A ciphertext retrieval method for a domestic database

Abstract

The application relates to a ciphertext retrieval method for a localization database, which is based on hierarchical security targets of covering mode leakage and actual operation conditions, and is designed under the guidance of strong backward security to support an index structure with an immediate deletion function and a searchable encryption scheme meeting the strong backward security, and a trusted execution environment is applied to improve the scheme performance. The application effectively supports the immediate deletion of index information and document information, reduces the redundant information on the server, in the scheme design aspect, the trusted execution environment is adopted to optimize the searchable encryption scheme performance, reduce the client storage, and greatly improve the ciphertext retrieval efficiency.

Description

Technical Field

[0001] This invention belongs to the field of information security technology, and relates to searchable encryption technology, and in particular to a method for ciphertext retrieval for domestically produced databases. Background Technology

[0002] Database systems, as crucial computer software, serve as the storers and protectors of information in the era of massive data. With the introduction of information security laws, regulations, and standards, data protection has reached unprecedented levels of importance. Data owners typically encrypt their data and outsource it to cloud-based secure database systems to ensure data security. Secure database systems employ reliable encryption algorithms to store data in encrypted form on cloud servers, representing a vital approach to addressing both massive information storage and privacy protection. Semantic security encryption algorithms guarantee data confidentiality, preventing hackers from directly accessing users' plaintext data by attacking cloud servers. However, operations such as queries that could previously be performed on plaintext data are no longer possible, limiting data availability. Therefore, in secure database systems, ensuring data availability while protecting data privacy is a pressing issue that needs to be addressed.

[0003] Secret database systems typically employ searchable encryption techniques to implement keyword retrieval. While some research exists on searchable encryption techniques, many unresolved security and efficiency issues remain. Leaks of searchable encryption schemes can be characterized by data volume patterns, search patterns, and access behavior patterns. Current definitions of forward and backward security do not fully cover these three leakage patterns, and even if a scheme achieves forward and backward security, it still cannot defend against leak-abuse attacks. Searchable encryption schemes designed using blind storage technology offer high security, but face enormous communication, computational, and client storage costs, rendering them impractical.

[0004] Therefore, there is an urgent need for a practical and efficient solution that can meet the requirements of database keyword query, high security objectives, and lightweight storage of massive amounts of data. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a encrypted retrieval method for domestically produced databases. The design focuses on three aspects: the definition of hierarchical security objectives, the design of an index structure that supports immediate deletion, and strong backward-secure searchable encryption. The aim is to achieve innovative results in security objective definition and index design, providing a practical encrypted keyword retrieval scheme for encrypted database systems. Furthermore, the design and implementation are based on domestically produced database software, further realizing the independent controllability of encrypted database systems.

[0006] The technical problem solved by this invention is achieved through the following technical solution:

[0007] A method for encrypted retrieval of domestically produced databases, based on the hierarchical security objectives of coverage pattern leakage and actual operation, designs an index structure supporting immediate deletion and a searchable encryption scheme that meets strong backward security requirements under the guidance of strong backward security, and applies a trusted execution environment to improve the performance of the scheme; the steps of the method are as follows:

[0008] S1. Formal Definition of Hierarchical Security Objectives

[0009] Based on pattern leakage, we define graded security objectives. By assessing the severity of data volume pattern leakage, search pattern leakage, and access behavior pattern leakage, we refine the security definition of searchable encryption and redefine broad security objectives by drawing on the graded approach.

[0010] Based on actual access behavior, security definitions are refined, and strong backward security is defined by assessing the information leakage caused by immediate deletion operations to searchable encryption schemes.

[0011] S2. Immediate Deletion Index Structure Design

[0012] A hybrid index structure based on forward index and inverted index is proposed, where the inverted index is used for search operations and the forward index is used for update operations;

[0013] A hybrid index fusion strategy based on trapdoor technology is proposed. The inverted index stores the correspondence between keywords and document identifiers, while the forward index stores the correspondence between document identifiers and keyword trapdoors. Through keyword trapdoors, the specific position in the inverted index can be located directly without traversing the entire inverted index, thus providing support for immediate deletion and strong backward security.

