A method and system for retrieval of privacy information in multiple files against collusion

Abstract

The present invention proposes a collusion-resistant multi-file privacy information retrieval method and system, which relates to the field of privacy information retrieval, including: distributing and storing a number of files on a plurality of servers in a distributed manner through file packaging and coding storage to form a distributed storage system; obtaining retrieval information of a required file, selecting a query round together with storage parameters, and constructing a plurality of alternative retrieval schemes; selecting an optimal retrieval scheme from a plurality of alternative retrieval schemes according to a PIR code rate of the retrieval scheme; obtaining the content of the required file by executing the optimal retrieval scheme; based on marginal information query and demand information query, the present invention efficiently realizes multi-file privacy information retrieval that can resist collusion of any number of servers in a distributed storage system based on an arbitrary code.

Description

Technical Field

[0001] The present invention belongs to the field of privacy information retrieval, and in particular relates to a method and system for retrieval of multi-file privacy information against collusion. Background Art

[0002] The rapid development of big data and cloud computing technologies, while providing great convenience for people to obtain information, also brings the risk of user privacy information leakage; for example, most current search engines will analyze users' search behavior and make corresponding personalized customization or advertising promotions; when searching some sensitive databases, users' personal privacy needs to be protected; private information retrieval (PIR) allows users to send queries to the server to retrieve the required files, while ensuring that the server cannot obtain any clues to the retrieved file numbers.

[0003] The PIR problem originated in the field of cryptography and is closely related to research directions such as oblivious transfer, privacy computing, and secure multi-party computing. Its main optimization indicators are communication volume (upload and download required to retrieve a single bit) and algorithm complexity. In their groundbreaking work in 2017, Sun and Jafar extended the PIR problem in the field of cryptography to distributed storage systems, and the retrieval target was also derived from a single bit to a sufficiently large file.

[0004] The general description of the PIR problem in distributed systems is as follows: In the distributed storage system of the code, The files are encoded and stored in In a server, the user hopes to successfully retrieve a single required file while ensuring that any The colluding servers cannot obtain any information about the retrieved file number; the efficiency measure of the PIR scheme of the distributed storage system is called the PIR code rate, which is the ratio between the size of the required file and the total download amount.

[0005] In the past decade, the PIR problem of distributed systems has been one of the hot issues in the forefront of information science, and nearly 100 articles on related topics have been published in the flagship journal of international information theory, IEEE Transactions on Information Theory. However, the vast majority of studies only consider single-file retrieval. In terms of multi-file PIR retrieval, Banawan and Ulukus pointed out in their 2018 work that the efficiency of retrieving multiple files simultaneously is strictly better than that of repeatedly executing a single-file PIR scheme, which shows that the research on multi-file PIR retrieval schemes has practical application value. In the work of Banawan and Ulukus, if the number of retrieved files is Strictly less than , the PIR code rate of their scheme is still far behind that of the information theory community. At the same time, their work only targets This degenerate case (i.e., replicated storage system, no collusion between servers) does not take into account the general parameters and .

[0006] Therefore, existing multi-file privacy information retrieval schemes have limitations. Summary of the invention

[0007] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a method and system for anti-collusion multi-file privacy information retrieval, which is based on marginal information query and demand information query, and efficiently realizes the Distributed storage system with arbitrarily resistant Retrieval of private information from multiple files using collusion between servers.

[0008] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0009] A first aspect of the present invention provides a method for retrieval of multi-file privacy information against conspiracy.

[0010] A method for retrieval of multi-file privacy information against conspiracy, comprising:

[0011] By means of file subpackaging and encoding storage, several files are distributed and stored on multiple servers to form a distributed storage system;

[0012] Obtain the search information of the required documents, together with the storage parameters, select the query round, and build multiple alternative search plans;

[0013] According to the PIR code rate of the retrieval scheme, the best retrieval scheme is selected from multiple alternative retrieval schemes;

[0014] Obtain the content of the required document by executing the optimal retrieval solution;

[0015] The retrieval scheme consists of multiple rounds of demand information query and marginal information query and one round of only demand information query. The results of marginal information query are used to eliminate non-demand files in the demand information query results of subsequent rounds to obtain the content of the demand files.

