A Traceable Agent Re-encrypted Data Sharing Method for Satellite Networks

By employing the intra-cell product proxy re-encryption method, the problem of tracking and revoking user-level leakage behavior in satellite networks is solved. This enables secure data sharing and verification of re-encryption results in dynamic user environments, adapting to dynamic changes in satellite network links and updates in user authorization status.

CN123074684APending Publication Date: 2026-11-10CHANGCHUN UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611556602.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-09
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

In low-Earth orbit satellite constellations and integrated space-air-ground networks, existing data sharing schemes are unable to effectively track user-level leaks, user revocation controls are inconsistent, re-encryption results lack independent verification, and security is insufficient in quantum computing environments.

Method used

The lattice product proxy re-encryption method is adopted. The lattice product proxy re-encryption parameters are generated through the key generation center. The public tracking vector matrix, homomorphic signature private key and verification public key are used to realize user-level revocation control and independent verification of re-encryption key. Malicious users are located through black box tracking technology.

Benefits of technology

It achieves user-level revocation control and plaintext recovery control in a dynamic user environment, ensuring quantum-resistant security of data sharing, and can independently verify re-encryption results, adapting to dynamic changes in links and dynamic updates of user authorization status in satellite networks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123074684A_ABST
    Figure CN123074684A_ABST
Patent Text Reader

Abstract

This invention relates to the field of satellite network security technology, specifically to a traceable proxy re-encryption data sharing method for satellite networks. Addressing the problems of difficult-to-track leaks of decryption capabilities, unstable revocation control, and difficulty in verifying re-encryption results in satellite networks, this invention establishes lattice-product proxy re-encryption parameters and a full-rank public tracking vector matrix. A revocation verification vector is generated based on the revocation set, and the plaintext is encoded into a vector message. A re-encryption key is generated and signed according to an attribute access policy. The cloud server performs ciphertext evaluation and matrix switching. The delegate verifies the conversion result and uses a function to decrypt the key and recover the plaintext. Malicious users are located by challenging the black-box output probability difference of the ciphertext, and the revocation set is updated. This method is applicable to cross-node, cross-domain secure data sharing in low-Earth orbit satellites and integrated space-air-ground networks.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of satellite network security technology, specifically to a traceable proxy re-encrypted data sharing method for satellite networks. Background Technology

[0002] In low-Earth orbit satellite constellations and integrated space-air-ground networks, satellite nodes, ground stations, and cloud servers exchange remote sensing data, status data, and operational data via dynamic satellite links. Existing satellite network data sharing typically involves the data owner encrypting the plaintext operational data, transmitting the ciphertext via relay nodes to the cloud server, and then the cloud server performing access control and ciphertext conversion based on the receiving user's attributes and preset access policies. Finally, authorized users decrypt and use the data, maintaining data confidentiality and limiting access scope across open links and cross-domain nodes.

[0003] Satellite node movement and link switching cause dynamic changes in shared paths, and multiple users may have the same or similar access attributes. Existing solutions typically rely on keys bound to attribute sets and attribute-level revocation mechanisms. When decryption capabilities are leaked, it is difficult to further pinpoint the specific registered user. Furthermore, as the number of registered users changes and user revocation statuses are updated, existing solutions may also suffer from inconsistent user-level revocation relationships and decryption capabilities. Meanwhile, re-encryption results are usually generated by cloud servers and relayed, leaving the recipient without independent verification of the ciphertext conversion process. The computational difficulties relied upon by some traditional public-key cryptography schemes face security risks in quantum computing environments. Therefore, a traceable proxy re-encryption data sharing method for satellite networks is needed to improve existing technologies in areas such as user-level leakage tracking and revocation, revocation control in dynamic user environments, re-encryption result verification, and cryptographic security in quantum computing environments. Summary of the Invention

[0004] In view of the above-mentioned problems in the prior art, the present invention provides the following technical solution: A method for traceable agent re-encrypted data sharing in satellite networks includes: S1. The key generation center establishes the lattice product proxy re-encryption parameters, generates the lattice trapdoor, user master key, public trace vector matrix, homomorphic signature private key and verification public key, and decrypts the key according to the user public trace vector derivation function. S2. The data owner constructs a revocation verification vector based on the public tracking vector matrix and the revocation set, multiplies the plaintext scalar component by component into the revocation verification vector to form a vector message and encrypts it into the original ciphertext. S3. The key generation center generates a re-encryption key according to the authorization request, the principal user's master key, the trapdoor, and the entrusted party's public key, and generates a line signature for the re-encryption key line by line using the homomorphic signature private key. S4. The cloud server performs attribute evaluation on the original ciphertext. When the attribute vector corresponding to the original ciphertext satisfies the attribute access policy and the delegate has not revoked, the re-encryption key is used to perform matrix switching to obtain the re-encrypted ciphertext. S5. The entrusted party performs homomorphic evaluation on the line signature based on the original ciphertext and the attribute access policy to obtain the transformed signature, and uses the verification public key to verify the consistency between the transformed signature and the re-encrypted ciphertext. S6. After successful verification, the entrusted party uses the function decryption key to obtain the inner product of the vector message and the entrusted party's public tracking vector from the re-encrypted ciphertext, and recovers the plaintext based on the inner product of the revoked verification vector and the entrusted party's public tracking vector. S7. The tracking party compares the probability difference of the black-box output of the encrypted data to locate and remove malicious users.

[0005] Optionally, the publicly disclosed tracking vector matrix specifically includes: The elements of the publicly disclosed tracking vector matrix belong to a finite field modulo a prime number; The number of rows in the publicly tracked vector matrix is ​​equal to the number of registered users, with each row corresponding to one registered user; The columns of the public tracking vector matrix correspond to the public tracking vector components, and the number of columns is greater than the number of rows and has full row rank; After forming the public tracking vector matrix, the row rank is checked. If the row rank is insufficient, the public tracking vector for the corresponding registered user is regenerated, the row of the registered user in the public tracking vector matrix is ​​updated, and the function decryption key bound to the updated public tracking vector is re-derived.

[0006] Optionally, the data owner constructs a revocation verification vector based on the public tracking vector matrix and the revocation set, specifically including: The constraint vector corresponding to the revocation verification vector is indexed by the registered user. The constraint components of registered users within the cancellation set are set to zero, and the constraint components of registered users who have not cancelled their registrations are set to preset non-zero field elements. The publicly tracked vector matrix is ​​the first The dot product of the row and the undo verification vector is equal to the first constraint vector. One component; Gaussian elimination is performed on the linear constraint equations formed by the above inner product relationship in the finite field, and the free variables are set to zero to obtain the undo verification vector.

[0007] Optionally, the re-encryption key specifically includes: The authorization request includes the delegator index, the delegate index, the delegate public key, and the attribute access policy; The re-encryption key is simultaneously bound to the delegator index, the delegate index, the delegate public key, and the attribute access policy; The homomorphic signature private key generates a line signature for the re-encryption key line by line.

[0008] Optionally, the generation of the re-encrypted ciphertext specifically includes: Based on the attribute access policy, attribute evaluation is performed on the attribute vector and attribute ciphertext components of the original ciphertext to obtain the policy ciphertext components; Bit decomposition is performed after concatenating the first ciphertext component of the original ciphertext with the strategy ciphertext component. The bit decomposition result is concatenated with the message ciphertext component carrying the vector message in the original ciphertext to form a matrix input that matches the dimension of the re-encryption key; The matrix input is multiplied by the re-encryption key to obtain the re-encrypted ciphertext.

[0009] Optionally, the consistency verification specifically includes: The attribute evaluation, concatenation, bit decomposition, and matrix switching jointly determine the re-encryption calculation function. The input of the re-encryption calculation function includes the original ciphertext, the attribute access strategy, and the re-encryption key, and the output is the re-encrypted ciphertext. The transformed signature is obtained by performing homomorphic evaluation on the row signatures corresponding to each row of the re-encryption key according to the re-encryption calculation function; The consistency of the transformed signature, the re-encrypted ciphertext, and the output of the re-encryption calculation function is verified using the verification public key. In the event of a dispute, the certification authority independently performs verification based on the original ciphertext, the attribute access policy, the re-encrypted ciphertext, the line signature, and the verification public key, and outputs the verification result.