[0014] S3, a strongly backward-secure searchable encryption scheme design

[0015] In a cloud computing environment, a trusted execution environment is introduced to reduce the communication cost of the solution;

[0016] Based on the above index structure, a state partitioning strategy and a data and program ambiguity strategy for a trusted execution environment are designed to achieve a practical and searchable encryption scheme with strong backward security. This scheme aims to minimize information leakage while reducing client storage and improving efficiency, thereby achieving the goal of strong backward security.

[0017] Furthermore, the pattern leakage model proposed in S1 reduces the leakage of statistical information such as keyword distribution by padding the data volume pattern leakage. In the searchable encryption scheme that protects the data volume pattern, the added noise is identified. On the one hand, the identification information or specific flag bits are saved on the client side, and on the other hand, the identification information is stored in the encrypted data information. In this way, the padding data can be filtered out on the client side to obtain the correct result.

[0018] Furthermore, S1 proposes a forward query security model to further prevent attackers from knowing whether a search operation on a newly added document matches a previous search operation; it analyzes and summarizes methods to resist adaptive file injection attacks and proposes the concept of forward-opposite security, where attackers cannot know the timestamp of the search result update or its storage location in the database, thus eliminating the threat of injection attacks caused by pattern leakage at the source. Forward-opposite security only leaks the number of keywords contained in the updated document during the update operation, and only leaks the search pattern during the search operation.

[0019] Furthermore, the forward index of S2 retrieves keywords through document identifiers, and the inverted index retrieves all document identifiers containing the keywords. During the update operation, the inverted index needs to traverse all keyword indices; during the search operation, the forward index needs to traverse all document indices. Therefore, the search encryption scheme can employ either a forward index or an inverted index structure, specifically:

[0020] A hybrid index structure combining forward and inverted indexes is designed to support immediate deletion operations. The hybrid index consists of a forward index and an inverted index, wherein the inverted index is used for search operations and the forward index is used for update operations.

[0021] The counter pattern is used to obfuscate the same keywords contained in different documents. That is, a counter is stored for each keyword. If a document contains the keyword, the counter value is incremented. The forward index also uses the counter method to generate the keys in the index. Unlike the inverted index, the counter value here is the number of keywords contained in a document.

[0022] During the add operation, index information is added to both the forward index and the inverted index simultaneously in counter mode, while the status information is updated.

[0023] Furthermore, the immediate deletion based on trapdoor technology in S2 specifically refers to:

[0024] The forward index stores the correspondence between document identifiers and keyword trapdoors, while the inverted index stores the correspondence between keywords and document identifiers. Keywords are the keys of the inverted index. After obtaining the keyword trapdoors stored in the forward index, the index information in the inverted index can be located directly.

[0025] During the deletion operation, a deletion token (the key of the forward index) is generated based on the status information. The value of the forward index is retrieved, and then the key of the inverted index is obtained. The index information and documents are immediately deleted based on the keys of the forward and inverted indexes. This process does not traverse the inverted index, thus achieving the security objective.

[0026] During the query operation, a search token is generated based on the status information. By simply traversing the inverted index, the matching document identifier can be obtained, and the keyword search results can be obtained. To avoid breaking the chain-like inverted index structure due to immediate deletion operations, the client needs to maintain the status information and have the ability to calculate the search token for the entire inverted index.

[0027] Furthermore, the S3 guarantee of backward security for immediate deletion based on state information specifically includes:

[0028] During the insertion operation, the hybrid index structure ensures that index information can be added without traversing the index, meaning that the information of the newly inserted document will not be leaked before it is searched, thus meeting the requirements of forward security.