[0016] Furthermore, the file is divided into The original data packets are divided into two parts, each of which is transmitted through Code, generate The encoded data packets are stored in in a server.

[0017] Furthermore, the storage parameters include the number of segments , Number of servers , Total number of files ;

[0018] The search information of the required files, including the number of required files , the number of each requirement file and the number of colluding servers .

[0019] Further, the retrieval scheme includes several rounds, each round includes several stages, and in each stage, the user is required to send a query request to each server, and the server returns a query response;

[0020] The inquiry request is based on The form of the sum of the search symbols of the files, batch request these The file is stored in a file package on the server. is the round number;

[0021] The query response is the server returning the file data packet pointed to by the search symbol.

[0022] Furthermore, the query round is selected from In each round, select rounds, among which The first round includes demand information query and marginal information query. Each round only includes demand information queries;

[0023] The marginal information query is a query that only contains the sum of search symbols of non-required documents;

[0024] The demand information query is a query that includes the sum of search symbols of the demand documents.

[0025] Furthermore, the construction of the alternative search scheme also includes: calculating the scheme parameters based on the storage parameters and the search parameters. and ;

[0026] Using solution parameters and , perform the following calculations respectively:

[0027] (1) Calculate the selected The number of stages in each round;

[0028] (2) Using solution parameters and , set the search symbol for all files;

[0029] (3) Utilizing solution parameters and , calculate the PIR code rate of each scheme.

[0030] Furthermore, the setting of the search symbol is performed by preselecting The encoding matrix encodes the data packet of each file to obtain the retrieval symbol of the file.

[0031] A second aspect of the present invention provides a collusion-resistant multi-file privacy information retrieval system.

[0032] A collusion-resistant multi-file privacy information retrieval system includes a file storage module, a solution construction module, a solution selection module and a retrieval execution module:

[0033] The file storage module is configured to: distribute and store a number of files on multiple servers by means of file subpackaging and encoding storage, thereby forming a distributed storage system;

[0034] The solution building module is configured to: obtain the search information of the required document, together with the storage parameters, select the query round, and build multiple alternative search solutions;

[0035] The scheme selection module is configured to: select the best retrieval scheme from multiple candidate retrieval schemes according to the PIR code rate of the retrieval scheme;

[0036] The retrieval execution module is configured to: obtain the content of the required document by executing the optimal retrieval scheme;

[0037] The retrieval scheme consists of multiple rounds of demand information query and marginal information query and one round of only demand information query. The results of marginal information query are used to eliminate non-demand files in the demand information query results of subsequent rounds to obtain the content of the demand files.

[0038] The third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the anti-collusion multi-file privacy information retrieval method as described in the first aspect of the present invention.

[0039] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the anti-collusion multi-file privacy information retrieval method as described in the first aspect of the present invention are implemented.

[0040] One or more of the above technical solutions have the following beneficial effects:

[0041] 1) The present invention constructs a retrieval scheme consisting of multiple rounds of demand information query and marginal information query and one round of demand information query only, and uses the results of marginal information query to eliminate non-demand files in the demand information query results of subsequent rounds to obtain the content of the demand file, thereby achieving the goal of A distributed storage system with cipher codes that can resist arbitrary A multi-file private information retrieval scheme with server collusion can be used for any suitable parameters and , which allows the system to The replication storage is generalized to any and coded storage and allows server collusion.

[0042] 2) In the previous work of Banawan and Ulukus, It's your turn The default selection in round The present invention takes into account the differences and similarities of the selections in each round, and can efficiently calculate the PIR code rate of the solution that can be obtained by selecting each round, so as to give a given parameter , , , , , the optimal round Q can be selected. For example, , , , When , Banawan and Ulukus choose the fifth round by default, and the PIR code rate of the scheme is ; The present invention selects the fourth round, and the PIR code rate is , so it can be seen that the retrieval efficiency of this scheme has been greatly improved.

[0043] 3) The retrieval efficiency of the present invention is much better than that of a multiple-repeated single-file retrieval solution, and thus the retrieval efficiency can be effectively improved in an actual multi-file retrieval process.

[0044] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0046] Figure 1This is a flow chart of the method of the first embodiment.

[0047] Figure 2 The diagram is a schematic diagram of the interactions between the parties involved in the first embodiment.

