PLM-based electronic image-text file remote multi-warehouse collaborative storage method

Abstract

The invention belongs to the technical field of product production cycle management, and particularly relates to an electronic image-text file remote multi-bin collaborative storage method based on a PLM. Comprising the steps that a 1 + N hybrid storage architecture is constructed, one center bin is deployed in an enterprise headquarter data center and stores full document metadata and latest version files, and N area bins are deployed according to geographic areas and store cache copies and incremental update data of local high-frequency access documents; the remote multi-warehouse document increment synchronization is realized through an intelligent synchronization engine, and the intelligent synchronization engine adopts a content awareness increment synchronization algorithm and a dynamic synchronization strategy based on network quality; version management is realized through a version cooperative control module, and the version cooperative control module is used for establishing a global version number mapping mechanism and a conflict resolution decision tree; cross-warehouse process automation is realized through a cross-warehouse workflow engine, and the cross-warehouse workflow engine integrates a BPM process engine and a message queue; the method can adapt to a remote network environment, and the large file synchronization efficiency is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of product production cycle management, and particularly relates to a method for collaborative storage of electronic graphic documents in multiple warehouses in different places based on PLM. BACKGROUND

[0002] As a core tool for enterprise product lifecycle management, the electronic graphic document storage is a key function of supporting R&D and production collaboration in the product lifecycle management (PLM) system. With the deepening of the globalization of enterprises, the demand for real-time collaboration and access to electronic graphic documents in different places has significantly increased in multi-site R&D centers and production bases, especially the cross-regional synchronization of large files such as GB-level engineering drawings and 3D models, which has become a core link to ensure R&D efficiency. In the prior art, the electronic graphic document storage in different places mainly adopts centralized or simple distributed architecture. The centralized storage needs to retrieve all files through a remote network, and when multiple users access concurrently, it is easily limited by network bandwidth, and the transmission efficiency is extremely low. Although the simple distributed storage disperses the storage pressure, it uses the full synchronization method, and when the document is updated, the complete file needs to be transmitted. In the fluctuating environment of remote networks, the synchronization delay of large files often takes several hours, which seriously affects the collaboration progress of cross-regional teams. Therefore, the prior art cannot meet the needs of enterprises for efficient synchronization of large files in different places.

[0003] Based on the above problems, there is an urgent need for an electronic graphic document collaborative storage technology scheme that can adapt to the remote network environment and improve the synchronization efficiency of large files. SUMMARY

[0004] The purpose of the present application is to solve the problems existing in the prior art, and to propose a method for collaborative storage of electronic graphic documents in multiple warehouses in different places based on PLM, comprising: S1: Construct a 1+N hybrid storage architecture, wherein 1 central warehouse is deployed in the data center of the enterprise headquarters and stores the full amount of document metadata and the latest version of the file, and N regional warehouses are deployed according to geographical regions and store the cache copies and incremental update data of the local high-frequency access documents; S2: Realize incremental synchronization of documents in multiple warehouses in different places through an intelligent synchronization engine, which adopts a content-aware incremental synchronization algorithm and a dynamic synchronization strategy based on network quality; S3: Realize version management through a version collaboration control module, which is used to establish a global version number mapping mechanism and a conflict resolution decision tree; S4: Realize cross-warehouse process automation through a cross-warehouse workflow engine, which integrates a BPM process engine and a message queue; realize data security through a multi-layer security protection system, which includes transmission layer encryption, storage layer encryption and access control.

[0005] Preferably, the content-aware incremental synchronization algorithm of the intelligent synchronization engine achieves incremental transmission through file block hash comparison. Specifically, it dynamically adjusts the block granularity according to document type, blocks text files according to a preset size, blocks 3D models according to assembly, and generates a unique hash value for each block. The change blocks for document updates are determined by hash value comparison, and only the change blocks are transmitted to reduce the amount of data transmitted. The dynamic synchronization strategy based on network quality includes real-time monitoring of network bandwidth in different locations. When the network bandwidth is lower than a preset threshold, it automatically switches to asynchronous compressed transmission mode to improve transmission stability.

[0006] More preferably, the global version number mapping mechanism of the version collaboration control module generates a unique global identifier for each document, and the global identifier includes a timestamp, a repository identifier, and a document type code; Before updating a document, a regional repository needs to send a version lock application to the central repository. The central repository reviews the application and provides feedback on the review results. Once the review is approved, the regional repository obtains a version lock for a preset duration. Within the validity period of the version lock, the regional repository completes the update and synchronizes the updated data with the central repository. When a version conflict is detected, the conflict resolution decision tree judges the conflict in the following order: central repository priority, modification time priority, and approval status priority. Based on the judgment results, the conflicting versions are automatically merged. If automatic merging cannot be completed, the conflict information is pushed to the administrator for manual arbitration.

[0007] Further optimized, the BPM process engine of the cross-warehouse workflow engine predefines cross-warehouse approval templates. The cross-warehouse approval templates include design change templates, version update templates, and permission adjustment templates. Each template clearly defines the process node and the corresponding processing warehouse for each node. When a cross-warehouse approval event is triggered, the BPM process engine generates an approval task according to the corresponding template and pushes the approval task to the terminal of the person in charge of the corresponding warehouse. Message queues are used to implement cross-warehouse event notifications. When a document is checked in, updated, or its approval status changes, the message queue collects event information in real time and pushes the event information to the subscriber terminals of the associated warehouse according to preset subscription rules.

[0008] More preferably, the intelligent synchronization engine further includes a synchronization priority calculation step, wherein the synchronization priority is calculated using the following formula: ; Where P is the synchronization priority, ranging from 0 to 10; Scurrent is the current instantaneous bandwidth; Srated is the rated bandwidth; δjitter is the network jitter coefficient, ranging from 0 to 1, with the larger the network jitter, the closer δjitter is to 1; Ht0 is the initial access popularity; t is the time since the last access; λ is the popularity decay coefficient, ranging from 0.01 to 0.1; Tapproval is the remaining response time of the approval node; α, β, and γ are weighting coefficients, and α + β + γ = 1.

[0009] More preferably, the intelligent synchronization engine further includes a block size calculation step, wherein the block size is determined based on the synchronization priority and calculated using the following formula: ; Where L is the actual block size; Lbase is the base block size; P is the synchronization priority; θ is the block adjustment coefficient, ranging from 0.3 to 0.7; μ is the complexity influence coefficient, ranging from 0.2; Ccomplex is the file complexity, ranging from 0 to 2, with values ​​of 0.1 to 0.5 for text files and 1.5 to 2 for 3D models.

