CN115935390A - Attribute-based method for controlling and encrypting dynamic access to files in secure sandbox - Google Patents

Abstract

The invention provides a dynamic access control and encryption method based on attributes, which is oriented to a safe sandbox file. The method collects various attributes of the current operating environment such as time, file operating path, user identity, file keywords, file authority field and the like to form an attribute set. Extracting dynamic attributes from the attribute set, and realizing dynamic access control of the file by adopting a policy engine; and static attributes are extracted from the attribute set to implement attribute-based encryption of the file, so that double protection of the file is realized. The invention uses dynamic attributes, on one hand, an administrator can change the attributes or strategies to realize the adjustment of file authority as required; and on the other hand, the dynamic attribute is beneficial to real-time judgment of the current operating environment, and more flexible and safe access control is realized. The invention encrypts the file by multiplexing the attribute set of the access control, and realizes the sharing of the encrypted file under the condition of not sharing the key.

Description

An Attribute-Based Method for Dynamic Access Control and Encryption of Files in Security Sandbox

technical field

The invention relates to electronic file protection in the field of information security, in particular to electronic file access control and electronic file encryption protection.

Background technique

Electronic documents refer to the internal documents of all enterprises, governments and other organizational units on the terminal. Internal documents will contain important decisions and confidential information. Once leaked, it will adversely affect the corresponding organization.

Most of the existing electronic file protection technologies rely on manual management or forced blocking. The manual method relies on managers and corresponding internal policies to restrict. On the one hand, manual management has personnel costs, and on the other hand, there are many subjective factors, which are likely to cause document leakage. The method of forced blocking refers to the disabling of possible ways of file transmission on the terminal, such as USB, network devices, etc. This method has great disadvantages in today's "Internet + Office" mode, which is not conducive to file sharing within the allowed range , greatly reducing work efficiency.

At present, data leakage prevention systems (Data Leakage Prevention, DLP) are implemented in some organizations. DLP is a system architecture applied in data leakage prevention scenarios. DLP classifies and divides data and assesses risks during the data life cycle, and also controls the flow of data. However, at present, DLP focuses on the network side, and mainly checks sensitive information carried in different network communication protocols. It can detect data leakage during network transmission, but the obvious disadvantage is that the leakage path that can be detected is relatively simple, and if the virtual private The network (Virtual Private Network, VPN) or the use of encrypted traffic will be difficult to detect, and the scope of detection needs to be expanded. However, most of the current terminal DLP products only have the function of encryption, and lack the ability to execute control policies on the terminal. Once the encryption method is cracked or leaked, there is still a risk of leakage. Aiming at the problem of lack of management and control of terminal files, the present invention combines sandbox and file access control, pre-access control policy engine in file transparent encryption and decryption operation, and attaches access-allowing conditions such as time, operation path, user identity, etc. to each file . At the same time, the sandbox policy effectively protects files from being illegally read and destroyed by untrusted applications. Aiming at the problem that the key of the traditional encryption method is easily leaked, the present invention adopts the method of attribute-based encryption, which avoids the memory of the key by the user.

In the prior art, the patent specification of application number 202110739746.4 proposes a system and method for file redirection encryption and decryption. The application performs encryption and decryption in memory when reading and writing files by hooking functions related to file operations in the Linux operating system. Significant difference with the present invention is: this application is applied to the Linux operating system, and the present invention is aimed at the Windows operating system, wherein the relevant file operation API is significantly different; Identification and access control capabilities; the application requests to decrypt the file in memory for each file, and the present invention optimizes the performance of encryption and decryption based on the buffer; the application encrypts the file using traditional symmetric encryption, and the present invention is based on Attribute-based encryption implements a file encryption method and a ciphertext sharing scheme that does not require memory keys.

Contents of the invention

Aiming at the file security sandbox, the invention provides an attribute-based file dynamic access control and encryption method. In view of the lack of policy control in the file security sandbox environment, dynamic access control and encryption are combined to provide a dynamic protection file system in a secure isolation environment. Administrators can adjust the attributes of access objects in real time, and dynamically adjust file permissions through the policy engine on the terminal. Attribute-based encryption is implemented at the kernel layer, and users perform imperceptible encryption and decryption operations when operating files, and at the same time avoid the problem of key leakage during file sharing. It aims to solve the problems that there is a delay in the effect of the file control policy when the file is accessed within the organization, the file is not easy to share, and the file is easy to leak.