[0048] Figure 3 This is an example diagram of the retrieval scheme of the first embodiment. DETAILED DESCRIPTION

[0049] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0050] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0051] Embodiment 1

[0052] In one embodiment of the present disclosure, a method for retrieval of multi-file privacy information against collusion is provided. Figure 1 As shown, the following steps are included:

[0053] Step S1: by means of file packaging and encoding storage, a number of files are distributed and stored on multiple servers to form a distributed storage system;

[0054] Furthermore, the file is divided into The original data packets are divided into two parts, each of which is transmitted through Code, generate The encoded data packets are stored in in a server.

[0055] Step S2: Obtain the search information of the required document, together with the storage parameters, select the query round, and construct multiple alternative search solutions;

[0056] Furthermore, the storage parameters include the number of segments , Number of servers , Total number of files ;

[0057] The search information of the required files, including the number of required files , the number of each requirement file and the number of colluding servers .

[0058] Further, the retrieval scheme includes several rounds, each round includes several stages, and in each stage, the user is required to send a query request to each server, and the server returns a query response;

[0059] The inquiry request is based on The form of the sum of the search symbols of the files, batch request these The file is stored in a file package on the server. is the round number;

[0060] The query response is the server returning the file data packet pointed to by the search symbol.

[0061] Furthermore, the query round is selected from In each round, select rounds, among which The first round includes demand information query and marginal information query. Each round only includes demand information queries;

[0062] The marginal information query is a query that only contains the sum of search symbols of non-required documents;

[0063] The demand information query is a query that includes the sum of search symbols of the demand documents.

[0064] Furthermore, the construction of the alternative search scheme also includes: calculating the scheme parameters based on the storage parameters and the search parameters. and ;

[0065] Using solution parameters and , perform the following calculations respectively:

[0066] (1) Calculate the selected The number of stages in each round;

[0067] (2) Using solution parameters and , set the search symbol for all files;

[0068] (3) Utilizing solution parameters and , calculate the PIR code rate of each scheme.

[0069] Furthermore, the setting of the search symbol is performed by preselecting The encoding matrix encodes the data packet of each file to obtain the retrieval symbol of the file.

[0070] Step S3: selecting the best retrieval scheme from multiple candidate retrieval schemes according to the PIR code rate of the retrieval scheme;

[0071] Step S4: Obtain the content of the required document by executing the optimal retrieval scheme;

[0072] The retrieval scheme consists of multiple rounds of demand information query and marginal information query and one round of only demand information query. The results of marginal information query are used to eliminate non-demand files in the demand information query results of subsequent rounds to obtain the content of the demand files.

[0073] In the above method, the total The files are stored in the system through encoding, and each file is cut into The original data packets are transmitted through the same Code Generation The encoded data packets are stored in the distributed storage system There is no mixed encoding between files on the same server.

[0074] The number of files that the user needs to retrieve (i.e., demand files) is , and there is ;like Figure 2 As shown, the user generates several query requests and sends them to various servers. The servers respond to the user's queries. The user integrates all the feedback response results and decodes them to obtain the file content to be retrieved.

[0075] In this process, any The PIR scheme requires that users will not send any requests to these servers as long as they can perform the search correctly. A colluding server leaks any information about the requested file number; depending on the specific parameters and , the user selects different query request construction strategies and constructs different retrieval schemes, specifically:

[0076] (1) When , the user's query request construction strategy is as follows:

[0077] 1) Set the search symbol of the file: random permutations, respectively The coordinates of each file are shuffled, and the replaced coordinates are used as the retrieval symbols of the file.

[0078] 2) The search plan is divided into several "rounds", each of which consists of several "stages"; Rounds ( ) in a certain stage, the specific form of the user's retrieval to each server is Any in the file The sum of a certain coordinate of the files is Search.

[0079] 3) The number of stages in each round must follow specific rules - in the It's your turn In the rounds, the user only selects one of the rounds (referred to as round), the It's your turn The number of stages in the remaining rounds in the round is set to 0; the selected round , and before The rounds together constitute a complete query solution.

[0080] The number of stages in these rounds is non-zero, and their specific design follows a certain ratio. Their essential purpose is to make full use of the marginal information brought by non-demand files to improve the PIR code rate. The calculation details can be found in the "Definition and Setting of Rounds and Stages" section below.