[0010] Optionally, the inner product of the vector message and the publicly disclosed tracking vector of the delegated party specifically includes: The inner product function is performed on the re-encrypted ciphertext using the decryption key described in the function, and only the scalar inner product of the vector message and the publicly disclosed tracking vector of the entrusted party is output, without outputting the complete content of the vector message; The function decryption key is issued only to the entrusted party and is bound to the entrusted party's user master key and public tracking vector.

[0011] Optionally, the function decryption key performs an inner product function evaluation on the re-encrypted ciphertext, specifically including: The scalar inner product of the decryption key output of the function is used as the message inner product in a finite field modulo a prime number; The inner product of the revocation verification vector and the publicly disclosed tracking vector of the entrusted party is taken as the verification inner product; When the verification inner product is non-zero, the message inner product is multiplied by the multiplicative inverse of the verification inner product to obtain the plaintext; When the verification inner product is zero, plaintext is rejected.

[0012] Optionally, the black-box output probability difference specifically includes: The revocation set is denoted as The number of users who have not withdrawn their registration is recorded as ,in Choose two different finite field scalar messages and Output the black box to be tracked As a fixed output event; The two scalar messages are used to form different vector test messages, and the probability between the black box to be tracked outputting the fixed output event in response to the corresponding challenge ciphertext has a difference of not less than a preset discrimination threshold. Users whose registration has not been revoked are sorted by index. Construct a candidate user set as well as ; make Let be the zero vector. for The revocation verification vector is doubled; right ,make Satisfy: the candidate user set With the revocation set Public tracking vectors of registered users in China The inner product is zero, and the public tracking vectors of the remaining users who have not revoked their registration are... The inner product equals The product of the registered user's public tracking vector and the doubling verification vector.

[0013] Optionally, the black-box tracing of the vector test message specifically includes: Define the vector test message as ,in The revocation verification vector; For each The vector test message is independently and randomly encrypted to generate multiple challenge ciphertexts; The multiple challenge ciphertexts are submitted to the black box to be tracked, and the probability estimate of the black box outputting the fixed output event is calculated. The upper bound of the statistical error is determined based on the number of queries and the confidence level. ; The probability estimate is determined by the number of times the fixed output event is output in independent repeated queries and the number of queries. Each candidate user set obtains the probability estimate using the same number of queries and confidence level. right Compare the probability estimates corresponding to adjacent candidate user sets, and calculate the statistical lower bound of the probability estimate difference after deducting the upper bound of the statistical error. ; When the statistical lower bound is greater than the preset tracking threshold, because Registered users Determined to be a malicious user and added to the subsequent revocation set, among which, This represents the set difference operation, i.e., from set... Remove from set Users.

[0014] The technical solution provided by this invention has the following beneficial effects: 1. By forming a finite field public tracking vector matrix by rows of the public tracking vectors of registered users, the row rank of the matrix is ​​checked. If the row rank is insufficient, the public tracking vector of the corresponding registered user is regenerated and the function decryption key bound to it is re-derived. Then, the constraint vector is set according to the revocation set and the revocation verification vector is solved so that the verification inner product of revoked users is zero and the verification inner product of non-revoked users is a non-zero field element, thereby realizing user-level revocation control and plaintext recovery control.

[0015] 2. By binding the re-encryption key to the delegator index, the delegate index, the delegate public key, and the attribute access policy, and generating a row signature for the re-encryption key line by line, and then determining the re-encryption calculation function by attribute evaluation, concatenation, bit decomposition, and matrix switching, the row signature is homomorphically evaluated and consistency verification is performed, so that the delegate and the certification authority can independently verify the ciphertext conversion result generated by the cloud server.

[0016] 3. By using lattice product proxy re-encryption parameters to complete ciphertext sharing, and performing independent random encryption and repeated black-box queries on vector test messages, the probability estimation difference between adjacent candidate user sets is compared based on the number of queries, confidence level, and upper bound of statistical error. Registered users who meet the tracking threshold are identified as malicious users and added to the subsequent revocation set. This gives data sharing a quantum-resistant security foundation based on lattice cryptography mechanism and enables statistical tracking and subsequent revocation of leaked users. Attached Figure Description

[0017] Figure 1 A schematic diagram illustrating the steps of a traceable proxy re-encrypted data sharing method for satellite networks provided by the present invention; Figure 2 The present invention provides an overall flowchart of a traceable proxy re-encrypted data sharing method for satellite networks; Figure 3 A schematic diagram illustrating the construction of the publicly disclosed tracking vector matrix, row rank detection, and revocation verification vector generation provided by this invention; Figure 4 The graph showing the relationship between the number of attributes and the runtime of re-encryption key generation is provided for this invention. Figure 5 A graph showing the relationship between the number of attributes and the overall processing runtime provided by this invention; Figure 6 The graph showing the relationship between the number of users and the black-box tracking runtime is provided for this invention. Detailed Implementation

[0018] 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 a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] This implementation method can be applied to cross-node and cross-domain data sharing scenarios in low-Earth orbit satellite constellations and integrated space-air-ground networks. In this application scenario, satellite nodes, ground stations, and cloud servers transmit service data through inter-satellite links, satellite-to-ground links, or relay links composed of satellites and ground stations. The data to be shared may include remote sensing data, satellite status data, link status data, and cross-domain service data. Different receiving users can configure corresponding access conditions based on mission attributes, regional attributes, node types, time windows, or access levels.

[0020] Specifically, when a data owner generates satellite service data to be shared, the data can be transmitted to the cloud server via a satellite relay node after being encrypted as plaintext and having corresponding attribute information and access policies set. The entrusting party submits an authorization request to the key generation center based on the data sharing needs. The key generation center generates a corresponding re-encryption key based on the entrustment relationship and the target entrusted party. The cloud server performs attribute evaluation on the original ciphertext. If it confirms that the attribute vector corresponding to the original ciphertext satisfies the access policies and the entrusted party has not been revoked, it performs ciphertext conversion and generates re-encryption ciphertext. After verifying the conversion result, the entrusted party recovers the plaintext data it is entitled to. During the above process, the satellite relay node only forwards the ciphertext, key materials, and verification materials; it does not obtain the satellite service plaintext, user master key, trapdoor, or function decryption key.

[0021] In situations where the number of users changes dynamically, user authorization status is updated, or there is key leakage or abnormal access, the data owner can reconstruct the revocation verification vector based on the current revocation set, making the verification inner product corresponding to the revoked user zero, thereby restricting their continued recovery of valid plaintext. When an illegal decryptor or suspected decryption capability leakage is discovered, the tracer can further identify the corresponding registered user through black-box tracking and add the identified malicious user to the subsequent revocation set. Therefore, this implementation method can adapt to data sharing scenarios in satellite networks with dynamic link changes, cross-domain user access, and dynamic updates to user authorization status.

[0022] For example, in a low-Earth orbit (LEO) satellite remote sensing data sharing scenario, after a LEO remote sensing satellite completes data acquisition of a target area, it can send the remote sensing data to a ground service processing node via an inter-satellite link or a satellite-to-ground link. The ground service processing node, acting as the data owner, encapsulates the remote sensing data to be shared and forms corresponding attribute vectors and attribute access policies based on information such as the remote sensing mission number, target area, data acquisition time, receiving node type, and access level. A key generation center can be deployed in a trusted ground management node to generate user keys, public tracking vectors, and corresponding function decryption keys for each registered user, and maintain the public tracking vector matrix and its version information. The data owner constructs a revocation verification vector based on the current public tracking vector matrix and the revocation set, encodes the plaintext corresponding to the remote sensing data into a vector message, generates the original ciphertext, and then sends the original ciphertext to a ground cloud platform or satellite edge cloud node via a satellite relay node.

[0023] When a registered user needs to obtain the remote sensing data, the entrusting party can submit an authorization request based on the data identifier, target entrusted party, and access policy of this remote sensing mission. The key generation center generates a re-encryption key bound to the entrusting party, the entrusted party, and the attribute access policy based on the authorization request, and generates a line signature for each line of the re-encryption key. The cloud server performs attribute evaluation on the original ciphertext. After confirming that the attribute vector corresponding to the original ciphertext satisfies the attribute access policy and that the entrusted party is not in a revoked state, it performs bit decomposition and matrix switching to generate re-encrypted ciphertext for the entrusted party. The entrusted party uses the verification public key to verify the consistency of the ciphertext conversion result, and after successful verification, uses the function to decrypt the key to obtain the corresponding message inner product, and then recovers the plaintext remote sensing data based on the verification inner product corresponding to the revoked verification vector.

