Object-storage-oriented efficient AES-GCM encryption and decryption method

Abstract

The invention discloses an object-storage-oriented efficient AES-GCM encryption and decryption method, relates to the technical field of data encryption and decryption, and aims to integrally solve the problems of confidentiality, integrity and encryption and decryption performance of data. The key points of the technical scheme of the method comprise: (1) an encryption stage: after a client uploads an object, a system divides data into blocks and allocates a unique IV to each block; next, a key is obtained from the KMS through the Key ID, each data block is encrypted by adopting an AES-GCM algorithm to obtain a ciphertext and an authentication tag (Tag), and the ciphertext and the authentication tag (Tag) are stored in metadata together with the Key ID and IV; (2) a decryption stage: when a user downloads an object, the system obtains the Key ID, the IV and the Tag from the metadata, and requests a corresponding key from the KMS; and decrypting each data block in parallel and verifying the Tag, and after the verification is passed, splicing the data into a complete object and returning the complete object. According to the method, the authentication encryption characteristic of the AES-GCM is utilized, the data integrity can be guaranteed without additional calculation, and the safety performance and the data processing efficiency in a large-scale object storage scene are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of data encryption and decryption technology, and in particular to an efficient AES-GCM encryption and decryption method for object-oriented storage. Background Technology

[0002] With the rapid development of cloud computing and big data technologies, object storage has become a key technology for storing unstructured data. As data security demands increase, data security and privacy protection requirements are becoming increasingly prominent, especially in sensitive data storage scenarios. It is necessary not only to ensure data confidentiality but also to guarantee data integrity and authenticity. Traditional simple encryption schemes cannot meet these security requirements. In object storage systems, data is stored and managed based on objects as the basic unit, possessing high scalability, high reliability, and high availability. Through the separate management of data and metadata, it provides a flexible storage model. However, with the increasing demands for data security, ensuring the confidentiality and integrity of data in object storage systems has become a crucial issue. However, traditional technologies have the following problems:

[0003] Traditionally, many object storage systems use AES-CBC for data encryption, but AES-CBC has limitations in authentication and performance. It relies on an additional Message Authentication Code (MAC) to ensure data integrity and authenticity, increasing system complexity and processing overhead. Furthermore, the serial processing nature of AES-CBC limits its performance in parallelized scenarios.

[0004] As the closest existing technology, "object storage + AES-CBC + HMAC" can be regarded as the baseline solution. However, it faces problems such as insufficient parallelism, large overhead of additional MAC maintenance and verification, and high complexity of key and IV management in high concurrency and large object block scenarios. It is difficult to balance performance and security consistency, thus forming the objective technical problem that this invention aims to solve.

[0005] AES-GCM is a highly efficient authentication and encryption algorithm that combines encryption and authentication functions. It effectively ensures the confidentiality and integrity of data while supporting parallel processing. This makes AES-GCM an ideal choice for object storage data encryption and decryption.

[0006] AES-GCM not only guarantees data confidentiality but also provides protection for data integrity and authenticity. However, directly using AES-GCM in an object storage environment still faces several challenges, including: how to efficiently and flexibly manage keys, how to store and quickly retrieve encryption parameters for each data block, and how to efficiently process and verify each data block during both object upload and download phases. Therefore, it is necessary to propose an improved object storage data encryption and decryption method based on the AES-GCM algorithm. Through reasonable key management, data block processing, and authentication tag management strategies, this method can achieve secure and reliable data storage in the backend and provide efficient encryption, decryption, and integrity verification capabilities during object upload and download.

[0007] Therefore, it is necessary to provide an efficient AES-GCM encryption and decryption method for object-oriented storage to solve the above problems. Summary of the Invention

[0008] The purpose of this invention is to provide an efficient AES-GCM encryption and decryption method for object storage, which is suitable for object storage systems.

[0009] This method encrypts data into multiple blocks during object upload, generating initialization vectors (IVs) and authentication tags for each block and storing corresponding encrypted metadata to ensure data confidentiality and integrity. During object download, the corresponding key, IV, and tag are retrieved from the metadata, and each block is decrypted and verified in parallel to ensure data integrity and rapid recovery of the original plaintext. This method maintains high performance and low additional storage overhead even with large-scale objects and high concurrency.