[0029] To ensure strong backward security, the client caches tokens for all deleted inverted indexes, thereby generating search tokens that match the current query results. This prevents search operations from linking to the indexes of deleted files and their original storage locations, effectively reducing information leakage caused by search operations.

[0030] Furthermore, the S3 trusted execution environment efficiency optimization strategy based on state partitioning and ambiguity technology is specifically as follows:

[0031] The efficiency of the solution can be improved by placing the client of the searchable encryption scheme in a trusted execution environment, that is, by storing the state information that was originally stored on the client in trusted memory;

[0032] The Trusted Execution Environment (TEE) guarantees the confidentiality and integrity of data computations within it and supports remote authentication, allowing remote clients to establish trust within the trusted memory. Clients can establish secure channels with verified trusted memory. The interaction between the client and server is transformed into an interaction between the server and the TEE. Index and token information, which previously required client computation, are now generated by the TEE, resulting in constant client storage and low search / query costs per interaction. However, this approach suffers from several drawbacks. First, it can lead to excessive data storage in the trusted memory, resulting in significant performance overhead once the capacity is exceeded. Second, this method fails to achieve strong backward security; attackers can observe not only the data stored on the server but also changes in the memory bus and memory addresses. Directly placing searchable, encrypted clients within the TEE results in data volume and access behavior pattern leaks, giving attackers a greater advantage and violating the strong backward security objective.

[0033] By employing a block-based processing approach, deleted search tokens are stored in a trusted execution environment (TEA). This reduces client-side storage and avoids additional communication overhead. The TEA, deployed on the server, is responsible for secure computation and relaying communication between the client and the server. Bitwise operations are used to implement the code running within the TEA.

[0034] The advantages and beneficial effects of this invention are as follows:

[0035] This invention addresses the security and practicality of searchable encryption technology, focusing on three aspects: security objectives, immediate deletion, and the construction of practical solutions. Regarding security objectives, it draws inspiration from the design of differential privacy-level security objectives, refining security goals to cover actual user behaviors such as access behavior patterns, search patterns, data volume patterns, and immediate deletion. In terms of index design, it applies trapdoor technology to integrate hybrid indexes, combining the characteristics of forward and inverted indexes to design a hybrid index that effectively supports the immediate deletion of index and document information, reducing redundant information on the server. Finally, in terms of solution design, it employs a trusted execution environment to optimize the performance of the searchable encryption scheme, reducing client-side storage and significantly improving encrypted retrieval efficiency. Attached Figure Description

[0036] Figure 1 This is a technical roadmap of the present invention;

[0037] Figure 2 This is a diagram illustrating the pattern leakage model of the present invention;

[0038] Figure 3 This is a schematic diagram of the forward index and inverted index of the present invention;

[0039] Figure 4 This is a schematic diagram of the hybrid index structure of the present invention;

[0040] Figure 5 This is a diagram of the searchable encryption architecture based on a trusted execution environment according to the present invention. Detailed Implementation

[0041] The present invention will be further described in detail below through specific embodiments. The following embodiments are merely descriptive and not limiting, and should not be used to limit the scope of protection of the present invention.

[0042] An innovative method for encrypted retrieval of domestically produced databases, characterized by: Figure 1 The technical approach adopted in this invention is as follows: Figure 1 As shown, based on the hierarchical security objectives of overlay pattern leakage and actual operation, an index structure supporting immediate deletion and a searchable encryption scheme satisfying strong backward security are designed under the guidance of strong backward security. A trusted execution environment is applied to improve the performance of the scheme.

[0043] (1) Formal definition of graded security objectives

[0044] This invention formally defines hierarchical security objectives in two steps:

[0045] The first step is to define graded security objectives based on pattern leakage. By assessing the severity of data volume pattern leakage, search pattern leakage, and access behavior pattern leakage, the security definition of searchable encryption is further refined, and a broad range of security objectives is redefined by drawing on the graded approach.