[0024] During task execution, if a registered user's authorization is revoked, that user can be added to the revocation set during subsequent data sharing. A new revocation verification vector is then constructed, making the verification inner product for that user zero, thus preventing them from obtaining valid plaintext based on the plaintext recovery relationship. If it is further discovered that an external black-box decryptor can abnormally recover protected remote sensing data, the tracking party can construct vector test messages for different candidate user sets, independently and randomly encrypt and repeatedly query the corresponding challenge ciphertext using black-box methods, determine the corresponding registered user based on the output probability estimation difference between adjacent candidate user sets, and add the identified malicious user to the subsequent revocation set.

[0025] Please see Figure 1 and Figure 2 This embodiment of a traceable proxy re-encrypted data sharing method for satellite networks includes: S1. The key generation center establishes the lattice product proxy re-encryption parameters, generates the lattice trapdoor, user master key, public trace vector matrix, homomorphic signature private key and verification public key, and decrypts the key according to the user public trace vector derivation function. Specifically, during the system initialization phase, the key generation center sets security parameters, the maximum number of registered users, the number of rows in the lattice matrix, the number of columns in the lattice matrix, the number of attributes, and the lattice modulus. Plaintext Modulus The plaintext encoding order and error distribution are determined, and parameter identifiers are configured for the parameter set generated during this initialization. The lattice modulus is used for lattice matrix and ciphertext operations, the plaintext modulus is used for finite field inner products and plaintext recovery, the number of attributes is used to determine the size of the attribute matrix, and the plaintext encoding order is determined by how well the plaintext encoded values ​​can be embedded into the modulus. The conditions for the ciphertext space are set; the error term is generated according to a discrete Gaussian distribution, and its distribution parameters are determined based on security parameters, lattice modulus, and the noise limit allowed for correct decoding, so that the cumulative error generated by normal encryption and re-encryption is within the allowable range for message decoding.

[0026] For each registered user, the key generation center establishes key materials corresponding to the user index, generates a user matrix, a trapdoor, and an auxiliary matrix, and uses the trapdoor to perform short matrix sampling to obtain the user short matrix. The user public key consists of the user matrix and the auxiliary matrix, while the user master key includes the user short matrix and secret parameters bound to the user index. The user index serves as the identity association identifier for the key materials, used to distinguish the data authorization, ciphertext conversion, and plaintext recovery processes participated in by different registered users.

[0027] Each registered user is also configured with a public tracking vector. A one-to-one correspondence is established between the public tracking vector and the user index, and the vector is registered in the public directory. The public tracking vector is used for user revocation of constraints, function decryption key derivation, and black-box tracking, but does not form a complete plaintext decryption capability on its own.

[0028] The function decryption key is derived from the user's master key and the corresponding public tracking vector, and is bound to the user index and the current public tracking vector. The function decryption key is used to evaluate the inner product function on the ciphertext after proxy re-encryption, outputting the scalar inner product associated with the corresponding public tracking vector, but not the complete vector message. After the public tracking vector is updated, the function decryption key corresponding to the original public tracking vector is no longer used, and a new function decryption key is derived according to the updated public tracking vector.

[0029] The key generation center also generates homomorphic signature private keys and verification public keys. The homomorphic signature private key is used to form row signatures for each row of data in the subsequently generated re-encryption key, while the verification public key is used to verify the transformed signature formed by homomorphic evaluation of the row signatures. A correspondence is established between signature materials and parameter identifiers, authorization identifiers, and re-encryption key row indices to distinguish signature materials in different data sharing tasks.

[0030] The public directory records the user index, user public key identifier, public tracking vector, parameter identifier, and public tracking vector matrix version number. The matrix version number is updated synchronously when the public tracking vector is updated. Parameters, public tracking vectors, and related key materials in the same data sharing task use consistent parameter identifiers and matrix version numbers; materials with inconsistent versions are not included in the current data sharing process.

[0031] Optional, please refer to Figure 3 The publicly disclosed tracking vector matrix specifically includes: The elements of the publicly disclosed tracking vector matrix belong to a finite field modulo a prime number; The number of rows in the publicly tracked vector matrix is ​​equal to the number of registered users, with each row corresponding to one registered user; The columns of the public tracking vector matrix correspond to the public tracking vector components, and the number of columns is greater than the number of rows and has full row rank; After forming the public tracking vector matrix, the row rank is checked. If the row rank is insufficient, the public tracking vector for the corresponding registered user is regenerated, the row of the registered user in the public tracking vector matrix is ​​updated, and the function decryption key bound to the updated public tracking vector is re-derived.

[0032] Specifically, using prime numbers Establish a finite field for the model The total number of registered users is recorded as follows: The dimension of the publicly tracked vector is denoted as and set For registered users Generate public tracking vectors ,in, .

[0033] Arrange the public tracking vectors as rows in a matrix according to the registered user index to form a public tracking vector matrix:

[0034] The first in the matrix The fixed row corresponds to the first Register users and ensure that the matrix row index is consistent with the registered user index.

[0035] After the public tracking vector matrix is ​​generated, in a finite field Calculate the row rank of a matrix. When: When the current publicly available tracking vector matrix is ​​accepted, and the matrix row rank is less than 1, the current publicly available tracking vector matrix is ​​accepted. At that time, the public tracking vector of the corresponding registered user is regenerated, the corresponding row in the matrix is ​​replaced with the newly generated public tracking vector, and the row rank is checked again until the matrix meets the full row rank condition.

[0036] After the public tracking vector is regenerated, the function decryption key bound to the original public tracking vector is no longer used. The function decryption key is then regenerated using the corresponding registered user's master key and the updated public tracking vector, and the public tracking vector and matrix version number in the public catalog are updated simultaneously. After the update is complete, the public tracking vector and function decryption key corresponding to the current version for the same registered user index will remain unchanged; materials from older versions will not participate in the inner product function evaluation of the current version.

[0037] The public tracking vector matrix has more columns than rows and maintains full row rank, ensuring that the matrix rows corresponding to each registered user are linearly independent. This allows for the establishment of finite-field constraints corresponding to user indices based on the user's withdrawal status during subsequent processing. When a registered user's withdrawal status changes, the corresponding constraint components can be adjusted without regenerating the unchanged public tracking vector for that user.

[0038] S2. The data owner constructs a revocation verification vector based on the publicly available tracking vector matrix and the revocation set. The plaintext scalar is multiplied component by component into the revocation verification vector to form a vector message, which is then encrypted back into the original ciphertext. Specifically, the data owner obtains the plaintext of the satellite service to be shared, the attribute vector corresponding to this data sharing task, and the current revocation set. Satellite service plaintext can be remote sensing data, satellite status data, link status data, or cross-domain service data; attribute vectors can represent mission attributes, regional attributes, node types, time windows, or access levels. The data owner reads the currently valid public tracking vector matrix based on the parameter identifier and the version number of the public tracking vector matrix, and forms constraint information corresponding to the registered user index based on the revocation status of each registered user in the revocation set.

[0039] After canceling the generation of the verification vector, the plaintext to be shared is converted into a plaintext scalar in a finite field according to the plaintext encoding rules. The plaintext scalar is then multiplied by each component of the revocation verification vector to form a vector message. :

[0040] in, To cancel the verification vector, The plaintext modulus. For the _th ...__ of the revocation verification vector. Each component The corresponding vector message components are:

[0041] Thus, a product relationship is established between the plaintext scalar and the revocation verification vector, so that the message inner product obtained subsequently based on the user's public tracking vector simultaneously includes the plaintext scalar and the verification inner product corresponding to the user's revocation status.

[0042] The data owner encrypts the vector message according to the lattice encryption parameters, generating a message ciphertext component carrying the vector message; simultaneously, an attribute ciphertext component is generated based on the attribute vector, and a first ciphertext component used for ciphertext conversion is also generated. The message ciphertext component, attribute ciphertext component, and first ciphertext component are combined in a preset component order to form the original ciphertext.