[0010] Furthermore, in a more preferred embodiment, the transport layer encryption of the multi-layer security protection system employs dynamic key encryption, with the encryption key length determined based on the block size and calculated using the following formula: ; Where K is the actual encryption key length; K is the base key length; L is the actual block size; L is the base block size B; T_approval is the remaining response time of the approval node; T_approval is the base approval time; ξ is the impact coefficient of block size on encryption, with a value of 0.4; η is the impact coefficient of approval time on encryption, with a value of 0.6.

[0011] In a further preferred embodiment, the storage layer encryption of the multi-layer security protection system adopts a document block encryption storage method. After the document is divided into blocks, a unique encryption key is generated for each block. The encryption key is generated using the national cryptographic algorithm. Metadata and file entities are stored separately. Metadata is stored in the metadata database of the central repository, while file entities are stored in the entity storage area of ​​the corresponding repository. Metadata includes document identifiers, block information, encryption key indexes, and access permission information. Add digital watermarks to sensitive documents. The digital watermark contains user IP, operation timestamp, and operation type information. The digital watermark is embedded in the document through a steganography algorithm and remains visible when the document is copied, transmitted, or printed.

[0012] Furthermore, the access control of the multi-layered security protection system adopts a hybrid model of RBAC and ABAC. The RBAC model predefines roles and the basic permissions corresponding to each role. The basic permissions include read, modify, delete, and approve; The ABAC model determines dynamic permissions based on attributes, including the user's repository, document storage area, and document security level. When a user initiates an access request, the access control module first uses the RBAC model to determine whether the user has basic permissions, and then uses the ABAC model to determine whether the user's attributes meet the dynamic permission requirements. Only when both basic and dynamic permission requirements are met can the user be allowed to perform the corresponding access operation.

[0013] Furthermore, the 1+N hybrid storage architecture also includes a load balancing mechanism, which monitors the CPU utilization, memory utilization, and network bandwidth utilization of each regional warehouse in real time. When the load parameters of a regional repository exceed a preset threshold, the load balancing mechanism will migrate some of the frequently accessed document cache copies of the regional repository to an adjacent regional repository with a lower load. At the same time, the allocation of incremental synchronization tasks in the regional warehouses will be adjusted, and some synchronization tasks will be transferred to the central warehouse for execution. Once the load parameters of the regional warehouses drop below the threshold, their original cached copies and synchronization task allocations will be gradually restored.

[0014] The technical effects achieved by this invention compared to traditional technical solutions include: This invention addresses the primary problem of low synchronization efficiency for large files across different locations in the prior art through innovative technologies such as a 1+N hybrid storage architecture, an intelligent synchronization engine, and a version collaborative control module. Collaborative storage between the central and regional repositories reduces the pressure of cross-regional transmission; content-aware incremental synchronization transmits only change blocks; and combined with a dynamic network quality adjustment strategy, it significantly improves the efficiency of document synchronization across multiple warehouses, meeting the needs of cross-regional collaboration. Attached Figure Description

[0015] Figure 1 This is a flowchart of the PLM-based method for collaborative storage of electronic documents across multiple warehouses in different locations. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] Traditional technical solutions have the following technical problems: Traditional centralized storage has low transmission efficiency and strong network dependence when accessing large-scale documents in different locations and when multiple users operate concurrently; simple distributed storage makes it difficult to guarantee data consistency and has unclear cross-regional collaboration processes. Moreover, existing solutions have deficiencies in data synchronization, version control, cross-warehouse processes and security protection, and cannot meet the requirements of enterprises for efficient and reliable document management.

[0018] Based on this, please refer to Figure 1 This embodiment provides a PLM-based method for collaborative storage of electronic documents across multiple warehouses in different locations, including: constructing a 1+N hybrid storage architecture, where one central warehouse is deployed in the enterprise headquarters data center and stores all document metadata and the latest version files, and N regional warehouses are deployed according to geographical regions and store cached copies of frequently accessed local documents and incremental update data; achieving incremental synchronization of documents across multiple warehouses through an intelligent synchronization engine, which adopts a content-aware incremental synchronization algorithm and a dynamic synchronization strategy based on network quality; implementing version management through a version collaboration control module, which establishes a global version number mapping mechanism and a conflict resolution decision tree; achieving cross-warehouse process automation through a cross-warehouse workflow engine, which integrates a BPM process engine and a message queue; and achieving data security through a multi-layer security protection system, which includes transmission layer encryption, storage layer encryption, and access control.

[0019] This solution utilizes a "1+N" hybrid architecture combining four core modules to specifically address the shortcomings of traditional storage models: the collaborative architecture between the central and regional warehouses balances data integrity with local access efficiency; the intelligent synchronization engine overcomes the limitations of full-scale synchronization; the version collaboration control module resolves version conflicts across multiple warehouses; the cross-warehouse workflow engine eliminates process gaps; and a multi-layered security protection system enhances data security. This multi-module collaborative architecture design forms a complete solution covering storage, synchronization, versioning, processes, and security, achieving a systematic optimization of document management in a geographically dispersed, multi-warehouse environment compared to traditional centralized or simple distributed storage.

[0020] It's worth noting that in the 1+N hybrid storage architecture, the central repository serves as the global data benchmark to ensure data consistency, while regional repository stores frequently accessed documents to reduce the pressure of cross-regional transmission. The intelligent synchronization engine's incremental synchronization and dynamic strategies adapt to differences in cross-regional networks. The version collaboration control module reduces conflicts at the source through global identifiers and version locks. The cross-repository workflow engine's BPM and message queues enable process automation and real-time notifications. A multi-layered security protection system ensures security from multiple dimensions, including transmission, storage, and access. These components do not exist in isolation but collaborate through data interaction. For example, data transmitted by the synchronization engine must be encrypted by the security system, and version updates trigger the workflow engine's notification mechanism.

[0021] The technical effects achieved by the above solution include: addressing the inherent defects of centralized and simple distributed systems through a 1+N architecture; improving data access efficiency through the division of labor between the central warehouse and regional warehouses; reducing data transmission volume and adapting to different network environments through an intelligent synchronization engine; avoiding version conflicts and ensuring data consistency through a version collaboration control module; automating approval and other processes through a cross-warehouse workflow engine, reducing delays; and reducing the risk of data leakage through a multi-layered security protection system.

[0022] The modules work together to meet the high efficiency and reliability requirements of collaborative storage of electronic documents under the globalization of enterprises, and provide support for document management in the digital transformation of enterprises.

[0023] Traditional technical solutions have the following technical problems: Data synchronization efficiency is low in existing technologies. Traditional distributed storage adopts a full synchronization method, which requires the transmission of the entire file when updating a document. This is time-consuming in remote network environments, especially for large files where the synchronization delay can reach several hours. Furthermore, it lacks the ability to dynamically adapt to changes in network quality.