[0046] The second step is to refine the security definition based on actual access behavior, and further define strong backward security by assessing the information leakage caused by the immediate deletion operation to the searchable encryption scheme.

[0047] This invention addresses the shortcomings of existing forward-secure and backward-secure searchable encryption definitions, which fail to cover real-world scenarios such as immediate deletion and deletion after query, as well as the unclear hierarchical definitions. It proposes a pattern leakage model, such as... Figure 2 As shown.

[0048] Searchable encryption employs semantically secure encryption algorithms, ensuring that ciphertext does not directly reveal plaintext information. However, access to indexes and documents during update and search operations can lead to information leaks. These leaks are characterized as access behavior pattern leaks, search pattern leaks, and data volume pattern leaks. Current forward and backward security definitions only formally define some search pattern leaks and cannot effectively limit access behavior pattern and data volume pattern leaks. Data volume pattern leaks are often accompanied by the leakage of statistical information, leading to inference attacks.

[0049] The leak pattern involves the following two steps:

[0050] The first step of this invention is to reduce the leakage of statistical information such as keyword distribution by filling in the gaps in the data volume pattern.

[0051] The second step is to identify the added noise in the searchable encryption scheme of the data volume protection mode. On the one hand, the identification information or specific flag bits are saved on the client, and on the other hand, the identification information is stored in the encrypted data information. In this way, the padding data can be filtered out on the client to obtain the correct result.

[0052] Current forward security only requires that new insert operations cannot be linked to previous search operations before a search operation is performed. Once a search operation is performed, the new insert operation is exposed. This makes forward-secure encrypted retrieval schemes only resistant to adaptive file injection attacks, but not to statistical inference attacks. If new update operations cannot be linked after a search operation, attackers will find it difficult to infer the keywords being searched, thus hindering defense against statistical attacks.

[0053] Forward safety design involves the following two steps:

[0054] First, based on traditional forward security, this invention proposes a forward query security model to further prevent attackers from knowing whether a search operation on a newly added document matches a previous search operation.

[0055] The second step, since forward search security only leaks the update time of search results during search operations, limiting the leakage of search patterns, this invention analyzes and summarizes methods for resisting adaptive document injection attacks and proposes the concept of forward-opposite security. Attackers cannot know the timestamp of search result updates or their storage location in the database, thus eliminating the threat of injection attacks caused by pattern leakage at the root. Forward-opposite security only leaks the number of keywords contained in the updated document during update operations, and only leaks the search pattern during search operations.

[0056] Current backward security requirements stipulate that document insertion and deletion operations must occur between two search operations, and subsequent search operations should not reveal the document's identifier. The backward security definition assumes that the document's index and document information are not immediately deleted, failing to consider the consequences of the server obtaining the document identifier during immediate deletion. If a document is immediately deleted and its storage space is released, the physical address of the document's storage space will be exposed to an untrusted server, allowing an attacker to obtain the deleted document's identifier. This invention proposes immediate deletion backward security, which only reveals the timestamp of search results being updated to the database. Immediate deletion backward security is stronger than forward security but weaker than level one backward security.

[0057] (2) Index structure design that supports immediate deletion

[0058] To address the current situation where logical deletion or punched encryption technologies do not support immediate deletion and cannot achieve immediate deletion by modifying the searchable encrypted index structure, the solution involves the following two steps:

[0059] The first step is index structure design, proposing a hybrid index structure based on forward index and inverted index, where the inverted index is used for search operations and the forward index is used for update operations.

[0060] The second step proposes a hybrid index fusion strategy based on trapdoor technology. The inverted index stores the correspondence between keywords and document identifiers, while the forward index stores the correspondence between document identifiers and keyword trapdoors. Through keyword trapdoors, the specific location within the inverted index can be directly located without traversing the entire inverted index, thus supporting immediate deletion and strong backward safety.