[0043] After the validation vector is generated, the data owner will revoke the validation vector. As an auxiliary parameter bound to the original ciphertext, the parameter identifier, data identifier, and revocation set version number are associated with the revocation verification vector. The parameter identifier is used to determine the lattice parameters and finite field parameters used in this ciphertext processing; the data identifier is used to associate the authorization request and ciphertext conversion process corresponding to the same business data; and the revocation set version number is used to identify the user revocation status used when generating the current revocation verification vector. The revocation verification vector does not participate in lattice encryption as a ciphertext component. The original ciphertext and its associated auxiliary parameters maintain consistency between the data identifier and the revocation set version number during subsequent forwarding; if any associated information is missing or the version is inconsistent, the corresponding ciphertext conversion will not continue.

[0044] When the revocation set contains all registered users, there are no valid recipients who have not been revoked. The data owner ends the generation of the original ciphertext and does not generate any original ciphertext for the current revocation set.

[0045] Optionally, the data owner constructs a revocation verification vector based on the public tracking vector matrix and the revocation set, specifically including: The constraint vector corresponding to the revocation verification vector is indexed by the registered user. The constraint components of registered users within the cancellation set are set to zero, and the constraint components of registered users who have not cancelled their registrations are set to preset non-zero field elements. The publicly tracked vector matrix is ​​the first The dot product of the row and the undo verification vector is equal to the first constraint vector. One component; Gaussian elimination is performed on the linear constraint equations formed by the above inner product relationship in the finite field, and the free variables are set to zero to obtain the undo verification vector.

[0046] Specifically, let the total number of registered users be... The current cancellation set is Construct a constraint vector based on the registered user index:

[0047] No. The constraint components corresponding to registered users are set according to their cancellation status as follows:

[0048] in, and Users whose registrations are cancelled correspond to zero-constraint components, while users whose registrations are not cancelled correspond to non-zero-constraint components.

[0049] Let the public tracking vector matrix be:

[0050] The verification vector to be revoked is Then, establish linear constraints in the finite field:

[0051] For the The above matrix relationship corresponds to the registered users as follows:

[0052] Therefore, when hour, ;when hour, .

[0053] Therefore, the revocation status of each registered user is converted into a finite field inner product relationship between the public tracking vector and the revocation verification vector, with revoked users corresponding to zero verification inner product and non-revoked users corresponding to non-zero verification inner product.

[0054] In a finite field Chinese system of linear equations Perform Gaussian elimination. During the elimination process, row operations are performed according to the rules of addition, multiplication, and multiplicative inverse in finite fields. When there are free variables in the system of equations, all free variables are set to zero. Then, the remaining variables are determined based on the elimination results to obtain the cancellation verification vector corresponding to the current cancellation set. .

[0055] Before solving for the revocation verification vector, verify that the version number of the public tracing vector matrix matches the version number of the matrix in the current public catalog. If the versions match, perform a finite field solution; if the versions do not match, stop the current solution, reread the currently valid public tracing vector matrix and version information, and reconstruct the constraint vector and revocation verification vector according to the current revocation set.

[0056] Because the vector message satisfies For the first For registered users, the inner product between their public tracking vector and vector message satisfies:

[0057] When the registered user has not been deregistered For non-zero finite field elements, the corresponding verification inner product has a multiplicative inverse, which can participate in plaintext recovery when access policies, authorization relationships, and ciphertext verification conditions are subsequently met; when a registered user is revoked... The corresponding verification inner product is zero, which does not meet the condition for recovering the plaintext using the multiplicative inverse.

[0058] S3. The key generation center generates a re-encryption key according to the authorization request, the principal user's master key, the trapdoor, and the entrusted party's public key, and generates a line signature for the re-encryption key line by line using the homomorphic signature private key. Specifically, the delegation policy generates authorization requests for shared data. These requests include the delegator's index, the delegate's index, the delegate's public key, attribute access policy, data identifier, parameter identifier, and authorization validity period. Upon receiving the authorization request, the key generation center first verifies the parameter identifier and authorization validity period, and then reads the key material corresponding to this authorization based on the delegator's index and the delegate's index. If the parameter identifiers are inconsistent or the authorization validity period has expired, the generation of the re-encryption key is terminated.

[0059] The key generation center performs policy evaluation on the attribute matrix according to the attribute access policy, obtaining a combined attribute matrix corresponding to the current access conditions. Then, the delegating user matrix and the combined attribute matrix are concatenated according to a predetermined column order to form a combined matrix used for pre-image sampling. The combined attribute matrix retains the matrix structure corresponding to the attribute access policy, ensuring that the generated re-encryption key is only applicable to the access policy specified in the current authorization request.

[0060] The key generation center uses the user's master key and trapdoor corresponding to the delegator's index to perform left basis sampling and preimage sampling on the combined matrix, obtaining a short preimage matrix that satisfies the short matrix norm constraint. The target of the preimage sampling is determined by the auxiliary matrix in the delegator's user public key, so that the short preimage matrix matches the input components in the delegator's ciphertext space.

[0061] Let the number of rows in the lattice matrix be... The number of columns in the lattice matrix is The dimension of the publicly tracked vector is The user matrix corresponding to the entrusted party is denoted as... The auxiliary matrix is ​​denoted as ,satisfy:

[0062] The short preimage matrix formed by performing left basis sampling and preimage sampling on the combined matrix obtained according to the attribute access strategy is denoted as . ,in Indicates the principal's index. This indicates the attribute access strategy. The short preimage matrix, after being extended to binary bits, is denoted as... ,in, Indicates in the model The following binary exponentiation expansion operation is performed according to the preset expansion bit depth and bit order, and the dimension of the expanded result is... The expansion rules are used in conjunction with the subsequent bit decomposition operations.

[0063] From the discrete Gaussian error distribution Error matrices are selected respectively:

[0064]

[0065] According to the user matrix of the entrusted party Auxiliary matrix The binary bit extension results of the error matrix and the short preimage matrix are used to construct the system by the client. To the entrusted party Re-encryption key:

[0066] in, for The zero matrix, for The identity matrix; the number of rows in the two upper matrix blocks is 1. The number of columns is Therefore, the overall re-encryption key satisfies:

[0067] The bit decomposition process uses and The same positional order and expansion rules ensure that the subsequent matrix input is the same as... The row dimensions are matched; the upper left matrix block is used to introduce the entrusted user matrix into the ciphertext transformation, the upper right matrix block simultaneously introduces the transformation relationship corresponding to the entrusted auxiliary matrix, error term and short preimage matrix, the lower left zero matrix block isolates the ciphertext coordinates that do not participate in the transformation, and the lower right identity matrix block retains the ciphertext coordinates of the message that needs to be directly transmitted.

[0068] After the re-encryption key is generated, each row is numbered sequentially according to the matrix row order, and a row signature is generated for each row of data using the homomorphic signature private key. Each row signature is associated with at least the authorization identifier, the delegator index, the delegate index, the row index, and the binding digest of the corresponding row data, ensuring that subsequent homomorphic signature evaluations correspond to the matrix operations in the re-encryption calculation according to the determined row order. The key generation center sends the re-encryption key and the corresponding row signatures as the conversion materials for this authorization task to the cloud server.

[0069] Optionally, the re-encryption key specifically includes: The authorization request includes the delegator index, the delegate index, the delegate public key, and the attribute access policy; The re-encryption key is simultaneously bound to the delegator index, the delegate index, the delegate public key, and the attribute access policy; The homomorphic signature private key generates a line signature for the re-encryption key line by line.

[0070] Specifically, the delegator index in the authorization request is used to identify the authorization initiator and its corresponding user master key and trapdoor; the delegate index is used to identify the target registered user receiving the conversion result; the delegate public key is used to establish a matrix mapping for the target user's ciphertext space; and the attribute access policy is used to determine the attribute conditions that allow the ciphertext conversion to be performed in this authorization.

[0071] When generating the re-encryption key, the delegator index, delegate index, delegate public key identifier, and attribute access policy identifier are recorded as binding fields along with the re-encryption key. After generation, each binding field is checked against the corresponding field in the authorization request; if any field is inconsistent, the current re-encryption key is not output. With this binding method, re-encryption keys generated by the same delegator for different delegates or different access policies have different authorization associations and cannot be directly replaced for other authorization tasks.

[0072] The re-encryption key is divided into multiple rows of data units according to the matrix rows, and row signatures are generated for each row according to a fixed row number. For the first row... Line-by-line encryption key data The signature is recorded as follows:

[0073] in, For homomorphic signature private keys, For authorization identifier, For the client's index, For the agent's index, For row index, For the first The key data is re-encrypted.