[0010] To achieve the above objectives, the technical solution adopted by the present invention includes the following steps:

[0011] An efficient AES-GCM encryption and decryption method for object-oriented storage includes an encryption phase and a decryption phase.

[0012] (1) During the encryption process:

[0013] (1.1) During the process of transferring objects to the storage system, encryption parameters are configured in the object metadata header, and the AES-GCM block algorithm is adopted. At the same time, the 256-bit key length and 128-bit tag length are explicitly set to achieve dual protection of data confidentiality and integrity.

[0014] (1.2) The object storage system divides each object data to be stored into multiple data blocks, and then assigns a unique initialization vector (IV) to each data block to be encrypted. The IV is 96 bits long. The IV adopts a structured design, containing a fixed part and a variable part. The fixed part contains a node identifier (16 bits) and a timestamp accurate to milliseconds (32 bits), while the variable part contains a high-entropy random number (32 bits) and an intra-object sequence counter (16 bits). The IV generation process utilizes a hybrid entropy source deterministic random bit generator (DRBG), combined with a hardware entropy source to improve the randomness quality. The system maintains a distributed IV registry, using multi-level Bloom filters to quickly detect potential IV conflicts, ensuring that IVs are globally unique within the same Key ID domain in a large-scale distributed environment. To this end, registry writes follow strong consistency (such as linearized writes, based on Raft or Paxos consensus), and Bloom filters are only used for fast screening. After a hit, a strong consistency KV secondary verification is performed. The registry key space displays Key ID and IV structure fields to avoid cross-key misjudgment. The IV serves as the initial value for the AES-GCM counter, ensuring that the counter space does not overlap when encrypting multiple data blocks in parallel.

[0015] (1.3) The data block is encrypted using the AES-GCM algorithm to obtain the encrypted data block and the corresponding tag. During system deployment, a Key Management Service (KMS) is configured, and the KMS is responsible for generating, storing, distributing, and rotating the keys.

[0016] (1.4) When data is needed during the encryption process, only the affected data blocks need to be re-encrypted and the metadata updated, without affecting the validity and integrity of the remaining data blocks;

[0017] (1.5) Finally, the object storage system stores each encrypted data block in the backend object storage system, while simultaneously recording the key ID, IV, and tag information in the object's metadata, completing the entire encrypted upload process. The metadata itself employs an independent encryption protection mechanism to ensure metadata security, and uses an efficient index structure to optimize query performance.

[0018] (2) During the data decryption stage:

[0019] (2.1) When a user downloads an object, based on the requested object identifier and the required data range, the key identifier (Key ID), IV and Tag information required for the corresponding data block are obtained from the object metadata, and the corresponding key is obtained from the Key Management Service (KMS);

[0020] (2.2) Read out the object in data blocks of a set size. For each data block, use the AES-GCM algorithm along with the IV and Tag to decrypt and verify the integrity of the encrypted data block. If the authentication verification passes, return the successfully decrypted plaintext data to the requester. After all data blocks have been authenticated and decrypted, the entire object download process is complete.

[0021] (2.3) In high-concurrency scenarios, multiple data blocks can be decrypted simultaneously, and the data blocks can be reassembled to restore the original object data after successful verification, thereby improving download response speed. The system will establish a performance monitoring mechanism to dynamically adjust the degree of parallelism based on the load and optimize resource utilization.

[0022] The specific encryption process is as follows:

[0023] (1) For each data block, before encryption, the Key ID is obtained from the storage policy where the object resides, and then a 96-bit random IV value is generated using a random number generator. The IV structure is: 16-bit node identifier + 32-bit timestamp + 32-bit random number + 16-bit sequence number, ensuring that the entropy of the IV is high enough to prevent collisions and predictions. The system maintains a global IV registry to avoid IV reuse in a distributed environment. The global IV registry uses sharded storage and a hierarchical Bloom filter structure to support high-concurrency queries and updates. Each storage node first checks the IV in its local cache and then verifies its global uniqueness with the coordinating node. Then, a 256-bit AES key Key is obtained from KMS using the Key ID;

[0024] (2) Using the AES-GCM encryption algorithm, with Key as the key and IV as the initialization vector input, the plaintext data of the data block is encrypted to generate ciphertext and a 128-bit Tag;

[0025] The tag generation process first divides the plaintext data into 128-bit data blocks, then performs Galois field multiplication on each data block, and incorporates Additional Authentication Data (AAD). The AAD contains: object metadata hash, data block location information, and access control policy hash. Finally, a GHASH value is generated and combined with an encryption counter to obtain the complete tag. To ensure cross-implementation consistency, the AAD fields use a fixed order and encoding (TLV or CBOR is recommended for unified byte order), the hash algorithm uses SHA-256, and a version number field is included in the AAD. When the version is upgraded or rolled back, version and encoding conventions ensure verification consistency.

