Online document co-processing method, device, medium and system in confidential environment
By combining MAC, DAC, IDEA encryption and a hybrid audit chain architecture, the problem of collaborative efficiency and security in the collaborative processing of classified documents is solved, achieving efficient, secure collaborative document operation and traceability.
Patent Information
- Application Number
- CN202511933567.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-19
- Publication Date
- 2026-03-17
AI Technical Summary
Existing file collaboration technologies cannot achieve efficient multi-person collaboration while ensuring the security of classified information, resulting in problems in both collaboration efficiency and security.
A dual authorization mechanism combining mandatory access control (MAC) and discretionary access control (DAC) is adopted. The multi-round operation of the IDEA encryption algorithm is distributed in parallel across various sub-threads. A hybrid audit chain architecture using zero-knowledge proof (ZKP), hash chains, and Merkle trees is used to build an attack-resistant audit mechanism, achieving encrypted recording and traceability of document operations.
Without relying on external cloud services, it enables cross-terminal and cross-role document collaboration, ensuring the confidentiality, integrity, and traceability of classified documents, and improving encryption speed and security.
Smart Images
Figure CN121682893A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of computer application, more particularly, to an online document collaborative processing method, device, medium and system in a secret-related environment. BACKGROUND
[0002] In today's digital age, efficient circulation and collaboration of information are crucial to the operation and development of each organization, and massive data is stored in the form of electronic information in mobile carriers, flows in network space, and stays in office equipment. Meanwhile, in daily office work, the demand for cross-department and cross-level information collaboration is increasingly prominent, and collaborative editing of documents has become a normal state to support efficient business development. However, for military enterprises, government agencies, scientific research units and other secret-related units, the information involved in the document often relates to core secrets, and information leakage will cause irreparable damage to national security and enterprise interests. Therefore, as an important carrier of secret-related information, how to realize online multi-person collaboration of documents while ensuring the security of secret-related information has become a problem to be solved.
[0003] Existing file collaboration technology often cannot balance collaboration efficiency and security at the same time, either overemphasizing collaboration efficiency and weakening fine control of permissions, or causing a significant decline in collaboration experience due to strict security restrictions. SUMMARY
[0004] The purpose of the present application is to overcome the shortcomings of the prior art and provide an online document collaborative processing method, device, medium and system in a secret-related environment, which effectively ensures the confidentiality, integrity and traceability of secret-related documents during collaboration.
[0005] The purpose of the present application is achieved by the following scheme: An online document collaborative processing method in a secret-related environment, comprising the following steps: First, the text content of the document is preprocessed to construct a semantic feature vector of the text sensitive information, and the vector similarity is calculated, and the secret level of the secret-related document with the largest value in the calculation result is taken as the secret level of the to-be-tested document; Second, the assisted control combining mandatory access control method MAC and discretionary access control method DAC is adopted, the dual authorization mechanism of the document directory and the document object is constructed by the discretionary access control method DAC, and the two authorization modes of inheriting the preset permissions of the parent directory and customizing the permissions configuration are adopted; on the basis of discretionary control permissions, different security levels are set for personnel and documents and a secret level matrix is constructed, and whether to enable access restriction is determined according to the sensitive markers of personnel and documents; meanwhile, the multi-round operation of the IDEA encryption algorithm is distributed in parallel on each sub-thread, the main thread is responsible for key initialization and parameter management of the sub-thread, and the RSA asymmetric encryption algorithm is used to encrypt the key for protection; Finally, in a confidential environment, all document operations are recorded in the server local log in an encrypted form, and a hybrid audit chain architecture of zero-knowledge proof ZKP, hash chain and Merkle Tree is adopted to construct an anti-attack audit mechanism without exposing the operation plaintext, and the audit log is only allowed to be decrypted and reviewed by authorized administrators in a trusted environment.
[0006] Further, the calculation of the vector similarity is obtained by calculating the cosine distance between the vectors.
[0007] Further, the hybrid audit chain architecture of zero-knowledge proof ZKP, hash chain and Merkle Tree specifically includes the following sub-steps: the MHT module encapsulates all operation records as leaf nodes, builds a hash tree, and takes the root hash as a global fingerprint; the ZKP proof generator generates zero-knowledge proof that each operation record belongs to the current MHT; and the ZKP verifier receives the proof and the root hash to verify the authenticity of the record without obtaining the entire chain.
[0008] An online document collaborative processing device in a confidential environment includes a processor and a memory, and the memory stores a computer program.
[0009] A computer readable storage medium stores a computer program, and the computer program is loaded and executed by a processor to perform the method of any one of the above.
[0010] An online document collaborative processing system in a confidential environment includes the online document collaborative processing device in a confidential environment as described above.