[0074] Each row of signatures is arranged in the row order of the re-encryption key and associated with the corresponding row data. When switching matrices later, the re-encryption key rows involved in the calculation can be mapped to their respective row signatures in the same row order. When row data is deleted, replaced, rearranged, or mixed with row data from other authorized tasks, the row index and authorization binding information become inconsistent, and the subsequent converted signature cannot form a consistent verification relationship with the current re-encryption calculation result.

[0075] When the re-encryption key and row signature are sent to the cloud server via the satellite relay node, the original row order is maintained. During the forwarding process, the link identifier and forwarding time can be carried to record the transmission path of the key materials, but the link identifier and forwarding time do not participate in the matrix operation of the re-encryption key, nor do they change the authorization binding relationship corresponding to the row signature.

[0076] S4. The cloud server performs attribute evaluation on the original ciphertext. When the attribute vector corresponding to the original ciphertext satisfies the attribute access policy and the delegate has not revoked, the re-encryption key is used to perform matrix switching to obtain the re-encrypted ciphertext. Specifically, the cloud server obtains the original ciphertext, re-encryption key, line signature, and attribute access policy corresponding to the current data identifier and parameter identifier, and verifies the parameter versions used by the original ciphertext, re-encryption key, and attribute access policy. If the parameter identifier or version information is inconsistent, the current ciphertext conversion is stopped, and ciphertext materials generated by different parameter versions are not mixed.

[0077] During ciphertext conversion, the cloud server reads the attribute vector information and attribute ciphertext components from the original ciphertext, performs attribute evaluation according to the attribute access policy, and checks whether the delegated party belongs to the current revocation set. If the attribute vector corresponding to the original ciphertext does not satisfy the attribute access policy or the delegated party belongs to the current revocation set, a conversion failure flag is output, and no re-encrypted ciphertext is generated for that delegated party; if the attribute vector corresponding to the original ciphertext satisfies the attribute access policy and the delegated party has not been revoked, the matrix switching process begins.

[0078] Attribute evaluation is performed in the ciphertext domain without recovering the plaintext content of the attribute vector. A policy ciphertext component corresponding to the current attribute access policy is generated based on the computational relationships between the attribute ciphertext components. Therefore, when the access policy participates in subsequent ciphertext transformations, the cloud server does not need to obtain the data owner's business plaintext or attribute vector plaintext.

[0079] After the attribute evaluation is completed, the cloud server organizes the first ciphertext component, policy ciphertext component, and message ciphertext component in the original ciphertext according to the predetermined ciphertext component order, and performs bit decomposition and matrix input construction based on the matrix structure corresponding to the re-encryption key. The constructed matrix input is then used to perform matrix operations with the re-encryption key corresponding to the current authorized task, converting the ciphertext data in the original ciphertext space into re-encrypted ciphertext oriented towards the delegate.

[0080] After the re-encrypted ciphertext is generated, the cloud server associates the re-encrypted ciphertext with the original ciphertext identifier, policy identifier, re-encryption key identifier, and corresponding line signature, and sends it to the entrusted party for subsequent consistency verification. The cloud server records the parameter version used in this conversion and the conversion result digest, which is used to determine the original ciphertext and re-encryption key corresponding to the re-encrypted ciphertext in the event of a verification dispute.

[0081] Optionally, the generation of the re-encrypted ciphertext specifically includes: Based on the attribute access policy, attribute evaluation is performed on the attribute vector and attribute ciphertext components of the original ciphertext to obtain the policy ciphertext components; Bit decomposition is performed after concatenating the first ciphertext component of the original ciphertext with the strategy ciphertext component. The bit decomposition result is concatenated with the message ciphertext component carrying the vector message in the original ciphertext to form a matrix input that matches the dimension of the re-encryption key; The matrix input is multiplied by the re-encryption key to obtain the re-encrypted ciphertext.

[0082] Specifically, the cloud server reads the attribute ciphertext components corresponding to the attribute vectors from the original ciphertext and determines the attribute positions and calculation order for evaluation under the current policy based on the attribute access policy. The attribute ciphertext components are evaluated within the ciphertext domain according to this calculation order, forming the policy ciphertext components. The policy ciphertext components represent the ciphertext calculation result of the original ciphertext under the current attribute access policy and do not contain the plaintext representation of the attribute vectors.

[0083] After obtaining the policy ciphertext component, the first ciphertext component of the original ciphertext is concatenated with the policy ciphertext component in a fixed component order, and bit decomposition is performed on the concatenation result. The number of bits, bit order, and component arrangement used in the bit decomposition are consistent with the binary bit expansion rules used when generating the re-encryption key, so that the bit decomposition result can be matched item by item with the corresponding matrix block in the re-encryption key.

[0084] After completing the bit decomposition, the bit decomposition result is concatenated with the message ciphertext components carrying the vector message in the original ciphertext in a predetermined order to obtain the matrix input. Each position in the matrix input corresponds to the first ciphertext component, the component obtained by bit decomposition of the strategy ciphertext component, and the message ciphertext component, respectively. The number of columns in the matrix input is consistent with the operational dimension corresponding to the re-encryption key.

[0085] The cloud server will input the matrix. Perform matrix multiplication with the re-encryption key to obtain the re-encrypted ciphertext. The binary bit extension in the re-encryption key corresponds to the bit decomposition result. The zero matrix block does not introduce the corresponding ciphertext component, and the identity matrix block retains the ciphertext coordinates that need to be preserved. This allows the policy calculation result in the original ciphertext and the vector message ciphertext to be mapped into the ciphertext space corresponding to the delegate according to the predetermined matrix mapping.

[0086] After the matrix switch is complete, the cloud server checks the data identifier, parameter identifier, policy identifier, and re-encryption key identifier corresponding to the generated result. If the identifiers match, the re-encrypted ciphertext is output; if any identifier does not match the current authorized task, the intermediate matrix input and conversion result generated this time are discarded, and the re-encrypted ciphertext is not sent to the delegate. For repeated conversion requests with the same data identifier and authorization identifier, if the original ciphertext digest, re-encryption key identifier, and policy version are all consistent, the already generated re-encrypted ciphertext can be used; if any item changes, the policy status check and ciphertext conversion are re-executed.

[0087] S5. The entrusted party performs homomorphic evaluation on the line signature based on the original ciphertext and the attribute access policy to obtain the transformed signature, and uses the verification public key to verify the consistency between the transformed signature and the re-encrypted ciphertext. Specifically, the entrusted party receives the re-encrypted ciphertext, original ciphertext identifier, policy identifier, re-encryption key identifier, and corresponding line signature sent by the cloud server, and determines the ciphertext component arrangement rule, bit decomposition rule, and matrix operation rule to be used for this conversion based on the parameter identifiers. If the parameter identifier, policy identifier, or re-encryption key identifier is inconsistent with the current authorized task, subsequent verification will not be performed.

[0088] Attribute evaluation, ciphertext component concatenation, bit decomposition, and matrix switching are combined into a re-encryption computation function according to a defined operational order. Let the original ciphertext be... The attribute access strategy is The re-encryption key is The re-encryption calculation function is expressed as:

[0089] in, This is the re-encrypted ciphertext that should be obtained according to this computational relationship. (Function) The calculation order is consistent with the order of attribute evaluation, concatenation, bit decomposition, and matrix switching used by the cloud server when generating re-encrypted ciphertext.

[0090] The entrusted party determines the function input order based on the arrangement of the attribute ciphertext components, the first ciphertext component, and the message ciphertext component in the original ciphertext, and determines the attribute evaluation relationship according to the attribute access strategy. After completing the attribute evaluation, the concatenation and bit decomposition relationship is determined according to the same bit order rules as in the re-encryption key generation stage, and then the participation position of each row signature in the matrix switching calculation is determined according to the matrix row index.

[0091] The homomorphic signature process includes line signature generation, homomorphic evaluation of the signature, and verification of the transformation result. For the re-encryption key... row data Generate the corresponding line signature using the homomorphic signature private key. The entrusted party calculates the encryption function accordingly. The operational relationships between the rows of the matrix are used to perform homomorphic signature evaluation on the corresponding row signature sets to obtain the transformed signature. Then, the public key was verified, and the encryption function was re-encrypted. Re-encrypting ciphertext and signature conversion As verification input, when the transformed signature can be verified and According to If the calculated output is consistent component by component, the output verification passes; otherwise, the output verification fails.