[0024] Based on this, the content-aware incremental synchronization algorithm of the intelligent synchronization engine achieves incremental transmission through file block hash comparison. Specifically, it dynamically adjusts the block granularity according to document type, blocks text files according to a preset size, blocks 3D models according to assembly, and generates a unique hash value for each block. The changed blocks of the document update are determined by hash value comparison, and only the changed blocks are transmitted to reduce the amount of data transmission. The dynamic synchronization strategy based on network quality includes real-time monitoring of remote network bandwidth. When the network bandwidth is lower than a preset threshold, it automatically switches to asynchronous compression transmission mode to improve transmission stability.

[0025] This solution addresses the inefficiency of full synchronization by combining block hash comparison with dynamic network adaptation: dynamic adjustment of block granularity adapts to the structural characteristics of different document types; for example, dividing 3D models into assembly blocks better suits their data organization and allows for precise location of changed parts; hash comparison ensures that only changed blocks are transmitted, avoiding redundancy from full transmission; real-time monitoring and mode switching of network bandwidth enable the synchronization process to adapt to the instability of remote networks. Compared to traditional full synchronization, this "block-comparison-dynamic adjustment" logic forms a more intelligent synchronization mechanism.

[0026] It's worth noting that the preset block size for text files can be set based on the average file size, ensuring a reasonable number of blocks. This avoids both excessively small blocks leading to increased hash calculation overhead and excessively large blocks affecting the accuracy of change block location. When dividing 3D models into assemblies, the hierarchical structure of the model can be considered, treating sub-assemblies as independent blocks. When a sub-assembly is updated, only the hash value and data of the corresponding block need to be transmitted. Network bandwidth monitoring is achieved by real-time collection of parameters such as transmission rate. The preset threshold can be set according to the enterprise network environment. The asynchronous compression transmission mode uses a high-efficiency compression algorithm to reduce data volume while ensuring compression and decompression efficiency and maintaining synchronization continuity.

[0027] The technical effects achieved by the above solution include: incremental transmission through block hash comparison, avoiding the large data transmission required for full synchronization and shortening the synchronization time for large files; dynamically adjusting the block granularity according to document type, improving the accuracy of changed block identification, and further reducing invalid transmission; and dynamic response to network bandwidth, ensuring the synchronization process remains stable under different network conditions and avoiding transmission interruptions or delays due to insufficient bandwidth. These effects collectively improve the efficiency and stability of data synchronization between geographically dispersed multi-warehouses, solving the time-consuming problem of traditional synchronization methods in geographically dispersed network environments, and meeting the enterprise's need for rapid synchronization of large files.

[0028] Traditional technical solutions have the following technical problems: version control is chaotic in existing technologies. When multiple warehouses are stored independently, teams in different regions may modify the same document at the same time, resulting in version conflicts. Furthermore, there is a lack of effective conflict resolution mechanisms. There have been cases where economic losses have been caused by version inconsistencies. At the same time, inconsistent version identification also increases the difficulty of version management.

[0029] Based on this, the global version number mapping mechanism of the version collaboration control module generates a unique global identifier for each document. The global identifier includes a timestamp, repository identifier, and document type code. Before updating a document, a regional repository must send a version lock application to the central repository. The central repository reviews the application and provides feedback on the review result. After the review is approved, the regional repository obtains a version lock for a preset duration. Within the validity period of the version lock, the regional repository completes the update and synchronizes the updated data with the central repository. When a conflict resolution decision tree detects a version conflict, it judges the conflict in the following order: central repository priority, modification time priority, and approval status priority. Based on the judgment result, the conflicting versions are automatically merged. When automatic merging cannot be completed, the conflict information is pushed to the administrator for manual arbitration.

[0030] This solution addresses the problem of chaotic version control by combining global identifiers, version locks, and a conflict resolution decision tree: global identifiers ensure the uniqueness and traceability of each document version, avoiding version confusion caused by inconsistent identifiers; the version lock mechanism prevents multiple repositories from simultaneously modifying the same document, reducing the likelihood of conflicts; and the conflict resolution decision tree provides clear conflict handling rules, enabling conflicts to be resolved systematically and avoiding the arbitrariness of manual handling. Compared to the lack of unified version management in existing technologies, this "prevention-handling" logic forms a complete version control system.

[0031] It is worth mentioning that: the timestamp in the global identifier can be accurate to the millisecond level, ensuring that different operations at the same point in time can be distinguished; the warehouse identifier corresponds to the geographical information of the regional warehouse, which facilitates tracing the source of the version; and the document type code corresponds to the enterprise document classification system, improving the standardization of version management.

[0032] The review of version lock applications includes verifying whether the current document has been locked and the operation permissions of the requested regional repository. The duration of the version lock after approval can be set according to the document type. For example, the version lock duration for complex 3D models can be appropriately extended.

[0033] The decision-making logic of the conflict resolution decision tree is pre-embedded in the system. The central warehouse priority principle ensures the authority of the global data benchmark, while the priority of modification time and approval status serve as supplements to adapt to the conflict handling needs in different scenarios. When manually arbitrating, the information pushed to the administrator includes the content of the conflicting version, the person who made the modification, the modification time, etc., to assist the administrator in decision-making.

[0034] The technical effects achieved by the above solution include: global identification gives each version a unique and identifiable identity, facilitating version traceability and management; the version locking mechanism effectively reduces conflicts caused by simultaneous modifications in multiple repositories, controlling version chaos from the source; and the conflict resolution decision tree enables conflicts to be handled automatically or in an orderly manner, avoiding problems such as misuse of drawings due to version conflicts.

[0035] These effects collectively enhance the orderliness and accuracy of version control, ensure consistency of document versions across multiple warehouses in different locations, and provide version-level assurance for collaborative work among cross-regional teams.

[0036] Traditional technical solutions have the following technical problems: existing solutions lack a unified workflow engine, and cross-warehouse collaborative processes such as document approval and change notification cannot be automatically processed, resulting in a break in the cross-warehouse collaborative process. For example, a certain airline company experienced a delay in the development cycle of a new aircraft due to cross-warehouse approval delays, which affected the company's R&D efficiency.

[0037] Based on this, the BPM process engine of the cross-warehouse workflow engine predefines cross-warehouse approval templates. These templates include design change, version update, and permission adjustment templates. Each template clearly defines the process node and the corresponding processing warehouse. When a cross-warehouse approval event is triggered, the BPM process engine generates an approval task according to the corresponding template and pushes the approval task to the terminal of the person in charge of the corresponding warehouse. The message queue is used to implement cross-warehouse event notifications. When a document is checked in, updated, or its approval status changes, the message queue collects event information in real time and pushes the event information to the terminals of subscribed users in the associated warehouse according to preset subscription rules.

[0038] This solution addresses the issue of discontinuity in cross-warehouse collaborative processes by combining a BPM process engine with a message queue: pre-defined approval templates clearly define the nodes and responsible parties for different types of processes, preventing process interruptions due to a lack of rules; the BPM engine automatically generates and pushes tasks when events are triggered, replacing manual transmission and reducing human delays; and the message queue ensures that event information is delivered to relevant users in real time and accurately, preventing information loss.