[0061] Forward indexes retrieve keywords by document identifiers, while inverted indexes retrieve all document identifiers containing the keyword, such as... Figure 3 As shown. In update operations, the inverted index needs to traverse all keyword indexes; in search operations, the forward index needs to traverse all document indexes. For efficiency reasons, searchable encryption schemes can use either forward or inverted index structures.

[0062] The design of a hybrid-structured index involves the following three steps:

[0063] The first step of this invention addresses the requirement for a highly secure, searchable, and encrypted scheme by designing a hybrid index structure that combines the advantages of forward and inverted indexes to support immediate deletion operations. The index consists of a forward index and an inverted index, where the inverted index is used for search operations and the forward index is used for update operations.

[0064] The second step of this invention is to use a counter model to obfuscate the same keywords contained in different documents. Specifically, a counter is stored for each keyword. If a document contains the keyword, the counter value is incremented ("keyword || counter value" becomes the new keyword). The forward index also uses a counter method to generate the keys in the index. Unlike the inverted index, the counter value here is the number of keywords contained in a document.

[0065] The third step involves adding index information to both the forward index and the inverted index simultaneously in counter mode, while also updating the status information.

[0066] The strategy of achieving immediate deletion based on the trapdoor technique consists of the following three steps:

[0067] The first step involves storing the mapping between document identifiers and keyword trapdoors in the forward index, and the mapping between keywords and document identifiers in the inverted index. Keywords are the keys of the inverted index; therefore, after obtaining the keyword trapdoors stored in the forward index, the index information in the inverted index can be directly located.

[0068] The second step, during the deletion operation, involves generating a deletion token (the key of the forward index) based on the status information, retrieving the value of the forward index, and then obtaining the key of the inverted index. The index information and documents are immediately deleted based on the keys of both the forward and inverted indexes. This process avoids traversing the inverted index, thus achieving a security goal. The logical process is as follows: Figure 4 As shown.

[0069] The third step, during the query operation, involves generating a search token based on the status information. By simply traversing the inverted index, matching document identifiers can be obtained, leading to the keyword search results. To prevent immediate deletion operations from breaking the chain-like inverted index structure, the client needs to maintain the status information and have the ability to calculate the search token for the entire inverted index.

[0070] (3) Design of a strongly backward-secure searchable encryption scheme

[0071] To address the current situation where robust backward-secure searchable encryption schemes based on blind storage technology suffer from low efficiency and a lack of practical research on robust backward-secure searchable encryption, this paper proposes a two-step approach to solve this problem:

[0072] The first step is to introduce a trusted execution environment in the cloud computing environment to reduce the communication cost of the solution.

[0073] The second step, based on the above index structure, is to design a state partitioning strategy and a data and program ambiguity strategy for the trusted execution environment, thereby implementing a practical and searchable encryption scheme with strong backward security. This scheme aims to minimize information leakage while reducing client storage and improving efficiency, thus achieving the goal of strong backward security.

[0074] Ensuring backward security through immediate deletion based on state information involves the following two steps:

[0075] The first step is that during the insertion operation, the index structure ensures that index information can be added without traversing the index, meaning that the information of the newly inserted document will not be leaked before it is searched, thus meeting the requirements of forward security.

[0076] The second step, to ensure strong backward security, is to cache tokens for all deleted inverted indexes on the client side, thereby generating search tokens that match the current query results. This prevents search operations from linking to the indexes of deleted files and their original storage locations, thus directly reducing information leakage caused by search operations.

[0077] The efficiency optimization strategy for trusted execution environments based on state partitioning and blind spot techniques consists of the following three steps:

[0078] The first step is to improve the efficiency of the solution by placing the client with searchable encryption schemes in a trusted execution environment, that is, by storing the state information originally stored on the client in trusted memory.