[0092] The delegated party uses the verification public key to check the verification relationship between the transformed signature and the current re-encryption calculation function. During verification, the number of re-encryption key rows and row signatures, row index order, authorization identifier, delegator index, delegated party index, and parameter identifiers are checked, and it is also checked whether the transformed signature can be verified by the verification public key. If a row signature is missing, row indexes are discontinuous, authorization fields are inconsistent, or the transformed signature verification fails, the current re-encrypted ciphertext is deemed invalid and will not proceed to the subsequent function decryption process.

[0093] Optionally, the consistency verification specifically includes: The attribute evaluation, concatenation, bit decomposition, and matrix switching jointly determine the re-encryption calculation function. The input of the re-encryption calculation function includes the original ciphertext, the attribute access strategy, and the re-encryption key, and the output is the re-encrypted ciphertext. The transformed signature is obtained by performing homomorphic evaluation on the row signatures corresponding to each row of the re-encryption key according to the re-encryption calculation function; The consistency of the transformed signature, the re-encrypted ciphertext, and the output of the re-encryption calculation function is verified using the verification public key. In the event of a dispute, the certification authority independently performs verification based on the original ciphertext, the attribute access policy, the re-encrypted ciphertext, the line signature, and the verification public key, and outputs the verification result.

[0094] Specifically, re-encryption computation function The complete computational order from the original ciphertext to the re-encrypted ciphertext is preserved, including policy evaluation of the attribute ciphertext components, concatenation of the first ciphertext component and the policy ciphertext component, bit decomposition of the concatenation result, re-concatenation of the bit decomposition result and the message ciphertext component, and matrix switching. The input and output positions of each operation step are fixed, so that each row of the re-encryption key and its row signature can correspond to a defined matrix computation position.

[0095] The entrusted party, according to the function The computational relationships between each row of signatures in homomorphic evaluation are determined, and the row signatures are combined according to the computational order of their corresponding matrix rows to generate a transformed signature. Each row signature corresponds one-to-one with a row of the re-encryption key; therefore, deleting or replacing a row, or changing the order of the rows involved in the computation, will change the computational relationship corresponding to the transformed signature, rendering it unable to continue corresponding to the current re-encrypted ciphertext.

[0096] When performing consistency verification using the verification public key, at least the following should be checked: the number of row signatures is consistent with the number of rows of the corresponding re-encryption key; the row indices are consecutive and consistent with the matrix operation order; the authorization identifier, delegator index, delegate index, and parameter identifier bound to each row signature are consistent; the converted signature satisfies the signature verification relationship corresponding to the verification public key; and each component of the re-encrypted ciphertext is consistent with the corresponding output of the re-encryption calculation function.

[0097] The result of the re-encryption calculation function calculated based on the current input is denoted as:

[0098] When the verification relationship is true and At this point, the conversion result passes the consistency verification; in this process, the two re-encrypted ciphertexts are compared item by item according to their corresponding components. If any corresponding component is inconsistent, the current conversion result is determined to be invalid, and subsequent inner product function evaluation and plaintext recovery are not performed.

[0099] In the event of a dispute regarding the conversion result, the certification authority receives the original ciphertext, attribute access policy, re-encrypted ciphertext, row signature, and verification public key, and re-determines the input order and computational relationships of the re-encryption calculation function based on the same parameter identifiers. The certification authority independently performs checks on the number of row signatures, row index, authorization fields, conversion signature verification, and consistency checks on the re-encrypted ciphertext output, and outputs either verification passed or verification failed based on the results of each check.

[0100] The certification authority records the data identifier, authorization identifier, parameter identifier, number of line signatures, line index check result, and verification result corresponding to this verification, so that the verification record can be mapped to the specific original ciphertext, authorized task, and re-encrypted ciphertext; when the verification fails, the current re-encrypted ciphertext is not used as a valid conversion result for decryption.

[0101] S6. After successful verification, the entrusted party uses the function decryption key to obtain the inner product of the vector message and the entrusted party's public tracking vector from the re-encrypted ciphertext, and recovers the plaintext based on the inner product of the revoked verification vector and the entrusted party's public tracking vector. Specifically, after the S5 consistency verification is successful, the entrusted party reads the function decryption key that matches the current parameter identifier, user index, and public trace vector version, and obtains the revocation verification vector bound to the original ciphertext based on the data identifier of the original ciphertext and the revocation set version number. The entrusted party verifies the data identifier, parameter identifier, and revocation set version number corresponding to the revocation verification vector. If any of these pieces of information is inconsistent with the current original ciphertext, the current decryption process is stopped, and plaintext recovery is not performed on the re-encrypted ciphertext.

[0102] The entrusted party uses the function decryption key to perform an inner product function on the re-encrypted ciphertext, obtaining a scalar result corresponding to the current public tracking vector. Subsequently, it calculates the finite-field inner product between the current public tracking vector and the revocation verification vector, and determines whether plaintext recovery should be performed based on whether this inner product is a non-zero field element. The recovered finite-field plaintext scalar is then reverse-mapped according to the plaintext encoding rules used in the data encryption phase to obtain the corresponding satellite service plaintext.

[0103] Optionally, the inner product of the vector message and the publicly disclosed tracking vector of the delegated party specifically includes: The inner product function is performed on the re-encrypted ciphertext using the decryption key described in the function, and only the scalar inner product of the vector message and the publicly disclosed tracking vector of the entrusted party is output, without outputting the complete content of the vector message; The function decryption key is issued only to the entrusted party and is bound to the entrusted party's user master key and public tracking vector.

[0104] Specifically, let the publicly available tracking vector corresponding to the entrusted party be... The function decryption key is denoted as When generating the decryption key, the user index is used. User master key and public tracking vector of the current version Establish a binding relationship and distribute the keys only to the corresponding delegates. After the public tracing vector is updated, the original function decryption key ceases to be used, and a new function decryption key is derived based on the updated public tracing vector.

[0105] After the entrusted party inputs the re-encrypted ciphertext and the function decryption key, the function decryption process forms the decoding input based on the current lattice modulus, plaintext modulus, and plaintext encoding order, and then performs message decoding. When the cumulative error during encryption and re-encryption does not exceed the correct decoding noise limit set in the initialization phase, the decoded message vector is obtained. Vector messages formed during the data encryption phase In a finite field The function maintains consistency with the target language; when the accumulated error exceeds the correct decoding noise limit, it outputs a decoding failure flag and stops plaintext recovery. Upon successful decoding, the function only calculates the public tracking vector for the decryption key. and scalar inner product No output Given the vector components, then:

[0106] Function evaluation only outputs scalars No output The vector components. When the decoding result satisfies the current parameter conditions. Since the vector message is generated during the corresponding data encryption stage, subsequent plaintext recovery only requires the inner product of the messages, and does not require the entrusted party to obtain the complete vector message.

[0107] When the function decryption key is called, the user index, parameter identifier, and public trace vector version are checked simultaneously. If any of these do not match the information corresponding to the current re-encrypted ciphertext, a valid message inner product is not output; if all match, the message inner product is output. As input for subsequent finite field plaintext recovery.

[0108] Optionally, the function decryption key performs an inner product function evaluation on the re-encrypted ciphertext, specifically including: The scalar inner product of the decryption key output of the function is used as the message inner product in a finite field modulo a prime number; The inner product of the revocation verification vector and the publicly disclosed tracking vector of the entrusted party is taken as the verification inner product; When the verification inner product is non-zero, the message inner product is multiplied by the multiplicative inverse of the verification inner product to obtain the plaintext; When the verification inner product is zero, plaintext is rejected.

[0109] Specifically, message accumulation In a finite field The process is as follows: The delegate uses a revocation verification vector that is bound to the current original ciphertext and is consistent with the version of the revocation set. Calculate the public tracking vector The inner product of validation with the undo validation vector:

[0110] The vector message generated during the data encryption phase satisfies:

[0111] in, Let be a plaintext scalar. Therefore, when the function decrypts to obtain the message inner product corresponding to the vector message, we have:

[0112] when hour, In a finite field There exists a multiplicative inverse. The commissioned party calculates:

[0113] Obtain plaintext scalar Then, the plaintext of the business data is restored according to the encoding method corresponding to the data encryption stage.