[0026] (3) Record the Key ID, IV, and Tag of the data block in the object metadata. In this way, even if the key corresponding to the Key ID is changed in the future, the metadata can still provide the correct reference, so that the old data block can be re-encrypted or correctly decrypted when necessary;

[0027] (4) Process all data blocks according to steps (1), (2), and (3), and finally update and submit the metadata of the object to the metadata storage system to complete the encryption of all data blocks of the object to be uploaded;

[0028] (5) If an anomaly occurs during the encryption process (e.g., the key cannot be obtained from KMS, or AES calculation fails), the system will record detailed error information in the security log for subsequent analysis and auditing. Simultaneously, the system will interrupt the upload process and return an error message to the client to ensure that data is not stored in an inconsistent encrypted state.

[0029] The specific decryption process is as follows:

[0030] (1) When a client requests to download object data, the system determines the set of data blocks to be read based on the requested offset and length. It then searches the metadata for the Key ID, IV, and Tag corresponding to each data block.

[0031] (2) Request the corresponding key from KMS using the key ID. For each data block to be decrypted, use the AES-GCM algorithm, taking the obtained key and the recorded IV as input, and inputting the ciphertext read from the storage system into the decryption function;

[0032] (3) If the key cannot be obtained from KMS during the decryption process, or the ciphertext is read incorrectly, or the AES decryption calculation is abnormal, the system will record detailed error information and may trigger an automatic recovery process. The system adopts a tiered retry strategy: for temporary failures (such as network timeout), it will automatically retry up to 3 times, with the interval increasing exponentially; for persistent failures, decryption will terminate and return a failure message.

[0033] (4) AES-GCM verifies the correctness of the tag during the decryption process. If the tag verification passes, the output is the original plaintext data of the data block. If the verification fails, the system will trigger a response mechanism and notify the upper-layer service to return an error response. At the same time, it will record a complete security event log for subsequent investigation, including the data block identifier, timestamp, failure type and access source information;

[0034] (5) Process all data blocks according to the previous steps (1), (2), (3), and (4). Concatenate all the decrypted plaintext data blocks in their original order to obtain the complete object data, which is then returned to the requester. The system uses an adaptive thread pool to dynamically adjust the number of concurrent decryption threads based on server CPU utilization, memory status, and IO load. Simultaneously, it incorporates data locality optimization, prioritizing the loading and decryption of consecutive data blocks to improve the download response speed of large objects.

[0035] In the context of object storage block encryption and decryption, this invention achieves overall synergy through the following combination of features:

[0036] (1) Object block parallel encryption and decryption, combined with the design of 96-bit structured IV (16 nodes + 32 timestamps + 32 random numbers + 16 sequences), ensures concurrency and uniqueness;

[0037] (2) The combination of strong consistency registry verification (linearized write) and hierarchical Bloom pre-screening reduces write amplification and latency under high concurrency while ensuring uniqueness;

[0038] (3) Standardized AAD encoding (TLV or CBOR, fixed order and uniform byte order, including version field, field value SHA-256) ensures cross-implementation consistency and stability of rollback / upgrade verification;

[0039] (4) Metadata only wraps the key material (AES-KW), avoiding wrapping the entire metadata, reducing overhead and narrowing the attack surface;

[0040] (5) Dynamic parallelism and load adaptation (based on CPU / memory / IO feedback control) maintain performance stability in both the upload and download stages.

[0041] The above elements are not simply juxtaposed, but rather revolve around the core security constraints of "IVs cannot be reused under each key" and "parallel counters are mutually exclusive". By introducing Key ID and IV structure fields into the registry key space, and combining Bloom pre-screening and strong consistency secondary verification, we can achieve secure scalability under high concurrency. At the same time, we use standardized AAD and the minimum wrapping strategy for critical materials to ensure robustness across versions and implementations.