[0079] The second step involves a Trusted Execution Environment (TEE) that ensures the confidentiality and integrity of data computations within it and supports remote authentication. This allows remote clients to establish trust within the trusted memory, enabling them to establish secure channels with verified trusted memory. The interaction between the client and server is transformed into an interaction between the server and the TEE. The index and token information, which previously required client computation, is now generated by the TEE, resulting in constant client storage and low search / query costs per interaction. However, this approach has two drawbacks. First, it can lead to excessive data storage in the trusted memory, resulting in significant performance overhead once the capacity is exceeded. Second, this method fails to achieve strong backward security. Attackers can observe not only the data stored on the server but also changes in the memory bus and memory addresses. Directly placing searchable, encrypted clients within the TEE results in data volume and access behavior pattern leaks, giving attackers a greater advantage and violating the strong backward security objective.

[0080] Thirdly, this invention employs a block-based processing approach, storing the deleted search tokens in a trusted execution environment. This reduces client-side storage without incurring additional communication overhead. Figure 5 As shown, the Trusted Execution Environment (TEE) deployed on the server is responsible for secure computation and relaying communication between the client and the server. In addition, bitwise operations are used to implement the code running internally within the TEE.

[0081] Although embodiments and drawings of the present invention have been disclosed for illustrative purposes, those skilled in the art will understand that various substitutions, variations and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the scope of the present invention is not limited to the contents disclosed in the embodiments and drawings.

Claims

1. A method for encrypted text retrieval in domestically produced databases, characterized in that: Based on the hierarchical security objectives of overlay pattern leakage and actual operation, and guided by strong backward security, an index structure supporting immediate deletion and a searchable encryption scheme satisfying strong backward security are designed. A trusted execution environment is applied to improve the scheme's performance. The steps of the method are as follows: S1. Formal Definition of Hierarchical Security Objectives Based on pattern leakage, we define graded security objectives. By assessing the severity of data volume pattern leakage, search pattern leakage, and access behavior pattern leakage, we refine the security definition of searchable encryption and redefine broad security objectives by drawing on the graded approach. Based on actual access behavior, we refine the security definition and define strong backward security by assessing the information leakage caused by the immediate deletion operation to the searchable encryption scheme. S2. Immediate Deletion Index Structure Design A hybrid index structure based on forward index and inverted index is proposed, where the inverted index is used for search operations and the forward index is used for update operations; A hybrid index fusion strategy based on trapdoor technology is proposed. The inverted index stores the correspondence between keywords and document identifiers, while the forward index stores the correspondence between document identifiers and keyword trapdoors. Through keyword trapdoors, the specific position in the inverted index can be located directly without traversing the entire inverted index, thus providing support for immediate deletion and strong backward security. S3, a strongly backward-secure searchable encryption scheme design In a cloud computing environment, a trusted execution environment is introduced to reduce the communication cost of the solution; Based on the above index structure, a state partitioning strategy and a data and program ambiguity strategy for a trusted execution environment are designed to achieve a practical and searchable encryption scheme with strong backward security. This scheme aims to minimize information leakage while reducing client storage and improving efficiency, thereby achieving the goal of strong backward security. The forward index of S2 retrieves keywords using document identifiers, while the inverted index retrieves all document identifiers containing the keywords. During the update operation, the inverted index needs to traverse all keyword indexes; during the search operation, the forward index needs to traverse all document indexes. The search encryption scheme can employ either a forward index or an inverted index structure, specifically: A hybrid index structure combining forward and inverted indexes is designed to support immediate deletion operations. The hybrid index consists of a forward index and an inverted index, wherein the inverted index is used for search operations and the forward index is used for update operations. The counter pattern is used to obfuscate the same keywords contained in different documents. That is, a counter is stored for each keyword. If a document contains the keyword, the counter value is incremented. The forward index also uses the counter method to generate the keys in the index. Unlike the inverted index, the counter value here is the number of keywords contained in a document. During the add operation, index information is added to both the forward index and the inverted index simultaneously in counter mode, while the status information is updated at the same time; The immediate deletion based on trapdoor technology in S2 specifically refers to: The forward index stores the correspondence between document identifiers and keyword trapdoors, while the inverted index stores the correspondence between keywords and document identifiers. Keywords are the keys of the inverted index. After obtaining the keyword trapdoors stored in the forward index, the index information in the inverted index can be located directly. During the deletion operation, a deletion token is generated based on the status information. This token is used to retrieve the value of the forward index, which in turn yields the key of the inverted index. The index information and documents are then immediately deleted based on both the forward and inverted index keys. This process avoids traversing the inverted index, thus achieving the security objective. During the query operation, a search token is generated based on the status information. By simply traversing the inverted index, the matching document identifier can be obtained, and the keyword search results can be obtained. To avoid breaking the chain-like inverted index structure due to immediate deletion operations, the client needs to maintain the status information and have the ability to calculate the search token for the entire inverted index.

