of the task node to the computing node within the same set time period. The product of the two is the direct trust. Based on the recommendations of other trusted nodes, the indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated.
[0132] After the computing power requirement task is released, candidate computing nodes that can bear the computing power requirement are screened out through a combination of direct trust and indirect trust. Direct trust is determined based on the historical interaction and feedback between the task node and the computing node, while indirect trust is generated based on the recommendation list obtained from the computing node's recommendation opinions. The combination of direct trust and indirect trust can reduce subjectivity as much as possible, minimize the deviation of evaluation results caused by malicious feedback, improve the accuracy of the trust evaluation process, and also allow computing nodes newly added to the platform to participate in transactions in a relatively fair manner without historical transaction data.
[0133] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A computing power trading platform based on blockchain, characterized by: include: The task node module is configured to: publish computing power demand tasks through the blockchain module; The computing node module is configured to: send a response message according to the computing power requirement task issued by the blockchain module, receive the trust evaluation result sent by the smart contract module after becoming a candidate node for this transaction, and provide a quotation according to the computing power requirement task; after the transaction is reached, execute the computing power requirement task and send the result ciphertext to the task node module; The blockchain module is configured to: match the corresponding computing nodes using the smart contract module according to the computing power demand tasks issued by the task node module; The smart contract module is configured to: generate trust based on the trust evaluation model, select computing nodes with trust exceeding a set threshold as candidate nodes, and determine transaction nodes through quotations; Trust includes direct trust and indirect trust. Based on the historical interactions and feedback with the computing nodes, the task node obtains the response ratio of the computing node to the task node's computing power request within a set time interval, and the trust level of the task node to the computing node within the same set time period. The product of the two is the direct trust. Based on the recommendations of other trusted nodes, indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated.
2. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: Direct trust and indirect trust adaptively adjust their respective weights according to the volatility of the task node's direct trust in the computing node compared with the previous direct trust during this transaction, and determine the final total trust. The computing nodes whose total trust exceeds the set threshold are selected as candidate nodes.
3. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: The response message includes signing the digest of the respective identity and computing power information using the latest timestamp. The computing node becomes a candidate computing node for this transaction by sending the response message.
4. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: Based on the recommendations of other trusted nodes, generate indirect trust for each candidate task node and update the trusted node recommendation list, including: According to the recommendation node r, the direct trust of the computing node s at time t is used as the recommendation opinion, and the indirect trust of the task node u on the computing node s is calculated, as shown in the following formula: Among them, DT r,s represents the direct trust of the recommending node r on the computing node s within Δt, t u,r Represents the recommendation credibility of task node u to recommend node r.
5. The computing power trading platform based on blockchain as claimed in claim 4, characterized in that: Based on the recommendations of other trusted nodes, generate indirect trust for each candidate task node and update the trusted node recommendation list, including: The inverse deviation between individual feedback and global feedback is set to reflect the credibility of the recommendation, as shown in the following formula: u,r =f u,s Where, f u,s For feedback credibility; The recommendation credibility is the degree of trust that the task node u has in the recommendation provided by the recommendation node r. To verify.
6. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: Provide quotations based on computing power demand tasks, specifically: Candidate nodes are computing nodes that meet the requirements of this transaction and are selected to participate; The set of candidate nodes that are higher than the set threshold is the candidate node set; According to the auction rules, bidder V i Interaction continues to give lower quotes P i , until no new bids appear, and the winning bidder V is selected * Provide computing power services for task nodes and return bidding results to task nodes.
7. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: After the transaction is concluded, the task node module uses the public key of the selected computing node to encrypt the data of the computing power demand task, and sends the data ciphertext hash value of the computing power demand task to the blockchain module, and sends the data ciphertext of the computing power demand task to the computing node module; The computing node module obtains the data ciphertext hash value of the computing power requirement task through the blockchain module. After verification, it uses the private key of the computing node to decrypt the data ciphertext of the computing power requirement task and executes the computing power requirement task.
8. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: After the task is completed, the computing node module encrypts the calculation result of the computing power requirement task through the public key of the task node, and stores the encrypted result ciphertext hash value in the blockchain; The task node module obtains the ciphertext hash value of the result of the computing power requirement task through the blockchain module. After verification, it uses the private key of the task node to decrypt the ciphertext of the result of the computing power requirement task and provides feedback based on the calculation results.
9. The computing power trading platform based on blockchain as claimed in claim 1, characterized in that: The blockchain module monitors the trust change trend of the computing node. If the trust changes by more than the set value within the set time period, the computing node is a potential malicious service node. Based on the node feedback score, the abnormality of the autonomous dynamic service behavior of the computing node is determined.
10. The working method of the computing power trading platform based on blockchain is characterized in that: The following steps are involved: The task node module publishes computing power demand tasks to candidate computing nodes through the blockchain module; The computing node module sends a response message based on the computing power demand task forwarded by the blockchain module, receives the trust evaluation result sent by the smart contract module, and provides a quotation based on the computing power demand task; after the transaction is reached, it executes the computing power demand task and sends the result ciphertext to the task node module; The blockchain module uses the smart contract module to match the corresponding computing nodes according to the computing power demand tasks issued by the task node module; The smart contract module generates trust based on the trust evaluation model, selects computing nodes with trust exceeding the set threshold as candidate nodes, and determines the transaction node through quotation; Trust includes direct trust and indirect trust. Based on the historical interactions and feedback with the computing nodes, the task node obtains the response ratio of the computing node to the task node's computing power request within a set time interval, and the trust level of the task node to the computing node within the same set time period. The product of the two is the direct trust. Based on the recommendations of other trusted nodes, indirect trust for each candidate task node is generated, and the trusted node recommendation list is updated.
Citation Information
Cited By
Computing power bidding method and computing power bidding system
CN119991228A