[0114] when If there is no corresponding multiplicative inverse in the finite field, the plaintext recovery operation is not performed, and a decryption rejection result is output. According to the revocation constraint relationship established in S2, the verification inner product corresponding to the registered user in the revocation set is zero, and the verification inner product corresponding to the registered user who has not revoked is a non-zero field element. Therefore, the revocation status of the registered user is transformed into a finite field determination relationship for whether the plaintext recovery operation condition is met.

[0115] Before plaintext recovery, the version of the revocation verification vector is checked against the version number of the revocation set of the original ciphertext record. If the versions match, plaintext recovery is performed using the current verification inner product; if the versions do not match, decryption is stopped to avoid mixing revocation verification vectors and message inner products corresponding to different revocation states in the same recovery operation.

[0116] S7. The tracking party compares the probability difference of the black-box output of the encrypted data to locate and remove malicious users.

[0117] Specifically, the tracker initiates black-box tracking upon discovering key leakage, unauthorized decryptors, or abnormal access, and does not repeat the tracking during normal data sharing. The tracking target is the black box capable of receiving challenge ciphertext and outputting decryption results. The tracker determines the unregistered users participating in the tracking based on the current revocation set, and constructs test data using two different finite-field scalar messages. By changing the vector inner product relationship between different candidate users, the tracker aims to make the black box produce statistically significant changes in output probability under adjacent test conditions.

[0118] During the tracking process, each challenge ciphertext is independently and randomly generated according to the currently valid lattice encryption parameters. Multiple independent queries are performed under the same test conditions, and the number of times the black box to be tracked outputs a specified scalar message is counted. Based on the output probability estimates corresponding to different candidate user sets, the probability changes between adjacent candidate user sets are compared. When the probability change still exceeds the tracking threshold after deducting statistical errors, the registered user who is the only one whose probability has changed between adjacent candidate user sets is identified as the tracking result and added to the subsequent cancellation set.

[0119] Optionally, the black-box output probability difference specifically includes: The revocation set is denoted as The number of users who have not withdrawn their registration is recorded as ,in Choose two different finite field scalar messages and Output the black box to be tracked As a fixed output event; The two scalar messages are used to form different vector test messages, and the probability between the black box to be tracked outputting the fixed output event in response to the corresponding challenge ciphertext has a difference of not less than a preset discrimination threshold. Users whose registration has not been revoked are sorted by index. Construct a candidate user set as well as ; make Let be the zero vector. for The revocation verification vector is doubled; right ,make Satisfy: the candidate user set With the revocation set Public tracking vectors of registered users in China The inner product is zero, and the public tracking vectors of the remaining users who have not revoked their registration are... The inner product equals The product of the registered user's public tracking vector and the doubling verification vector.

[0120] Specifically, the current set of cancellations is denoted as Exclude users who have resigned from the currently registered users, and then transfer the remaining users to the next list. Registered users are arranged by user index as follows: .

[0121] The tracking party selects two that satisfy... Scalar messages with finite fields, and fixed black-box output. This is recorded as an event to be counted. The two scalar messages use the same plaintext finite-field parameters as normal data sharing, enabling the test message to participate in challenge ciphertext processing according to normal encryption and decryption rules.

[0122] For the ranked users who have not withdrawn their registrations, construct a progressively smaller set of candidate users:

[0123]

[0124] Until .

[0125] Only one registered user is reduced between adjacent candidate user sets, and the following conditions are met:

[0126] Construct corresponding vectors for each candidate user set Boundary conditions are set to ,as well as ,in, This is the cancellation verification vector corresponding to the current cancellation set.

[0127] for In a finite field China Linear constraints. For those belonging to Registered users ,set up .

[0128] For those that do not belong to Unregistered users ,set up .

[0129] Based on the row indices of the publicly traced vector matrix, the above inner product conditions are used to form a system of linear equations in a finite field, and the solutions are obtained using the same finite field elimination rules as those for solving the undo verification vector. When there are free variables, set the free variables to zero to obtain the definite vector corresponding to the current candidate user set.

[0130] After applying the above inner product condition, the non-revoked users inside the candidate user set correspond to zero additional inner product, and the non-revoked users outside the candidate user set correspond to... Double the inner product for validation. This difference is used to ensure that adjacent candidate user sets are only validated when... The corresponding test relationships have changed.

[0131] Optionally, the black-box tracing of the vector test message specifically includes: Define the vector test message as ,in The revocation verification vector; For each The vector test message is independently and randomly encrypted to generate multiple challenge ciphertexts; The multiple challenge ciphertexts are submitted to the black box to be tracked, and the probability estimate of the black box outputting the fixed output event is calculated. The upper bound of the statistical error is determined based on the number of queries and the confidence level. ; The probability estimate is determined by the number of times the fixed output event is output in independent repeated queries and the number of queries. Each candidate user set obtains the probability estimate using the same number of queries and confidence level. right Compare the probability estimates corresponding to adjacent candidate user sets, and calculate the statistical lower bound of the probability estimate difference after deducting the upper bound of the statistical error. ; When the statistical lower bound is greater than the preset tracking threshold, because Registered users Determined to be a malicious user and added to the subsequent revocation set, among which, This represents the set difference operation, i.e., from set... Remove from set Users.

[0132] Specifically, for each Construct vector test messages .

[0133] For those belonging to the candidate user set Unregistered users ,because ,therefore .

[0134] When verifying that the inner product is nonzero, the test message is obtained by recovering the relation from the plaintext. .

[0135] For those who do not belong to the candidate user set And registered users who have not been revoked ,have ,therefore, .

[0136] Following the same plaintext recovery relation, the corresponding test results are: Therefore, when the candidate user set is composed of Change to At that time, only registered users The test plaintext changes from inside the set to outside the set. Change to The change in black-box output probability between adjacent test conditions can correspond to the registered user. .

[0137] For each vector test message Multiple independent random encryption operations are performed using the same encryption parameters, ensuring that each query uses an independent encryption random value and generates a corresponding challenge ciphertext. Let the number of queries corresponding to each candidate user set be... The black box outputs fixed events to be tracked. The number of times The probability estimate is denoted as .

[0138] Each candidate user set uses the same number of queries and confidence level, and the corresponding upper bound of statistical error is calculated according to the same statistical error determination rule. This avoids directly comparing probability estimates obtained under different query sizes or confidence conditions.

[0139] For adjacent and Calculate the statistical lower bound of the probability estimate difference:

[0140] The preset tracking threshold is denoted as .when At that time, according to The registered users corresponding to the probability change are determined as follows Index users Write to the tracking record and add it to the subsequent undo set.

[0141] When all If none of the tracking thresholds are exceeded, the current tracking process outputs no identified results, and no new or withdrawn users are added. When tracking needs to be executed again, the fixed output events, candidate user order, and confidence conditions remain unchanged, but independent query samples are added, and the probability estimates and statistical error upper bounds are recalculated.

[0142] For identified malicious users, the tracking logs must at least store the user index, the determination time, the corresponding probability estimate, the upper bound of the statistical error, and the threshold comparison results. The updated revocation set is used to reconstruct the revocation verification vector in the subsequent data sharing process, ensuring that the verification inner product corresponding to newly revoked users becomes zero in the subsequent plaintext recovery process.

[0143] To further illustrate the technical effects of this invention, tests were conducted on the runtime of the original ciphertext generation, re-encryption key generation, proxy re-encryption, verification decoding, comprehensive processing, and black-box tracing processes. The results are as follows: Figures 4 to 6 As shown.

[0144] Please see Figure 4 The horizontal axis represents the number of attributes, and the vertical axis represents the re-encryption key generation runtime. The legend includes comparison schemes 1, 2, 3, and 4, as well as the proposed method. As the number of attributes increases, the re-encryption key generation time of the proposed method gradually increases, reflecting that the computational load required for attribute access policy evaluation, short preimage matrix generation, matrix expansion, and row-by-row signature increases with the attribute size. The runtime curve changes continuously, indicating that the re-encryption key can be generated according to the same authorization process under different attribute sizes.