[0081] Generally, The selection of has a certain impact on the PIR bit rate. , , Afterwards, for each Calculate the bitrate one by one and select the best one design solutions; in particular, Divisibility hour, The selection of can be arbitrary.

[0082] (2) For general parameters , , the user's query request construction strategy is as follows:

[0083] 1) Select a specific Coding matrix, performs coding preprocessing on each file and generates the retrieval symbol of the file.

[0084] 2) In the Rounds ( ), the user's retrieval form to each server is Any in the file The sum of the symbols of the files after being encoded in the first step, totaling Search.

[0085] 3) The specific design of the ratio of the number of stages in each round is subject to , , , , impact.

[0086] In a specific search, the encoded symbols of each file involved must meet the following settings: the symbols of the files that the user needs to search for are linearly independent, and the combination of non-searched files in the sum of each search must be calculated through the data that has been searched on other servers; at the same time, in order to ensure the privacy of the solution, On a colluding server, the search symbols for any file must be linearly independent. For specific design details, see the "Setting of Search Symbols for Each File" section below.

[0087] The following describes in detail the implementation process of a method for retrieval of privacy information of multiple files that is resistant to collusion in this embodiment from the aspects of construction of a distributed storage system, definition and setting of "rounds and stages", setting of various file retrieval symbols, an overall solution example and a PIR code rate calculation formula.

[0088] 1. Construction of distributed storage system

[0089] Considered by A distributed storage system consisting of 10 servers. files, each via the same The code is stored in the following form:

[0090] Files: ,in, , that is, the length of each file is , expressed as a size The matrix of .

[0091] Code generator matrix : .

[0092] File The matrix represents the Row, as a data packet for cutting, recorded as a row vector ;use The generator matrix of the code is The row vector of Encoded as length Vector , No. The servers store the encoded vector ; Then traverse all the rows of the matrix representation of all files, such as Figure 2 As shown, Content stored on the server are all row vectors in The projection on

[0093]

[0094] Users need to search in this system files (i.e., requirement files), where , and ensure that any A colluding server cannot obtain any information about the retrieved file number.

[0095] II. Definition and Setting of “Rounds and Stages”

[0096] The structure of a search scheme includes several rounds, each round includes several stages, and in each stage, the user sends a query request to each server, and the specific form of the query request is the sum of the search symbols of several files.

[0097] No. Rounds ( ) indicates that the search is done by in the file The sum of the search symbols of different files; A phase of a round is defined as a Any file The module with the sum of the search symbols of files, totaling Search.

[0098] Among them, in front In round, Rounds ( ) is a stage in which A query consisting of only the sum of search symbols for non-required documents is called a marginal information query, and A query that contains the sum of the symbols of the required files is called a demand information query. To In round, Rounds ( ) is only one stage Request information query.

[0099] The round selection setting of the alternative search plan follows the following rules:

[0100] User in To Select one of the rounds, To The remaining rounds in the round are set to 0, that is, no actual search or query is performed; the selected rounds are compared with the previous The total number of rounds constitutes a complete search plan. Alternative search options.

[0101] The retrieval scheme adjusts the number of stages in each round to fully utilize the marginal information query of non-required documents and eliminate the relevant interference in the subsequent rounds of required information query. arrive No marginal information query is generated in any round, and any round can make full use of all marginal information queries by adjusting the number of stages. Therefore, given the retrieval information and storage parameters, we calculate them one by one. After the PIR code rates of the alternatives are calculated, only the arrive The round that makes the PIR code rate of the scheme the highest, and the previous The optimal search plan is composed of rounds.

[0102] The number of stages of the search plan is set according to the following rules:

[0103] make and To satisfy the following equation, the scheme parameters are set in this way: and The value of

[0104]

[0105] In an alternative search scenario, suppose the user selects wheel( ) Participate in the formation of alternative search plans, the number of rounds of the plan The calculation is as follows:

[0106]

[0107] when At the time of the plan The number of rounds is .when At that time, that is, the plan It's your turn Except the first The number of stages outside the round is set to 0. At that time, that is, the plan It's your turn The number of stages in a round is set to .