[0039] Compared to the lack of a unified workflow in existing technologies, this "template-push-notification" mechanism creates a closed loop for cross-warehouse processes from initiation to completion.

[0040] It's worth noting that the approval templates are designed to align with actual business processes. Modification templates can include nodes such as regional warehouse initial review, central warehouse verification, and expert review, clearly defining the processing time and output requirements for each node. Templates can be adjusted to meet specific business needs, supporting the addition or modification of nodes. When the BPM workflow engine generates approval tasks, it associates them with relevant document information and historical records, allowing responsible personnel to quickly understand the background. Task push notifications are delivered via in-system messages or mobile app notifications, ensuring timely receipt by responsible personnel.

[0041] The subscription rules for message queues can be set by users. For example, developers can subscribe to version update notifications for specific types of documents. The subscription rules are stored in the system, and the message queue pushes messages accurately according to the rules, avoiding interference from irrelevant information.

[0042] The technical effects achieved by the above solution include: preset approval templates provide a clear workflow for cross-warehouse approvals, avoiding process chaos; the BPM engine automatically generates and pushes approval tasks, reducing manual intervention and shortening approval time; and the message queue pushes event information in real time, ensuring that relevant users are aware of document status changes in a timely manner, facilitating collaborative work.

[0043] These effects collectively resolve the issue of discontinuity in cross-warehouse collaboration processes, enabling efficient operation of approval and notification processes, avoiding delays in R&D cycles, and improving the efficiency of cross-regional collaboration for enterprises.

[0044] Traditional technical solutions have the following technical problems: the existing synchronization mechanism lacks differentiation of synchronization task priorities, which may cause important documents or documents related to urgent approval to be delayed because resources are occupied by low-priority synchronization tasks, affecting the progress of critical business. In addition, the arrangement of synchronization tasks does not take into account actual factors such as network jitter and access popularity, which reduces the utilization efficiency of synchronization resources.

[0045] Based on this, the intelligent synchronization engine also includes a synchronization priority calculation step, which is calculated using the following formula: ; Where P is the synchronization priority, ranging from 0 to 10; Scurrent is the current instantaneous bandwidth, in Mbps; Srated is the rated bandwidth, in Mbps; δjitter is the network jitter coefficient, ranging from 0 to 1, with the larger the network jitter, the closer δjitter is to 1; Ht0 is the initial access popularity, in times per day; t is the time since the last access, in days; λ is the popularity decay coefficient, ranging from 0.01 to 0.1; Tapproval is the remaining response time of the approval node, in hours; α, β, and γ are weighting coefficients, and α + β + γ = 1.

[0046] This formula is used to quantify the priority of synchronization tasks. The higher the P value, the higher the priority of execution. By integrating three core factors, namely real-time network status, document access popularity, and approval urgency, it achieves dynamic sorting of synchronization tasks, solving the problem of unreasonable resource allocation caused by ambiguous task priorities in traditional synchronization.

[0047] The first part of the formula Focus on network adaptability: This represents the current instantaneous bandwidth, reflecting real-time transmission capability; This refers to the rated bandwidth, which is the basic bandwidth agreed upon between the enterprise and the operator. This is the network jitter coefficient, ranging from 0 to 1. The greater the jitter, the closer the value is to 1, used to quantify the degree of bandwidth fluctuation. This part calculates the actual available transmission capacity of the network by "real-time bandwidth / (rated bandwidth × jitter correction coefficient)". The more stable the network and the higher the real-time bandwidth, the higher the score of this part, ensuring that resources are allocated preferentially when network conditions permit.

[0048] For weighting coefficients, and The sum is 1, which can be adjusted according to the enterprise's emphasis on network stability; for example, it can be increased in cross-border transmission scenarios. .

[0049] Formula Part 2 Focusing on document access popularity: Initial access popularity, which is the frequency of access to the document in the most recent period; This is the time since the last visit; This is the heat decay coefficient (0.01-0.1), which controls the rate at which heat decays over time; the larger the decay rate value, the faster the decay. This part calculates the current effective heat using "initial heat / (initial heat + decay amount)". Documents that have been accessed frequently recently receive higher scores, avoiding the problem of "low-frequency documents occupying synchronization resources for a long time".

[0050] As a weighting coefficient, it can improve performance in R&D-intensive scenarios (such as the new product design stage). Priority will be given to ensuring the synchronization of frequently used engineering drawings.

[0051] Formula Part 3 Focus on the urgency of the approval process: This represents the remaining response time for the approval node; the shorter the time, the more urgent the approval. This section directly correlates the urgency level with "1 / remaining time," with documents related to urgent approvals receiving higher scores to avoid approval stalls due to synchronization delays. As a weighting factor, it can be improved during the mass production stage (where approval processes are intensive). Prioritizing approval-related documents. Overall, this formula breaks through the limitations of traditional single-factor priority determination. Through the weighted fusion of three types of factors, it achieves dynamic adaptation by prioritizing the transmission of documents that can be transmitted efficiently when the network is good, prioritizing the transmission of frequently accessed documents, and prioritizing the transmission of urgent approval documents. This provides a priority basis for subsequent block-based strategies and encryption strategies.

[0052] This solution addresses the issue of ambiguous priorities in synchronization tasks through a synchronization priority calculation model. The formula comprehensively considers actual network conditions, current bandwidth and jitter, document access popularity, initial popularity and attenuation, approval urgency, and remaining response time, making priority calculation more aligned with actual business needs. The weighting coefficients α, β, and γ can be adjusted according to the enterprise's business priorities. For example, when network stability is critical, α can be increased to ensure that tasks that can be effectively transmitted are prioritized during network jitter. Compared to existing technologies that use no priority or a single factor to determine priority, this multi-factor weighted calculation method is more scientific.

[0053] It's worth noting that: the current instantaneous bandwidth S is obtained through real-time monitoring of network transmission rates; the rated bandwidth S_rated is the bandwidth agreed upon between the enterprise and the network service provider; the network jitter coefficient δ_jitter is calculated by statistically analyzing bandwidth fluctuations over a certain period; the greater the fluctuation, the higher the δ_jitter value. This parameter allows priority to adapt to real-time network changes. Initial access popularity Ht0 counts the number of times a document is accessed within a set period; the time elapsed since the last access t records the time from the last access to the current time; the popularity decay coefficient λ controls the rate of popularity decay over time, giving higher priority to recently accessed documents. The remaining response time T_review for approval nodes is provided by the workflow engine; the shorter the time, the more urgent the approval, and the higher the priority of the corresponding synchronization task, ensuring that approval-related documents are synchronized first.