[0011] The beneficial effects of the present application include: The present application combines the information semantic-based document classification method, the dual access control method based on MAC and DAC, and the multi-threaded IDEA-ECB file encryption to automatically classify and control the document, effectively improving the encryption speed and security of the document. On the premise of not relying on external cloud services, cross-terminal and cross-role document collaborative operations are realized, and a document operation audit chain is formed, thereby effectively ensuring the confidentiality, integrity and traceability of the confidential document in the collaborative process. BRIEF DESCRIPTION OF DRAWINGS
[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0013] Figure 1 This is a flowchart of a document security level detection activity according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a cooperative access control principle according to an embodiment of the present invention; Figure 3 This is a flowchart of an embodiment of the IDEA-ECB encryption technology based on multithreading. Figure 4 This is a flowchart of a hybrid audit chain consisting of a hash chain, a Merkle tree, and zero-knowledge proofs (ZKP) according to an embodiment of the present invention. Detailed Implementation
[0014] All features disclosed in all embodiments of this specification, or steps in all methods or processes implied in the disclosure, may be combined and / or extended or replaced in any way, except for mutually exclusive features and / or steps.
[0015] In a preferred embodiment, such as Figure 1 As shown, a method for collaborative online document processing in a classified environment is provided, including the following steps: First, by preprocessing the text content of the document, semantic feature vectors of the sensitive information in the text are constructed, and the vector similarity is obtained by calculating the cosine distance between the vectors. The classification level of the classified document with the largest value in the calculation result is taken as the classification level of the document to be tested.
[0016] Secondly, a collaborative control method combining Mandatory Access Control (MAC) and Discretionary Access Control (DAC) is adopted. A dual authorization mechanism for document directories and document objects is constructed using DAC, supporting both document inheritance of parent directory preset permissions and custom permission configurations. Based on DAC permissions, different security levels are set for personnel and documents, and a security level matrix is constructed. The system determines whether to enable access restrictions based on the sensitivity markers of personnel and documents, thereby preventing the leakage of high-level confidential information to low-level personnel. Simultaneously, considering the sensitivity of classified documents, the system proposes a multi-threaded IDEA-ECB symmetric encryption algorithm. Based on the principle of trading space for time, the multiple rounds of IDEA encryption algorithm operations are distributed in parallel across various sub-threads. The main thread is responsible for key initialization and parameter management of sub-threads. The improved algorithm's key is encrypted using the RSA asymmetric encryption algorithm. This avoids the potential risks of plaintext file storage while ensuring the speed and efficiency of the file encryption algorithm.
[0017] Finally, in the classified environment, all document operations (CRUD operations) are recorded in encrypted form in the local server logs. A hybrid audit chain architecture of Zero-Knowledge Proof (ZKP), hash chain, and Merkle Tree is adopted to achieve operation verifiability and meet the "minimum necessary disclosure" privacy protection principle without exposing the plaintext of the operations. This constructs an attack-resistant audit mechanism that can withstand mass tampering. It achieves a traceable and non-repudiable operation audit chain throughout the entire lifecycle. Audit logs can only be decrypted and viewed by authorized administrators in a trusted environment, meeting the compliance requirements of Level 3 or above of the Information Security Protection System.
[0018] The above embodiments propose an online document collaborative processing scheme in a classified environment. Based on information security, this scheme realizes document collaborative processing in a classified environment from four perspectives: access control, storage encryption, document labeling, and audit tracing.
[0019] More specifically, in other embodiments, an online document collaborative processing method in a classified environment is provided. Based on the methods of the above embodiments, a document classification level detection model based on sensitive information semantics is specifically constructed, consisting of a training corpus and a training corpus. The training corpus is constructed from documents to be classified, and the training corpus is constructed from documents already classified. Through data preprocessing, feature word sets of the two corpora are extracted; through text feature representation, sensitive information semantic feature vector sets of the two corpora are obtained; through vector similarity calculation, a reference classification level of the document to be classified is obtained, specifically including the following steps: Step 1: Data Preprocessing: The original text structure is converted into a feature word set composed of independent terms. First, the plain text content of documents in different formats is extracted, removing non-text information such as images, icons, and punctuation marks. Second, a professional vocabulary dictionary in the relevant field is introduced, and word segmentation tools are used to segment the Chinese text content into several independent terms. Then, prepositions, conjunctions, auxiliary words, and other connecting words in the document are removed, and a feature word set for the document is established. All words in the document and their corresponding vocabulary information are saved.
[0020] Step 2: Construct Semantic Feature Vectors for Sensitive Information: Extract classified sensitive features from the text content and convert them into feature vectors. Match the feature word set from step 1 with a classified sensitive word library in the professional field, and construct semantic feature vectors for the matched sensitive information. The vector representation is as follows: where, Document sensitive words Co-occurring word pairs, express In the document Feature weights in the model.