[0042] Formal constraints and checks for non-overlapping counter spaces:

[0043] (1) Constraint definition: For the same Key ID, any two data blocks b i ≠b j Its GCM counter sequence interval C(b i ) and C(b j ) satisfies C(b) i )∩C(bj =∅. The starting point of the counter is determined by the 96-bit IV, and the counter for the k-th block within the block is ctr = inc32(IV, k), k∈[0, n]. i );

[0044] (2) Allocation strategy: The block number within the object is monotonically mapped to the sequence field in the IV, and wraparound is prohibited; when the sequence field approaches the upper limit, rekeying or object-level resharding is triggered;

[0045] (3) Detection mechanism:

[0046] Before generation: Perform a duplicate check on the (Key ID, IV structure field) in the registry;

[0047] At runtime: Maintain local Bloom and LRU sets for active sessions to detect short-lived duplicates;

[0048] Audit period: Periodic sampling statistical counter interval, abnormal triggering alarms and circuit breakers;

[0049] (4) Boundary handling: When any detection step determines a potential overlap or reuse risk, the IV is immediately discarded, and the metadata is regenerated and re-signed; if necessary, the Key ID is put into a rate-limiting or circuit-breaking state.

[0050] Misuse prevention and threshold strategies:

[0051] (1) Limits and rekeys: Set the maximum cumulative bytes of encrypted plaintext and the maximum number of blocks for each Key ID (e.g., no more than 232 blocks). When the threshold is reached, the Key will be automatically rotated and further allocation will be prevented.

[0052] (2) Replay / rollback protection: AAD includes version and object block indexes, and the server maintains the highest confirmed version, rejecting writes of older version tags that have passed verification;

[0053] (3) Entropy source degradation protection: Monitor DRBG health and hardware entropy source supply. When the health test fails, switch to the backup entropy source and reduce the allocation rate, and issue an alarm at the same time.

[0054] (4) Abnormal handling: When the registry is unavailable, Bloom FP surges, or consistency lags, the system enters degrade mode (small batch strong verification, reduced parallelism) or read-only protection to avoid unrecoverable IV reuse.

[0055] The present invention has the following beneficial effects and advantages:

[0056] (1) This invention encrypts object data stored in object storage using the AES-GCM encryption algorithm. In addition to encrypting the data, it also provides authentication encryption capabilities to ensure data integrity and avoid the risk of tampering. Compared with the AES-CBC mode, no additional MAC operation is required, reducing computational overhead and implementation complexity;

[0057] (2) The present invention uses the AES-GCM encryption algorithm to encrypt the object data stored in the object storage, supports parallel encryption and decryption operations, and can use multi-core CPU or GPU to accelerate the encryption and decryption processing of object data blocks, thereby improving the performance of large-scale data processing.

[0058] (3) This invention uses the AES-GCM encryption algorithm to encrypt the object data stored in the object storage. Through innovative structured IV design and multi-level collision detection mechanism, it ensures the global uniqueness of IV while maintaining high performance.

[0059] (4) This invention implements an adaptive encryption strategy, which dynamically adjusts the encryption strength and processing priority based on data sensitivity, access frequency, and storage environment, thereby optimizing system resource utilization while ensuring security. For highly sensitive data, the system automatically adopts a more complex AAD construction and a more stringent verification process, while optimizing performance for low-sensitivity data. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating the encryption process for an uploaded object according to an embodiment of the present invention;

[0061] Figure 2 This is a flowchart illustrating the decryption process for a downloaded object according to an embodiment of the present invention;

[0062] Figure 3 This is a flowchart of the tag generation process during encryption;

[0063] Figure 4 This is a flowchart of Tag verification during the decryption process. Detailed Implementation

[0064] To make the technical solution, the technical problem solved, and the technical effects of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with specific embodiments. The present invention achieves a high-efficiency, high-security data encryption and decryption system in an object storage environment through innovative architecture design and algorithm optimization.

[0065] Example 1: Security-enhanced object storage encryption and decryption system.

[0066] This embodiment proposes an AES-GCM-based data encryption and decryption method for object storage. Taking a 10MB object as an example, it systematically illustrates the entire process of data encryption and decryption. This embodiment fully integrates multiple innovative technologies to form a complete technical solution.