[0054] The technical effects achieved by the above solution include: calculating synchronization priority based on multiple factors, making the arrangement of synchronization tasks more reasonable and avoiding delays in important or urgent documents due to resource consumption; introducing a network jitter coefficient to adapt priority to network conditions, ensuring feasible synchronization tasks are prioritized when the network is unstable; and considering access frequency and approval urgency to ensure that frequently accessed and urgently approved documents are synchronized first, improving the utilization efficiency of synchronization resources. These effects collectively enhance the targeting and efficiency of the synchronization process, ensuring that critical documents are synchronized in a timely manner and supporting the advancement of core business operations.

[0055] Traditional technical solutions have the following technical problems: the existing file chunking mechanism uses a fixed chunk size and does not take into account the priority of synchronization tasks and differences in file types. This results in increased transmission time for high-priority documents due to excessively large chunks, or difficulty in accurately locating changed chunks for complex files due to unreasonable chunking. The chunking strategy is out of touch with the actual synchronization needs, affecting the efficiency and accuracy of incremental synchronization.

[0056] Based on this, the intelligent synchronization engine also includes a block size calculation step. The block size is determined based on the synchronization priority and is calculated using the following formula: ; Where L is the actual block size in MB; Lbase is the basic block size in MB; P is the synchronization priority; θ is the block adjustment coefficient, ranging from 0.3 to 0.7; μ is the complexity influence coefficient, ranging from 0.2; Ccomplex is the file complexity, ranging from 0 to 2, with values ​​of 0.1 to 0.5 for text files and 1.5 to 2 for 3D models.

[0057] This formula is used to dynamically determine the actual size of document chunks. By associating synchronization priority with the complexity of the file itself, it solves the problem that traditional fixed chunks cannot adapt to the urgency of tasks and file types, and achieves precise matching between chunk size and transmission requirements and file characteristics.

[0058] in, The base block size is the baseline value for block calculation. It can be set according to the common file sizes of enterprises, such as the default 10MB, to ensure that the block size has a unified reference baseline and avoid the calculation results deviating from the actual storage needs.

[0059] Synchronization priority adjustment item: The synchronization priority (0-10) calculated in claim 5 is such that the higher the priority (the larger the P value), the smaller the adjustment term and the smaller the block size. The block adjustment coefficient (0.3-0.7) is used to control the impact of priority on block size. The smaller the value, the smoother the impact of priority changes on blocks.

[0060] For example, the adjustment item for a high-priority document (P=10) is... The significantly smaller file size reduces the amount of data retransmitted during transmission interruptions; the adjustment item for low-priority documents (P=2) is... The block size can be appropriately increased to reduce the redundant overhead of hash calculation.

[0061] This adjustment option enables the adaptation logic of rapidly transmitting small chunks for urgent tasks and appropriately reducing costs by using larger chunks for non-urgent tasks.

[0062] For file complexity adjustment items: The file complexity is set between 0 and 2, depending on the file type. For text files, the value is 0.1-0.5 because the structure is simple, hence the lower value. For 3D models, the value is 1.5-2 because they contain complex structures such as assembly levels, hence the higher value. This is the complexity impact coefficient, a fixed value of 0.2, which controls the extent to which complexity affects the partitioning.

[0063] For example, 3D models ( The adjustment term is 1 + 0.2 × 2 = 1.4. The block division is increased by 40% on the basis of priority adjustment to make the block division fit the assembly structure better, such as natural division according to sub-assemblies. text file The adjustment term is 1 + 0.2 × 0.3 = 1.06, meaning the block division is almost unaffected by complexity, maintaining compactness. This adjustment term ensures that the block division of complex files better matches their data organization, improving the accuracy of subsequent hash comparisons and change block identification.

[0064] Overall, the formula achieves dynamic optimization of block size through a two-layer adjustment that uses priority as the basis and complexity as the adaptation. This ensures both fast transmission of high-priority documents in small blocks and adherence to the block structure for complex files, providing a foundation for efficient incremental synchronization.

[0065] This scheme addresses the adaptability issue of fixed block sizes by combining a block size calculation model that considers both synchronization priority and file complexity. In the formula, the synchronization priority P determines the basic adjustment direction of the blocks; higher priority results in smaller blocks, reducing retransmission costs during transmission interruptions. The file complexity C adjusts the block granularity based on the structural characteristics of different file types, ensuring that the blocks for complex files (such as 3D models) better match their data organization. The settings of the block adjustment coefficient θ and the complexity influence coefficient μ ensure that the block size adjustment better meets actual transmission needs, avoiding unreasonable block sizes caused by extreme values.

[0066] Compared to fixed-block partitioning, this dynamic partitioning strategy achieves a precise match between block size and synchronization requirements and file characteristics.

[0067] It is worth mentioning that: the base chunk size L can be set based on the average size of common enterprise files, serving as a benchmark value for chunk calculation; the synchronization priority P comes from the calculation result of claim 5, directly related to network conditions, access popularity, and approval urgency, ensuring that the chunking strategy is adapted to the urgency of the synchronization task. The file complexity C is determined based on the structural characteristics of the file type; text files have a simple structure and therefore a lower value, while 3D models, due to their complex structures such as assembly hierarchies, have a higher value, allowing for natural segmentation by assembly during chunking, facilitating subsequent hash comparison and change block identification. The chunking adjustment coefficient θ controls the impact of priority on chunk size, avoiding drastic fluctuations in chunk size caused by small changes in priority, ensuring the stability of the chunking strategy.

[0068] The technical effects achieved by the above solution include: the chunk size is dynamically adjusted according to the synchronization priority, with high-priority documents using smaller chunks to reduce the amount of retransmission data after transmission interruption, adapting to urgent synchronization scenarios; the introduction of file complexity makes the chunk granularity fit the file structure, making the chunking of complex files such as 3D models more accurate and improving the efficiency of changed block identification; and the reasonable adjustment of chunk size reduces invalid data transmission while reducing the redundant overhead of hash calculation. These effects enable the chunking strategy to not only meet the transmission needs of tasks with different priorities but also adapt to the characteristics of different types of files, providing a more efficient foundation for incremental synchronization and further improving the adaptability and efficiency of cross-regional multi-warehouse document synchronization.

[0069] Traditional technical solutions have the following technical problems: existing encryption mechanisms mostly use fixed key lengths, without considering file block size and approval time requirements. This results in small block files taking too long to encrypt due to multiple encryption times and fixed key lengths, or urgent approval documents being delayed due to high encryption redundancy. There is an imbalance between encryption strength and transmission efficiency and approval timeliness, making it impossible to balance security and efficiency.