[0021]
[0022] Step 3: Feature Weight Calculation: The feature vector weights are calculated using TF-IDF word frequency-inverse document frequency, as shown in the following formula. Where, It is a characteristic word The number of times a word appears in document i, i.e., word frequency; It is the inverse document frequency, indicating the presence of feature words. Number of documents The reciprocal of the ratio to the total number of training documents N.
[0023]
[0024] Step 4: Vector Similarity Calculation: The cosine distance between two vectors can be obtained by calculating the cosine similarity between them. The larger the cosine distance value, the more similar the vector structures are, i.e., the higher the degree of similarity. The formula for calculating the cosine similarity of vectors is:
[0025] In the formula, and and These represent different vector sets. The closer the cosine value Sim(A,B) is to 1, the closer the angle between the two vectors is to 0, meaning the two vectors are more similar.
[0026] refer to Figure 2This embodiment further proposes a collaborative access control method based on the methods described in the previous embodiments. It first performs discretionary access control, and then, based on custom authorization, determines whether to grant relevant permissions according to a mandatory access control method. Specifically, discretionary access control employs dual operation-level authorization control for both the document directory and document objects, supporting both document inheritance of parent directory preset permissions and custom permission configurations. Mandatory access control uses a security level matrix for determination, where the personnel's confidentiality level is determined by... This indicates the confidentiality level of the document object. The classification levels based on the sensitivity of the subject and object are shown in Table 1 below: Table 1
[0027] Define the document object as a quintuple DOC=(L,C,R,W,D,S), where: L: The security classification level of the document object; C: Document text and attribute information; R: Read permission granted by discretionary access control for documents; W: Write permissions granted by discretionary access control for documents; D: Download permissions granted by discretionary document access control; S: The deletion permission provided by discretionary access control for documents; Document permissions must be granted to the subject. and object After undergoing dual verification of permissions through discretionary access control and mandatory access control, the principal... To obtain the final object The corresponding operation permissions.
[0028] refer to Figure 3 This embodiment, based on the methods described above, further proposes to integrate multi-threading technology with the IDEA-ECB algorithm, aiming to effectively improve IDEA-ECB by shortening encryption runtime, increasing encryption speed, and improving throughput. Since the IDEA algorithm can only use a single-threaded sequential stream for plaintext block encryption, and the encryption process involves multiple iterations and mixed operations, the encryption speed is often slow and cannot handle large-scale file transfers and encryption. Therefore, this embodiment proposes the following improved algorithm: setting main thread management parameters. The specified program contains different execution threads, each with its own method execution program counter and stack, enabling multiple child threads to execute independently and concurrently, achieving the goal of sharing resources within the program. The system manages and controls the child threads through the main thread, allowing them to simultaneously serve multiple plaintext packet modules in time and space, realizing concurrent operations between child threads and maximizing the utilization of CPU resources.
[0029] Initial input: n-bit plaintext blocks ; Output: n-bit ciphertext blocks .
[0030] The specific steps are as follows: The Java Random function is called to randomly generate plaintext, and the plaintext is divided into n groups of 64 bits each. IDEA's key length is 128 bits, utilizing Generate key That is, 52 subkeys; Implement the Runnable interface and set thread management parameters. Override the run() method; Start a thread to perform concurrent encryption operations. , It performs encryption operations with key k and controls various state transitions of the thread.
[0031] refer to Figure 4 This embodiment further proposes a hybrid audit chain model based on hash chains, Merkle trees, and zero-knowledge proofs (ZKP), building upon the methods described above. This model consists of four parts: a Merkle HashTree (MHT), a ZKP proof generator, a ZKP validator, and audit chain metadata. The MHT encapsulates all operation records as leaf nodes, constructing a hash tree with the root hash serving as the global fingerprint. The ZKP proof generator generates a zero-knowledge proof for each operation record stating that "this record belongs to the current MHT." The ZKP validator receives the proof and the root hash, quickly verifying the record's authenticity without needing to obtain the entire chain. The specific implementation steps are as follows: Step 1: Operation Event Capture: When a user performs an operation, the system automatically captures the operation context and generates a structured audit event object. The model de-identifies fields in the original event that may leak privacy, generates pseudo-identifiers, and uses a hash function with a fixed salt value (such as SHA-256) to ensure that the same entity is always mapped to the same pseudo-identifier. It supports cross-event correlation analysis, but cannot reverse-engineer the original value.