[0108] No. The number of stages in a round is set to To ensure that the number of stages in each round of the program is an integer; The number of stages in a round is determined by The number of stages of the round is recursively obtained. A phase of the round requires Wheel provided The marginal information query of the first stage is used to eliminate the interference in the demand information query. The number of stages in a round is:

[0109]

[0110] By analogy, for the wheel( ), the number of stages is:

[0111]

[0112] 3. Setting of each file search symbol

[0113] By pre-selected The encoding matrix encodes the data packet of each file to obtain the retrieval symbol of the file, which is specifically:

[0114] Assume that in a sufficiently large domain There is The transpose of the code generator matrix is ​​recorded as ;make Indicates that the user All on Uniformly and independently selected in a full rank matrix A random matrix.

[0115] In general parameters and Next, In the round of phases, one The specific form of the query of the sum of the search symbols of the files is set to be isomorphic to an auxiliary matrix. The auxiliary matrix is ​​constructed by The size of the symbols is A matrix where each symbol appears in the auxiliary matrix times, and any columns share a common sign, that is, any The servers have an identical query.

[0116] The search symbol settings for non-required files are as follows:

[0117] Through a specific MDS coding matrix ,Will Non-required files The retrieval symbol is pre-coded as ,in From the matrix Selected from Ensure that the entire program The selected rows are not repeated.

[0118] For the required documents , whose search symbol is set to .

[0119] In particular, for In the degenerate case, the file retrieval symbol is set as follows:

[0120] The user shuffles the coordinates of each file through M random permutations and uses the coordinates directly as retrieval symbols.

[0121] 4. Overall Solution Example

[0122] , , , When the 5th round is selected as Wheel, with front The search scheme consisting of rounds is as follows Figure 3 As shown, for the retrieval scheme round, in each stage, the user retrieves a total of Any To simplify the description, the query is as follows: express The query form of each file is a query module that is isomorphic to the auxiliary matrix and represents a query request.

[0123] consider a requirement information query containing both required and non-required documents, and A marginal information query containing only the corresponding non-required files, each non-required file The search symbols are obtained by the pre-coding method described in 3); only one required file is included in the demand information query, and the required file The search symbol from In the requirement information query containing two or more required files, only the symbol of one required file is used as the symbol to be searched for in the query, and the remaining required files are set as interference in the query using the same precoding method as the non-required files, and are set as search symbols that can be determined from the search of other servers. From the perspective of a colluding server, the search symbols for arbitrary files are linearly independent.

[0124] 5. PIR bit rate calculation formula

[0125] The PIR rate of the retrieval solution refers to the ratio between the size of the required file and the total download amount, which is calculated by the following iterative method:

[0126] Record The total downloads in one round are , For the The sum of the downloads of the round itself and the downloads of the required previous rounds, the size of the user's retrieval file is , For the The sum of the size of the retrieval file in the round itself and the size of the retrieval file downloaded from other servers in the previous round. The round code rate is expressed as , with the Round code rate is the initial value, for the previous The iterative calculation formula can be written as:

[0127]

[0128] The code rate of the retrieval alternative is arrive The code rate of the round stage is calculated as follows:

[0129]

[0130] choose arrive The round phase with the highest code rate is The best search solution among the alternative search solutions.

[0131] Embodiment 2

[0132] In one embodiment of the present disclosure, a multi-file privacy information retrieval system is provided to resist collusion, including a file storage module, a solution construction module, a solution selection module and a retrieval execution module:

[0133] The file storage module is configured to: distribute and store a number of files on multiple servers by means of file subpackaging and encoding storage, thereby forming a distributed storage system;

[0134] The solution building module is configured to: obtain the search information of the required document, together with the storage parameters, select the query round, and build multiple alternative search solutions;

[0135] The scheme selection module is configured to: select the best retrieval scheme from multiple candidate retrieval schemes according to the PIR code rate of the retrieval scheme;

[0136] The retrieval execution module is configured to: obtain the content of the required document by executing the optimal retrieval scheme;

[0137] The retrieval scheme consists of multiple rounds of demand information query and marginal information query and one round of only demand information query. The results of marginal information query are used to eliminate non-demand files in the demand information query results of subsequent rounds to obtain the content of the demand files.

[0138] Embodiment 3

[0139] The purpose of this embodiment is to provide a computer-readable storage medium.

[0140] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in the anti-collusion multi-file privacy information retrieval method as described in the first embodiment of the present disclosure.