[0070] Based on this, according to the method described in the above embodiments, the transport layer encryption of the multi-layer security protection system adopts a dynamic key encryption method, and the encryption key length is determined based on the block size and calculated using the following formula: ; Where K is the actual encryption key length in bits; Kbase is the basic key length in bits; L is the actual block size in MB; Lbase is the basic block size in MB; Tapproval is the remaining response time of the approval node in hours; Tbaseapproval is the basic approval time in hours; ξ is the impact coefficient of block size on encryption, with a value of 0.4; η is the impact coefficient of approval timeliness on encryption, with a value of 0.6.

[0071] This formula is used to dynamically determine the actual key length for transport layer encryption. By associating block size with approval time requirements, it solves the problem of imbalance between traditional fixed key security strength and transmission efficiency, and achieves accurate matching between encryption strength and actual scenarios.

[0072] in, The base key length serves as the baseline for encryption strength, ensuring that even during dynamic adjustments, the key length does not fall below the security threshold, thus meeting the enterprise's basic encryption requirements for core documents. For block size adjustment items: The actual block size calculated in claim 6; Based on the block size; This represents the impact coefficient of block segmentation on encryption.

[0073] The smaller the block size, the smaller the value of this adjustment item, and the shorter the key length is accordingly. This is because smaller blocks require more encryption times. For example, a 1GB file divided into 1MB blocks requires 1000 encryptions, while a 10MB block only requires 100 encryptions. Shortening the key length can reduce the total encryption time and avoid the transmission delay caused by small blocks and long keys.

[0074] For example, when the actual block size is 1 / 2 of the base block size, the adjustment term is: The key length is reduced to 76% of the basic length, which reduces encryption overhead while ensuring basic security.

[0075] For the approval time limit adjustment item: Remaining response time for the approval node; The basic approval time, in hours, such as the default 24 hours; The impact coefficient of approval time on encryption is fixed at 0.6.

[0076] The shorter the remaining approval time, The smaller the value, the smaller the value of this adjustment item, and the shorter the key length accordingly. Because urgent approvals require rapid transmission, appropriately shortening the key can speed up encryption and avoid process stagnation caused by long keys and urgent approvals.

[0077] For example, when the remaining approval time is 1 / 4 of the base time, the adjustment item is: The key length has been reduced to 38% of the basic length to prioritize the timeliness of approval.

[0078] Overall, the formula achieves a dynamic balance by controlling encryption frequency through block size and controlling transmission urgency through a two-layer adjustment, which moderately reduces the key to improve efficiency in small blocks / urgent scenarios and uses long keys for strong security in large blocks / non-urgent scenarios, thus avoiding the imbalance between security and efficiency caused by a single encryption method.

[0079] This solution addresses the imbalance between encryption strength and actual needs by using a key length calculation model that correlates block size with approval timeliness. In the formula, the actual block size L reflects the frequency of encryption operations; smaller blocks require more encryption operations. The key length is adjusted using the block size's influence coefficient ξ to prevent efficiency degradation due to frequent encryption of small blocks. The remaining response time T of the approval node directly relates to the timeliness requirement of encryption, and the key redundancy of urgent documents is dynamically reduced using the approval timeliness's influence coefficient η. The base key length K ensures the minimum security strength of encryption, and the introduction of block size and approval factors strikes a balance between security and efficiency in the key length. Compared to a fixed key length, this dynamic adjustment mechanism achieves precise adaptation between encryption strategies and transmission scenarios.

[0080] It is worth mentioning that: the actual block size L comes from the calculation result of the block size calculation step, which is directly related to the file type and synchronization priority, ensuring that the key length adjustment and block strategy are coordinated; the basic block size L base corresponds to the basic key length K base, forming a baseline for encryption strength. The remaining response time T_review of the approval node is provided in real time by the cross-warehouse workflow engine. The shorter the time, the more urgent the approval, and the key length is adjusted accordingly to reduce encryption time consumption; the basic approval time T_review base is the enterprise's preset regular approval cycle, serving as a reference benchmark for timeliness judgment. The setting of the block impact coefficient ξ (0.4) and the approval timeliness impact coefficient η (0.6) reflects the priority consideration of approval timeliness, while retaining the basic impact of block size on encryption strength.

[0081] The technical effects achieved by the above scheme include: adjusting the key length according to the block size, appropriately reducing the key length of small block files, reducing the total time spent on encryption operations, and avoiding frequent encryption from affecting transmission efficiency; adjusting the key length according to the remaining approval time, reasonably optimizing the key length of urgent approval documents, and speeding up encryption while ensuring basic security; ensuring that the basic key length ensures that the encryption strength is not lower than the security standard, and the dynamic adjustment mechanism avoids efficiency loss caused by over-encryption.

[0082] These effects enable the encryption strategy to adapt to the transmission needs of different chunks and respond to the timeliness requirements of approval, achieving synergy between security protection, transmission efficiency, and timely approval, and improving the ability to balance security and efficiency in remote document transmission.

[0083] Traditional technical solutions have the following technical problems: existing storage encryption mostly adopts whole file encryption, the block encryption mechanism is imperfect, and metadata and file entities are not stored separately, which may lead to the risk of the entire file being leaked if a single encryption key is leaked; sensitive documents lack effective traceability identifiers, making it difficult to trace the source after leakage, and the accuracy and traceability of storage layer security protection are insufficient.

[0084] Based on this, the storage layer of the multi-layer security protection system adopts a document block encryption storage method. After the document is divided into blocks, a unique encryption key is generated for each block. The encryption key is generated using the national cryptographic algorithm. Metadata and file entities are stored separately. Metadata is stored in the metadata database of the central repository, and file entities are stored in the entity storage area of ​​the corresponding repository. Metadata includes document identifier, block information, encryption key index, and access permission information. Sensitive documents are marked with digital watermarks. The digital watermarks include user IP, operation timestamp, and operation type information. The digital watermarks are embedded in the document using a steganography algorithm and remain visible when the document is copied, transmitted, or printed.

[0085] This solution addresses the accuracy and traceability issues of storage layer security through a combined strategy of block encryption, separate storage, and digital watermarking: block encryption generates a unique key for each block, reducing the impact of a single key leak; the national cryptographic algorithm generates the key to ensure encryption strength; metadata and physical documents are stored separately, making it difficult for attackers to correlate complete document information through metadata even if they obtain the physical file; digital watermarking adds an irremovable traceability identifier to sensitive documents, enabling full tracking of the operation process. These three elements work together to form a comprehensive storage security system encompassing encryption, isolation, and traceability, offering more comprehensive and targeted protection compared to overall encryption and hybrid storage.

[0086] It's worth noting that during block encryption, a unique key for each block is generated using the national standard SM4 algorithm. This key is stored in an independent key management system, physically isolated from the block file. Accessing a block requires retrieving the corresponding key from the key management system via a key index. The encryption key index in the metadata only records the key's association information and does not directly store the key, further reducing the risk of leakage. The metadata database employs an independent access control policy, accessible only to authorized nodes. Digital watermarks are embedded in non-critical data areas of the document using steganography algorithms, without affecting normal document use. Embedded user IP addresses, operation timestamps, and other information can be extracted using specialized tools. Even if the document is modified, core traceability information can still be identified, ensuring accurate identification of the source of the operation after a leak.