[0145] Please see Figure 5 The horizontal axis represents the number of attributes, and the vertical axis represents the overall processing runtime. The legend includes comparison scheme 1, comparison scheme 2, comparison scheme 3, comparison scheme 4, and the method proposed in this paper. As the number of attributes increases from low to high, the overall processing runtime curve of the method proposed in this paper remains low and the change is relatively small, while some comparison schemes show a more obvious upward trend with the increase of the number of attributes. This result reflects that when the original ciphertext processing, authorization conversion, and verification decoding are executed in combination, the overall runtime of the method proposed in this paper responds relatively smoothly to the increase in the number of attributes.

[0146] Please see Figure 6The horizontal axis represents the number of users, and the vertical axis represents the black-box tracking runtime. As the number of registered users increases from 5 to 50, the black-box tracking runtime of the proposed method increases from approximately 1650 milliseconds to approximately 3600 milliseconds, showing an overall approximately linear upward trend. With the increase in the number of users, the number of candidate user sets, vector test messages, and corresponding challenge ciphertexts that need to be constructed also increases, thus increasing the processing time for repeated black-box queries and probability statistics. This relationship indicates that the runtime of the black-box tracking process grows in an orderly manner with the scale of registered users, enabling the comparison of probability differences step-by-step according to the candidate user set to locate the corresponding registered user.

[0147] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for sharing traceable, re-encrypted data via a traceable agent in satellite networks, characterized in that: include: S1. The key generation center establishes the lattice product proxy re-encryption parameters, generates the lattice trapdoor, user master key, public trace vector matrix, homomorphic signature private key and verification public key, and decrypts the key according to the user public trace vector derivation function. S2. The data owner constructs a revocation verification vector based on the public tracking vector matrix and the revocation set, multiplies the plaintext scalar component by component into the revocation verification vector to form a vector message and encrypts it into the original ciphertext. S3. The key generation center generates a re-encryption key according to the authorization request, the principal user's master key, the trapdoor, and the entrusted party's public key, and generates a line signature for the re-encryption key line by line using the homomorphic signature private key. S4. The cloud server performs attribute evaluation on the original ciphertext. When the attribute vector corresponding to the original ciphertext satisfies the attribute access policy and the delegate has not revoked, the re-encryption key is used to perform matrix switching to obtain the re-encrypted ciphertext. S5. The entrusted party performs homomorphic evaluation on the line signature based on the original ciphertext and the attribute access policy to obtain the transformed signature, and uses the verification public key to verify the consistency between the transformed signature and the re-encrypted ciphertext. S6. After successful verification, the entrusted party uses the function decryption key to obtain the inner product of the vector message and the entrusted party's public tracking vector from the re-encrypted ciphertext, and recovers the plaintext based on the inner product of the revoked verification vector and the entrusted party's public tracking vector. S7. The tracking party compares the probability difference of the black-box output of the encrypted data to locate and remove malicious users.

2. The method according to claim 1, characterized in that, The publicly disclosed tracking vector matrix specifically includes: The elements of the publicly disclosed tracking vector matrix belong to a finite field modulo a prime number; The number of rows in the publicly tracked vector matrix is ​​equal to the number of registered users, with each row corresponding to one registered user; The columns of the public tracking vector matrix correspond to the public tracking vector components, and the number of columns is greater than the number of rows and has full row rank; After forming the public tracking vector matrix, the row rank is checked. If the row rank is insufficient, the public tracking vector for the corresponding registered user is regenerated, the row of the registered user in the public tracking vector matrix is ​​updated, and the function decryption key bound to the updated public tracking vector is re-derived.

3. The method according to claim 2, characterized in that, The data owner constructs a revocation verification vector based on the public tracking vector matrix and the revocation set, specifically including: The constraint vector corresponding to the revocation verification vector is indexed by the registered user. The constraint components of registered users within the cancellation set are set to zero, and the constraint components of registered users who have not cancelled their registrations are set to preset non-zero field elements. The publicly tracked vector matrix is ​​the first The dot product of the row and the undo verification vector is equal to the first constraint vector. One component; Gaussian elimination is performed on the linear constraint equations formed by the above inner product relationship in the finite field, and the free variables are set to zero to obtain the undo verification vector.

4. The method according to claim 1, characterized in that, The re-encryption key specifically includes: The authorization request includes the delegator index, the delegate index, the delegate public key, and the attribute access policy; The re-encryption key is simultaneously bound to the delegator index, the delegate index, the delegate public key, and the attribute access policy; The homomorphic signature private key generates a line signature for the re-encryption key line by line.

5. The method according to claim 1, characterized in that, The generation of the re-encrypted ciphertext specifically includes: Based on the attribute access policy, attribute evaluation is performed on the attribute vector and attribute ciphertext components of the original ciphertext to obtain the policy ciphertext components; Bit decomposition is performed after concatenating the first ciphertext component of the original ciphertext with the strategy ciphertext component. The bit decomposition result is concatenated with the message ciphertext component carrying the vector message in the original ciphertext to form a matrix input that matches the dimension of the re-encryption key; The matrix input is multiplied by the re-encryption key to obtain the re-encrypted ciphertext.

6. The method according to claim 1, characterized in that, The consistency verification specifically includes: The attribute evaluation, concatenation, bit decomposition, and matrix switching jointly determine the re-encryption calculation function. The input of the re-encryption calculation function includes the original ciphertext, the attribute access strategy, and the re-encryption key, and the output is the re-encrypted ciphertext. The transformed signature is obtained by performing homomorphic evaluation on the row signatures corresponding to each row of the re-encryption key according to the re-encryption calculation function; The consistency of the transformed signature, the re-encrypted ciphertext, and the output of the re-encryption calculation function is verified using the verification public key. In the event of a dispute, the certification authority independently performs verification based on the original ciphertext, the attribute access policy, the re-encrypted ciphertext, the line signature, and the verification public key, and outputs the verification result.

7. The method according to claim 1, characterized in that, The inner product of the vector message and the publicly disclosed tracking vector by the entrusted party specifically includes: The inner product function is performed on the re-encrypted ciphertext using the decryption key described in the function, and only the scalar inner product of the vector message and the publicly disclosed tracking vector of the entrusted party is output, without outputting the complete content of the vector message; The function decryption key is issued only to the entrusted party and is bound to the entrusted party's user master key and public tracking vector.

8. The method according to claim 7, characterized in that, The function decryption key performs an inner product function evaluation on the re-encrypted ciphertext, specifically including: The scalar inner product of the decryption key output of the function is used as the message inner product in a finite field modulo a prime number; The inner product of the revocation verification vector and the publicly disclosed tracking vector of the entrusted party is taken as the verification inner product; When the verification inner product is non-zero, the message inner product is multiplied by the multiplicative inverse of the verification inner product to obtain the plaintext; When the verification inner product is zero, plaintext is rejected.

9. The method according to claim 1, characterized in that, The black-box output probability difference specifically includes: The revocation set is denoted as The number of users who have not withdrawn their registration is recorded as ,in Choose two different finite field scalar messages and Output the black box to be tracked As a fixed output event; The two scalar messages are used to form different vector test messages, and the probability between the black box to be tracked outputting the fixed output event in response to the corresponding challenge ciphertext has a difference of not less than a preset discrimination threshold. Users whose registration has not been revoked are sorted by index. Construct a candidate user set as well as ; make Let be the zero vector. for The revocation verification vector is doubled; right ,make Satisfy: the candidate user set With the revocation set Public tracking vectors of registered users in China The inner product is zero, and the public tracking vectors of the remaining users who have not revoked their registration are... The inner product equals The product of the registered user's public tracking vector and the doubling verification vector.

10. The method according to claim 9, characterized in that, The black-box tracing of the vector test messages specifically includes: Define the vector test message as ,in The revocation verification vector; For each The vector test message is independently and randomly encrypted to generate multiple challenge ciphertexts; The multiple challenge ciphertexts are submitted to the black box to be tracked, and the probability estimate of the black box outputting the fixed output event is calculated. The upper bound of the statistical error is determined based on the number of queries and the confidence level. ; The probability estimate is determined by the number of times the fixed output event is output in independent repeated queries and the number of queries. Each candidate user set obtains the probability estimate using the same number of queries and confidence level. right Compare the probability estimates corresponding to adjacent candidate user sets, and calculate the statistical lower bound of the probability estimate difference after deducting the upper bound of the statistical error. ; When the statistical lower bound is greater than the preset tracking threshold, because Registered users Determined to be a malicious user and added to the subsequent revocation set, among which, This represents the set difference operation, i.e., from set... Remove from set Users.