[0141] Embodiment 4

[0142] The purpose of this embodiment is to provide an electronic device.

[0143] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps in the anti-collusion multi-file privacy information retrieval method as described in the first embodiment of the present disclosure are implemented.

[0144] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for retrieval of multi-file privacy information against collusion, characterized in that: include: By means of file subpackaging and encoding storage, several files are distributed and stored on multiple servers to form a distributed storage system; Obtain the search information of the required documents, together with the storage parameters, select the query round, and build multiple alternative search plans; According to the PIR code rate of the retrieval scheme, the best retrieval scheme is selected from multiple alternative retrieval schemes; Obtain the content of the required document by executing the optimal retrieval solution; The retrieval scheme consists of multiple rounds of demand information query and marginal information query and one round of demand information query only, and the results of marginal information query are used to eliminate non-demand documents in the demand information query results of subsequent rounds to obtain the content of the demand documents; The query round is selected from In each round, select rounds, among which The first round includes demand information query and marginal information query. The round only includes demand information query, is the total number of files, is the number of required documents; The marginal information query is a query that only contains the sum of search symbols of non-required documents; The demand information query is a query that includes the sum of search symbols of the demand documents.

2. The method for retrieval of multi-file privacy information against collusion as claimed in claim 1, characterized in that: The file is divided into The original data packets are divided into two parts, each of which is transmitted through Code, generate The encoded data packets are stored in in a server.

3. The method for retrieval of multi-file privacy information against collusion as claimed in claim 1, characterized in that: The storage parameters include the number of segments , Number of servers , Total number of files ; The search information of the required files, including the number of required files , the number of each requirement file and the number of colluding servers .

4. The method for retrieval of multi-file privacy information against collusion as claimed in claim 3, characterized in that: The retrieval scheme includes several rounds, each round includes several stages, in each stage, the user is required to send a query request to each server, and the server returns a query response; The inquiry request is based on The form of the sum of the search symbols of the files, batch request these The file is stored in a file package on the server. is the round number; The query response is the server returning the file data packet pointed to by the search symbol.

5. The method for retrieval of multi-file privacy information against collusion as claimed in claim 1, characterized in that: The construction of the alternative search scheme also includes: calculating the scheme parameters based on the storage parameters and the search parameters. and ; Using solution parameters and , perform the following calculations respectively: (1) Calculate the selected The number of stages in each round; (2) Using solution parameters and , set the search symbol for all files; (3) Utilization of solution parameters and , calculate the PIR code rate of each scheme.

6. The method for retrieval of multi-file privacy information against collusion as claimed in claim 5, characterized in that: The setting of the search symbol is determined by pre-selecting The encoding matrix encodes the data packet of each file to obtain the retrieval symbol of the file.

7. A collusion-resistant multi-file privacy information retrieval system, characterized in that: It includes file storage module, solution building module, solution selection module and retrieval execution module: The file storage module is configured to: distribute and store a number of files on multiple servers by means of file subpackaging and encoding storage, thereby forming a distributed storage system; The solution building module is configured to: obtain the search information of the required document, together with the storage parameters, select the query round, and build multiple alternative search solutions; The scheme selection module is configured to: select the best retrieval scheme from multiple candidate retrieval schemes according to the PIR code rate of the retrieval scheme; The retrieval execution module is configured to: obtain the content of the required document by executing the optimal retrieval scheme; The retrieval scheme consists of multiple rounds of demand information query and marginal information query and one round of demand information query only, and the results of marginal information query are used to eliminate non-demand documents in the demand information query results of subsequent rounds to obtain the content of the demand documents; The query round is selected from In each round, select rounds, among which The first round includes demand information query and marginal information query. The round only includes demand information query, is the total number of files, is the number of required documents; The marginal information query is a query that only contains the sum of search symbols of non-required documents; The demand information query is a query that includes the sum of search symbols of the demand documents.

8. An electronic device, comprising: a memory for non-transitory storage of computer readable instructions; a processor for executing the computer-readable instructions; Wherein, when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 6 is executed.

9. A storage medium, characterized in that: The computer-readable instructions are non-transitorily stored, wherein when the computer-readable instructions are executed by a computer, the method according to any one of claims 1 to 6 is performed.

Images