[0067] (I) Reference Appendix Figure 1 and attached Figure 3 During the data encryption phase:

[0068] 1. Upon receiving an upload request, the object storage node confirms the object size is 10MB and plans to divide it into three data blocks: Chunk0 = 4MB, Chunk1 = 4MB, and Chunk2 = 2MB. The data block size can be dynamically adjusted according to system configuration to optimize encryption performance and storage efficiency.

[0069] 2. Obtain the Key ID from the system's currently effective key policy. Based on the adaptive security policy, the system determines the security level of this encryption operation to be "high sensitivity" and selects appropriate IV generation and Tag verification parameters accordingly.

[0070] 3. The system generates initialization vectors for each data block, starting with an IV0 for Chunk0. The following values ​​illustrate this: a 16-bit node ID (0x4D2A) is a unique identifier pre-assigned to this storage node; a 32-bit timestamp (0xA7C31D45) is the current system time converted to 32-bit hexadecimal; a 32-bit random number (0xF7E93C12) is generated in real-time by the TPM hardware security module and has cryptographic strength; and a 16-bit sequence number increments from 0x0001. The IV0 combines these elements to form a complete 96-bit value: 0x4D2AA7C31D45F7E93C120001 (Node ID + Timestamp + Random Number + Serial Number). This IV is submitted to the distributed registry and verified by a Bloom filter to be conflict-free.

[0071] The system generates IV1 (0x4D2AA7C31D45F7E93C120002) and IV2 (0x4D2AA7C31D45F7E93C120003) with incrementing sequence numbers for subsequent data blocks. All IVs undergo multi-layer verification to ensure global uniqueness in a distributed environment and eliminate the risk of IV reuse.

[0072] 4. Use the Key ID to request the corresponding key from KMS. After KMS returns the key through a secure channel, the object storage node temporarily holds a memory copy of the key. The key is deleted immediately after use in memory and is not stored for a long time, reducing the risk of key exposure.

[0073] 5. Encrypt Chunk0 using the AES-GCM algorithm to obtain the complete Ciphertext0, and generate a 16-byte Tag0. Only maintain the temporary GHASH accumulation state during streaming computation; do not persist it or write metadata.

[0074] 6. Write Ciphertext0 to backend storage (such as a distributed storage cluster), and record the Key ID, IV0, and Tag0 associated with Chunk0 in the metadata. The metadata is protected using an independent encryption mechanism, employing the AES-KW algorithm to protect key-related fields (only wrapping the key or key material, not the entire metadata), ensuring that even if the metadata is accessed, the actual encryption parameters cannot be obtained. The Key ID is associated with the wrapped material for easy decryption and location later.

[0075] 7. Repeat the above process to encrypt again using IV1 and IV2 respectively, to obtain Ciphertext1, Tag1 and Ciphertext2, Tag2. The metadata records the Key ID, IV1, Tag1 of Chunk1 and the Key ID, IV2, Tag2 of Chunk2;

[0076] 8. After all data blocks are encrypted, the object metadata is persisted to the metadata store. The upload process is complete, and the client receives a response indicating that the object has been successfully uploaded and securely encrypted and stored.

[0077] 9. The system records complete encrypted audit logs, including operation time, operation results, etc., but excluding sensitive key materials. All audit logs are stored in an append-only manner and protected by digital signatures to prevent tampering.

[0078] (II) Reference Appendix Figure 2 and attached Figure 4 During the data decryption phase:

[0079] 1. When a user initiates an object download request, the system first performs a comprehensive authentication and permission check. Based on the request parameters (such as object identifier and data range), the system determines the specific data block that needs to be decrypted. Through metadata services, the system extracts the object's structural information, confirming that the 10MB object consists of three encrypted data blocks: Chunk0 (4MB), Chunk1 (4MB), and Chunk2 (2MB).

[0080] 2. The system retrieves the encrypted parameters associated with the target object from the secure metadata store. First, it verifies the integrity of the metadata itself, checking the digital signature or HMAC of the metadata to ensure it has not been tampered with. Then, it extracts the Key ID, IV, and Tag information corresponding to each data block from the metadata.