[0087] The technical effects achieved by the above solution include: block encryption and unique keys reduce the impact of single-key leakage and improve the accuracy of storage encryption; separate storage of metadata and entities increases the difficulty for attackers to obtain complete documents and strengthens storage isolation; digital watermarking provides traceable identification for sensitive documents, enabling source tracing after leakage. These effects collectively enhance the security protection capabilities of the storage layer, reduce the risk of leakage during document storage, and provide reliable evidence for tracing the source of leakage incidents, meeting enterprises' high requirements for the security of electronic document storage.

[0088] Traditional technical solutions have the following technical problems: existing access control mostly adopts single role control (RBAC) or attribute control (ABAC). RBAC is difficult to deal with permission adjustments in dynamic scenarios, and ABAC lacks the support of a basic permission framework, resulting in permission management that is either too rigid or lacks standardization; the permission judgment logic of user access requests is simple and cannot accurately control the access scope by combining user roles and actual attributes, which poses a risk of unauthorized access.

[0089] Based on this, the access control of the multi-layered security protection system adopts a hybrid RBAC and ABAC model. The RBAC model predefines roles and the basic permissions corresponding to each role, including read, modify, delete, and approve. The ABAC model determines dynamic permissions based on attributes, including the user's repository, document storage area, and document security level. When a user initiates an access request, the access control module first uses the RBAC model to determine whether the user has basic permissions, and then uses the ABAC model to determine whether the user's attributes meet the dynamic permission requirements. Only when both basic and dynamic permission requirements are met is the user allowed to perform the corresponding access operation.

[0090] This solution addresses the rigidity and inconsistencies of a single access control model by employing a hybrid RBAC and ABAC model. RBAC provides a basic permission framework, ensuring clear benchmarks for permissions across different roles and preventing chaotic permission management. ABAC enables dynamic permission adjustments based on attributes, adapting to dynamic changes in user repositories, document security levels, and other scenarios. The two-layered judgment logic of "basic permissions + dynamic permissions" makes permission control both standardized and flexible. Compared to a single model, the hybrid model balances the standardization of permission management with scenario adaptability, resulting in more precise judgment logic.

[0091] It's worth noting that roles in the RBAC model are defined according to the enterprise's organizational structure and business needs, such as "Regional Warehouse Engineer" and "Central Warehouse Approver." Each role's basic permissions are clearly defined through a permission matrix. For example, a "Regional Warehouse Engineer" by default has read and modify permissions for local documents but no delete permissions. The attributes of the ABAC model are collected in real-time by the system. A user's warehouse is associated with their geographical location, document storage areas correspond to physical storage locations, and document security levels are categorized according to the enterprise's confidentiality levels (e.g., public, internal, secret). Dynamic permission rules are pre-configured, such as "Only users in the East China warehouse are allowed to access internal and lower-level security documents stored in the East China warehouse." When an access request is triggered, the access control module first queries the RBAC model to verify basic permissions, then calls the ABAC model's attribute matching engine to verify dynamic rules. An access permission instruction is generated only after both layers of verification pass.

[0092] The technical effects achieved by the above solution include: the RBAC model ensures clear benchmarks for access control, avoiding arbitrary permission allocation; the ABAC model enables permissions to be dynamically adjusted based on user and document attributes, adapting to the complex scenarios of multi-site warehouses; and the two-layer judgment logic strictly controls the access scope, reducing the risk of unauthorized access. These effects collectively improve the accuracy and flexibility of access control, ensuring that users in different regions and with different roles can only access documents within their authorized scope, providing reliable permission-level protection for document security in multi-site warehouse environments.

[0093] Traditional technical solutions have the following technical problems: existing regional warehouses lack an effective load balancing mechanism. When a regional warehouse experiences a peak in concurrent access, load parameters such as CPU, memory, and network bandwidth exceed the carrying capacity threshold, resulting in local cache loading delays, slow synchronization task processing, and the inability to quickly alleviate the overload state, affecting the document access and synchronization efficiency of the regional warehouse, and consequently impacting cross-warehouse collaboration.

[0094] Based on this, the 1+N hybrid storage architecture also includes a load balancing mechanism, which monitors the CPU utilization, memory utilization, and network bandwidth utilization of each regional warehouse in real time. When the load parameters of a regional warehouse exceed a preset threshold, the load balancing mechanism migrates some of the high-frequency access document cache copies of that regional warehouse to adjacent regional warehouses with lower loads. At the same time, it adjusts the incremental synchronization task allocation of that regional warehouse, transferring some synchronization tasks to the central warehouse for execution. Once the load parameters of the regional warehouse drop below the threshold, it gradually restores its original cache copies and synchronization task allocation.

[0095] This solution addresses the issue of excessive load on regional warehouses through a load balancing mechanism involving real-time monitoring, cache migration, and task transfer. Real-time monitoring ensures timely detection of load anomalies, preventing prolonged overload. Cache replica migration distributes high-frequency access pressure to adjacent regional warehouses, reducing local access latency. Synchronous task transfer reduces the processing pressure on regional warehouses, creating conditions for load shedding. Gradual migration after load recovery ensures service continuity. Compared to solutions without load balancing or with only simple adjustments, this closed-loop "monitoring-distribution-recovery" mechanism can quickly alleviate load pressure and adapt to dynamically changing access and synchronization needs. Notably, the load balancing mechanism collects load parameters in real-time through proxy nodes deployed in each regional warehouse, with a collection frequency of once per second, ensuring timely capture of instantaneous peak values. Preset thresholds are set according to the hardware configuration of the regional warehouses; for example, CPU utilization thresholds are set to 80%, memory utilization thresholds to 85%, and network bandwidth utilization thresholds to 90%. A load balancing operation is triggered when any parameter exceeds the threshold. The migration of frequently accessed document cache copies is based on access frequency statistics, selecting the top 20% of documents accessed in the most recent hour. The migration process uses incremental transmission, transferring only the differences in the cache copies to avoid consuming excessive network resources. Synchronization task transfer prioritizes non-urgent incremental synchronization tasks. After receiving the tasks, the central repository uses a dedicated processing thread to execute them, ensuring no impact on the central repository's core functions. After load recovery, cache copies and synchronization tasks are gradually migrated back using a canary approach, with each migration not exceeding 20% ​​of the original total to avoid triggering further load fluctuations.