[0032] Step 2: Merkle Tree Construction: MHT arranges all anonymized events (including the current event) in chronological order, generates a hash value H(event) for each event, and uses it as a leaf node of the Merkle Tree, constructing a binary tree structure from bottom to top. Assume the dataset is... Collision-resistant one-way hash function → The hash chain length is k, and the Merkle tree depth is... ,but
[0033]
[0034] In the formula, = , express Iterate i times, where || represents byte concatenation.
[0035] Step 3: Generate Zero-Knowledge Proof (ZKP Proof): When a user or external auditor wants to verify whether a record (such as L2) exists in the current audit chain, the system extracts the path proof of L2 from the MHT, i.e., the hash path from L2 to the root's sibling node; it constructs a proof circuit using zk-SNARKs, whose inputs are: Root Hash (known), Leaf Hash (L2), Authentication Path (path hash), and whether the path constitutes a valid MHT construction. A short ZKP proof is then generated and returned to the verifier along with the Root Hash.
[0036] Step 4: The validator receives and verifies ZKP: The validator receives: Root Hash (public) and ZKP proof (from the system), without needing to know the entire chain or other records. The verification process is as follows: Call the ZKP verification function, passing in the Root Hash and ZKP proof; if the verification is successful, it means that the record does exist in the current audit chain; if it fails, it means that the record is forged or the chain has been tampered with. Zero-knowledge proof (ZKP) interaction round t, challenge space. ,but
[0037]
[0038] The formula outputs 1 when the proposition is true and 0 when it is false. Based on the Merkle path verification results, ZKP verification confirms the integrity of the hash chain through hash iteration relationships.
[0039] The methods described above address the challenge of achieving efficient, real-time, and auditable multi-user collaborative editing in classified networks using traditional collaborative office systems. They combine a document security level detection method based on semantic information, a dual access control method based on MAC+DAC, and multi-threaded IDEA-ECB file encryption to automatically classify and control access to documents, effectively improving encryption speed and security. This enables cross-terminal and cross-role document collaboration without relying on external cloud services and allows for the formation of a document operation audit chain, effectively ensuring the confidentiality, integrity, and traceability of classified documents during the collaboration process.
[0040] The above description is only a preferred embodiment of the present invention and is not intended to limit the present application. 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 application.
[0041] The units described in the embodiments of the present invention can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.
[0042] According to one aspect of the present invention, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the methods provided in the various optional implementations described above.
[0043] In another aspect, embodiments of the present invention also provide a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.
Claims
1. An online document collaborative processing method in a secret environment, characterized in that, The method comprises the following steps: First, by preprocessing the text content of the document, a semantic feature vector of the text sensitive information is constructed, and a vector similarity is calculated, and the maximum value in the calculation result is used as the classified level of the document to be tested. Second, the combination of mandatory access control method MAC and discretionary access control method DAC is used for assisted control, the double authorization mechanism of document directory and document object is constructed by the discretionary access control method DAC, and the two authorization modes of inheriting the preset permissions of the parent directory and customizing the permissions configuration are used for the document; on the basis of the discretionary control permissions, different security levels are set for the personnel and the document, and a classified level matrix is constructed, and whether to enable access restriction is determined according to the sensitive marking of the personnel and the document; meanwhile, the multi-round operation of the IDEA encryption algorithm is distributed on each sub-thread in parallel, the main thread is responsible for key initialization and parameter management of the sub-thread, and the RSA asymmetric encryption algorithm is used for encryption protection of the key. Finally, in the secret environment, all document operations are recorded in the local log of the server in an encrypted form, a hybrid audit chain architecture of zero-knowledge proof ZKP, hash chain and Merkle Tree is used to construct an anti-attack audit mechanism without exposing the operation plaintext, and the audit log is only allowed to be decrypted and reviewed by the authorized administrator in a trusted environment.
2. The method according to claim 1, wherein, The vector similarity is obtained by calculating the cosine distance between the vectors.
3. The method of claim 1, wherein, The hybrid audit chain architecture of zero-knowledge proof ZKP, hash chain and Merkle Tree comprises the following sub-steps: the MHT module encapsulates all operation records as leaf nodes, constructs a hash tree, and uses the root hash as a global fingerprint; the ZKP proof generator generates zero-knowledge proof that each operation record belongs to the current MHT; and the ZKP verifier receives the proof and the root hash, and verifies the authenticity of the record without obtaining the entire chain.
4. An online document collaborative processing device in a secret-related environment, characterized in that, The computer program is stored in the memory and loaded by the processor to execute the method of any one of claims 1-3.
5. A computer-readable storage medium, characterized in that, The computer program is stored in the readable storage medium and loaded by the processor to execute the method of any one of claims 1-3.
6. An online document collaborative processing system in a secret environment, characterized in that, The online document collaborative processing device in the secret environment comprises the device according to claim 4. The computer program is stored in the memory and loaded by the processor to execute the method of any one of claims 1-3.