[0081] 3. The system uses the Key ID from the metadata to initiate a key request to KMS. KMS returns the corresponding key through a secure channel. The system temporarily stores the key in a secure memory area to ensure that key materials are not leaked or persistently stored;

[0082] 4. Read Ciphertext0 from the storage backend, use the same AES-GCM process to decrypt each data block, and calculate the authentication function GHASH in GCM mode to obtain the expected Tag'0;

[0083] 5. Compare Tag'0 with the Tag0 recorded in storage. If they match, authentication is successful, and the system transmits the decrypted plaintext data to the client. If they do not match, it indicates that the data may have been tampered with, decryption fails, the system interrupts the processing and returns a detailed error code and description, and triggers a security alarm.

[0084] 6. Similarly, decryption and verification are performed according to their respective IVs and tags. Performance optimization measures include: data prefetching, parallel decryption processing (dynamically adjusted according to the number of CPU cores), decryption result caching, and zero-copy data transfer technology, which significantly reduce latency and improve throughput;

[0085] 7. After all data blocks are decrypted and authenticated, the plaintext of Chunk0, Chunk1, and Chunk2 are concatenated into complete object data. The system uses intelligent buffer management, which may decrypt only a portion of the data blocks depending on the requested range. It also supports parallel decryption and streaming transmission for range requests, optimizing the experience of accessing large objects.

[0086] In summary, the efficient AES-GCM encryption and decryption method for object storage proposed in this invention can perform AES-GCM encryption on object data stored in an object storage system, thereby effectively ensuring data security.

[0087] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. These embodiments are only used to help understand the core technical content of the present invention and do not constitute a limitation on its protection scope. Based on the above specific embodiments, any improvements, modifications, or adjustments made by those skilled in the art without departing from the principles of the present invention should be considered to fall within the patent protection scope of the present invention.

Claims

1. A high-efficiency AES-GCM encryption and decryption method for object-oriented storage, characterized in that, It includes an encryption process and a decryption process; (a) During the encryption process, First, during object upload, encryption parameters are configured through the object metadata header, using the AES-GCM block designator algorithm. Then, the object storage system divides the object data to be stored into multiple data blocks, assigning each data block a unique 96-bit structured initialization vector (IV). Next, the AES-GCM algorithm is used to encrypt each data block in parallel, generating encrypted data blocks and corresponding authentication tags. Finally, the encrypted data blocks are stored in the backend object storage system, and the key ID, IV, and tag information are recorded in the object's metadata. (ii) During the decryption process, First, when a user downloads an object, the system retrieves the Key ID, IV, and Tag information of the corresponding data block from the object's metadata based on the request, and obtains the corresponding key from the Key Management Service (KMS). Then, for each data block, the system performs AES-GCM decryption and integrity verification in parallel using the obtained key, IV, and Tag. Finally, if the authentication verification passes, all successfully decrypted plaintext data blocks are reassembled in their original order to restore the complete object data and return it to the requester.

2. The method according to claim 1, characterized in that, The 96-bit structured initialization vector (IV) comprises a fixed part and a variable part, wherein the fixed part contains a node identifier and a timestamp, and the variable part contains a high-entropy random number and an in-object sequence counter.

3. The method according to claim 2, characterized in that, The system ensures the global uniqueness of IVs within the same Key ID domain by maintaining a distributed IV registry that employs multi-level Bloom filters and strong consistency checks.

4. The method according to claim 1, characterized in that, During the encryption process, the generation of the authentication tag also incorporates Additional Authentication Data (AAD), which includes the hash value of object metadata, data block location information, and access control policy hash value to enhance security.

5. The method according to claim 4, characterized in that, The AAD uses a fixed order and encoding, and includes a version number field to ensure consistency and stability across implementations and versions.

6. The method according to claim 1, characterized in that, The metadata of the object itself is protected by an independent encryption mechanism. The AES-KW algorithm is used to wrap only the key-related materials, rather than encrypting the entire metadata.

7. The method according to claim 1, characterized in that, In high-concurrency scenarios, the system dynamically adjusts the number of threads for parallel processing through a performance monitoring mechanism to optimize resource utilization and response speed.

8. The method according to claim 1, characterized in that, The system also includes misuse protection and threshold policies, setting a maximum limit on the amount of encrypted data for each Key ID and automatically rotating the key when the threshold is reached; at the same time, version and object block indexes are included in AAD to protect against replay or rollback attacks.

9. The method according to claim 1, characterized in that, If tag verification fails during the decryption process, the system will interrupt the process, trigger a security alarm, and record a complete security event log for subsequent investigation.

Images