2. The encrypted retrieval method for domestically produced databases according to claim 1, characterized in that: The proposed pattern leakage model in S1 addresses data volume pattern leakage by reducing the leakage of keyword distribution statistics through padding. In the searchable encryption scheme that protects data volume patterns, added noise is identified. On the one hand, the identification information or specific flag bits are saved on the client side, and on the other hand, the identification information is stored in the encrypted data information. In this way, the padding data can be filtered out on the client side to obtain the correct result.

3. The encrypted retrieval method for domestically produced databases according to claim 1, characterized in that: The S1 section proposes a forward query security model to further prevent attackers from knowing whether a search operation on a newly added document matches a previous search operation. It analyzes and summarizes methods to resist adaptive file injection attacks and proposes the concept of forward-opposite security. Attackers cannot know the timestamp of the search result update or its storage location in the database, thus eliminating the threat of injection attacks caused by pattern leakage at the root. Forward-opposite security only leaks the number of keywords contained in the updated document during update operations, and only leaks the search pattern during search operations.

4. The encrypted retrieval method for domestically produced databases according to claim 1, characterized in that: The S3 guarantee of backward security for immediate deletion based on state information specifically means: During the insertion operation, the hybrid index structure ensures that index information can be added without traversing the index, meaning that the information of the newly inserted document will not be leaked before it is searched, thus meeting the requirements of forward security. To ensure strong backward security, the client caches tokens for all deleted inverted indexes, thereby generating search tokens that match the current query results. This prevents search operations from linking to the indexes of deleted files and their original storage locations, effectively reducing information leakage caused by search operations.

5. The encrypted retrieval method for domestically produced databases according to claim 1, characterized in that: The S3 trusted execution environment efficiency optimization strategy based on state partitioning and blind spot technology is as follows: The efficiency of the solution can be improved by placing the client of the searchable encryption scheme in a trusted execution environment, that is, by storing the state information that was originally stored on the client in trusted memory; Trusted execution environments guarantee the confidentiality and integrity of data computations within them and support remote authentication, allowing remote clients to establish trust in trusted memory. Clients can establish secure channels with verified trusted memory. The interaction between the client and server is transformed into an interaction between the server and the trusted execution environment. The index and token information that the client originally needed to calculate and generate are now calculated and generated by the trusted execution environment. This results in constant client storage and search query costs per interaction. On the one hand, this leads to excessive data storage in trusted memory, which will cause huge performance overhead once the trusted memory capacity is exceeded. On the other hand, this method cannot achieve strong backward security goals. Attackers can observe not only the data stored on the server, but also changes in the memory bus and memory addresses. Directly placing the searchable encrypted client in the trusted execution environment leads to the leakage of data volume patterns and access behavior patterns, which gives attackers more attack advantages and violates the goal of strong backward security. By employing a block-based processing approach, deleted search tokens are stored in a trusted execution environment (TEA). This reduces client-side storage and avoids additional communication overhead. The TEA, deployed on the server, is responsible for secure computation and relaying communication between the client and the server. Bitwise operations are used to implement the code running within the TEA.

Images