[0096] The technical effects achieved by the above solution include: real-time monitoring can promptly detect high load conditions, buying time for balancing operations; cache migration distributes local access pressure, reducing loading latency for frequently accessed documents; synchronous task transfer reduces the processing burden on regional warehouses and accelerates load shedding; and gradual migration during the recovery phase avoids service interruptions. These effects collectively ensure that regional warehouses maintain stable operation even under peak concurrent access, reducing document access and synchronization latency caused by high load, providing stable storage and synchronization support for multi-warehouse collaboration in different locations, and guaranteeing the continuity of cross-warehouse collaboration.

[0097] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A PLM-based method for collaborative storage of electronic documents across multiple warehouses in different locations, characterized in that, include: S1: Construct a 1+N hybrid storage architecture, in which one central warehouse is deployed in the enterprise headquarters data center and stores full document metadata and the latest version of the file, and N regional warehouses are deployed according to geographical regions and store cached copies of frequently accessed local documents and incremental update data; S2: Incremental synchronization of documents across multiple warehouses in different locations is achieved through an intelligent synchronization engine, which employs a content-aware incremental synchronization algorithm and a dynamic synchronization strategy based on network quality. S3: Version management is achieved through a version collaboration control module, which is used to establish a global version number mapping mechanism and a conflict resolution decision tree; S4: Cross-warehouse process automation is achieved through a cross-warehouse workflow engine, which integrates a BPM process engine and a message queue; data security is achieved through a multi-layer security protection system, which includes transmission layer encryption, storage layer encryption, and access control.

2. The method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The content-aware incremental synchronization algorithm of the intelligent synchronization engine achieves incremental transmission through file block hash comparison. Specifically, it dynamically adjusts the block granularity according to document type, blocks text files according to preset size, blocks 3D models according to assembly, and generates a unique hash value for each block. The change blocks for document updates are determined by hash value comparison, and only the change blocks are transmitted to reduce the amount of data transmitted. The dynamic synchronization strategy based on network quality includes real-time monitoring of network bandwidth in different locations. When the network bandwidth is lower than a preset threshold, it automatically switches to asynchronous compressed transmission mode to improve transmission stability.

3. The method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The global version number mapping mechanism of the version collaboration control module generates a unique global identifier for each document. The global identifier includes a timestamp, a repository identifier, and a document type code. Before updating a document, a regional repository needs to send a version lock application to the central repository. The central repository reviews the application and provides feedback on the review results. Once the review is approved, the regional repository obtains a version lock for a preset duration. Within the validity period of the version lock, the regional repository completes the update and synchronizes the updated data with the central repository. When a version conflict is detected, the conflict resolution decision tree judges the conflict in the following order: central repository priority, modification time priority, and approval status priority. Based on the judgment results, the conflicting versions are automatically merged. If automatic merging cannot be completed, the conflict information is pushed to the administrator for manual arbitration.

4. The method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The BPM process engine of the cross-warehouse workflow engine predefines cross-warehouse approval templates. The cross-warehouse approval templates include design change templates, version update templates, and permission adjustment templates. Each template clearly defines the process node and the corresponding processing warehouse for each node. When a cross-warehouse approval event is triggered, the BPM process engine generates an approval task according to the corresponding template and pushes the approval task to the terminal of the person in charge of the corresponding warehouse. Message queues are used to implement cross-warehouse event notifications. When a document is checked in, updated, or its approval status changes, the message queue collects event information in real time and pushes the event information to the subscriber terminals of the associated warehouse according to preset subscription rules.

5. The method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The intelligent synchronization engine also includes a synchronization priority calculation step, wherein the synchronization priority is calculated using the following formula: ； Where P is the synchronization priority, ranging from 0 to 10; Scurrent is the current instantaneous bandwidth; Srated is the rated bandwidth; δjitter is the network jitter coefficient, ranging from 0 to 1, with the larger the network jitter, the closer δjitter is to 1; Ht0 is the initial access popularity; t is the time since the last access; λ is the popularity decay coefficient, ranging from 0.01 to 0.1; Tapproval is the remaining response time of the approval node; α, β, and γ are weighting coefficients, and α + β + γ = 1.

6. The method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 5, characterized in that, The intelligent synchronization engine also includes a block size calculation step, whereby the block size is determined based on the synchronization priority and calculated using the following formula: ； Where L is the actual block size; Lbase is the base block size; P is the synchronization priority; θ is the block adjustment coefficient, ranging from 0.3 to 0.7; μ is the complexity influence coefficient, ranging from 0.2; Ccomplex is the file complexity, ranging from 0 to 2, with values ​​of 0.1 to 0.5 for text files and 1.5 to 2 for 3D models.

7. A method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 6, characterized in that, The transport layer encryption in the multi-layered security protection system uses dynamic key encryption. The encryption key length is determined based on the block size and is calculated using the following formula: ； Where K is the actual encryption key length; K is the base key length; L is the actual block size; L is the base block size B; T_approval is the remaining response time of the approval node; T_approval is the base approval time; ξ is the impact coefficient of block size on encryption, with a value of 0.4; η is the impact coefficient of approval time on encryption, with a value of 0.

6.

8. A method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The storage layer encryption of the multi-layer security protection system adopts a document block encryption storage method. After the document is divided into blocks, a unique encryption key is generated for each block. The encryption key is generated using the national cryptographic algorithm. Metadata and file entities are stored separately. Metadata is stored in the metadata database of the central repository, while file entities are stored in the entity storage area of ​​the corresponding repository. Metadata includes document identifiers, block information, encryption key indexes, and access permission information. Add digital watermarks to sensitive documents. The digital watermark contains user IP, operation timestamp, and operation type information. The digital watermark is embedded in the document through a steganography algorithm and remains visible when the document is copied, transmitted, or printed.

9. A method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The access control of the multi-layered security protection system adopts a hybrid model of RBAC and ABAC. The RBAC model predefines roles and the basic permissions corresponding to each role. The basic permissions include read, modify, delete, and approve; The ABAC model determines dynamic permissions based on attributes, including the user's repository, document storage area, and document security level. When a user initiates an access request, the access control module first uses the RBAC model to determine whether the user has basic permissions, and then uses the ABAC model to determine whether the user's attributes meet the dynamic permission requirements. Only when both basic and dynamic permission requirements are met can the user be allowed to perform the corresponding access operation.

10. A method for collaborative storage of electronic documents across multiple warehouses based on PLM according to claim 1, characterized in that, The 1+N hybrid storage architecture also includes a load balancing mechanism that monitors the CPU utilization, memory usage, and network bandwidth utilization of each regional warehouse in real time. When the load parameters of a regional repository exceed a preset threshold, the load balancing mechanism will migrate some of the frequently accessed document cache copies of the regional repository to an adjacent regional repository with a lower load. At the same time, the allocation of incremental synchronization tasks in the regional warehouses will be adjusted, and some synchronization tasks will be transferred to the central warehouse for execution. Once the load parameters of the regional warehouses drop below the threshold, their original cached copies and synchronization task allocations will be gradually restored.

Images