In order to achieve the above object, the solution of the present invention is as follows: Referring to Fig. 1, after the user is authorized by the supported authentication method, a security sandbox is established for the user at the terminal; the file operation on the user terminal is captured, and by judging whether the file operation is in a sandbox environment And analyze the file data to determine whether the file is in a protected state; if the file is an unprotected file, only the file is isolated and protected; if the file is a protected file, it is sent to the access control policy engine for processing. Determine whether the user has the operation authority corresponding to the file; if the user does not have the corresponding authority, the file operation will be intercepted and recorded; if the user has the corresponding authority of the file, the file will be transparently encrypted and decrypted, the read and write operations of the file will be divided, and the encryption will be encrypted when writing And embed the user's attribute information, decrypt based on the user's key when reading, and realize transparent access based on double buffering.

Specifically include the following steps:

Step S1: System initialization, including system master key and public key generation, user identity registration;

Step S2: The user logs in using the supported authentication method on the terminal, and establishes a secure sandbox environment for the user after obtaining authorization and authentication;

Supported authentication methods include: password authentication, USB-Key authentication, and certificate authentication.

Step S3: intercepting file operations in the terminal security sandbox;

Step S4: judging whether the file corresponding to a file operation belongs to the current security sandbox;

If the file belongs to the security sandbox, proceed to step 6;

If the file does not belong to the security sandbox, proceed to step 5;

Step S5: the file self-identification process, analyzing the file, and matching based on the file name, file format and file content. If there is a corresponding file in the attribute library, the file will be encrypted through the attributes in the attribute library and moved into the sandbox; if there is no corresponding file, the file operation will be released;

Step S6: Collect the current operating environment, specifically, including time, file operation path, user identity information, file keywords and file permission fields, and obtain whether the file operation is allowed through the access control policy engine;

If the result of the policy engine is access denied, the file operation is discarded;

If the result of the policy engine is to allow access, it will distinguish the file read and write operations, and enter the transparent encryption and decryption process of the file;

Step S7: Perform encryption and decryption operations according to the file operation type.

For file read operations, read the file encryption flag;

If it is an encrypted file, it is judged whether there is a decryption cache. If there is a cached file, the plaintext file handle is returned directly. If there is no cache, the current user attribute is used for decryption. If the decryption is successful, the plaintext file handle is returned and added to the cache;

If it is a non-encrypted file, return the user's plaintext file handle;

For file write operations, the plaintext is updated in the cache, and the file is encrypted according to the corresponding attributes.

The beneficial effects of the present invention include:

(1) Design a two-layer protection mechanism for files

On the basis of the existing security sandbox file isolation, a dynamic access control system and transparent encryption and decryption of the kernel layer are introduced. Users must have both access rights and decryption keys to access files. If any layer cannot pass, the corresponding rights cannot be obtained.

(2) Administrators can dynamically adjust file permissions

Managers can modify the attributes of access objects (environment, personnel), and the policy engine on the terminal side can modify the access rights of files in real time, and instantly reclaim or authorize user rights. The user is not aware of the whole process.

(3) Improve security during encrypted file sharing

The traditional encryption file sharing process involves the sharing of keys. The present invention gets rid of the limitation of one key for one file, and multiple users can use multiple keys to decrypt the same file, thereby improving the security in the process of sharing encrypted files.

Description of drawings

Fig. 1 is a schematic flow chart of the authorized user sandbox environment delivery in the embodiment;

Fig. 2 is the schematic diagram of Windows file operation redirection in the embodiment;

Fig. 3 is a schematic diagram of the definition of similar file management and control strategies in the embodiment;

FIG. 4 is a schematic diagram of implementation of terminal-side access control in an embodiment;

Fig. 5 is the flow chart that realizes the transparent encryption and decryption protection of terminal side file in the embodiment;

Fig. 6 is the schematic diagram of file transparent encryption mode in the embodiment;

Fig. 7 is a schematic diagram of a file transparent decryption method in an embodiment;

FIG. 8 is a flow chart of an attribute-based dynamic access control and encryption method for files in a security sandbox proposed by the present invention.

Detailed ways

This embodiment provides an attribute-based dynamic file access control and encryption method.

The following is a description of related nouns:

Attribute: Refers to the attributes of various objects, such as file confidentiality level, system time or user level, etc.

File: refers to the protected file on the terminal, which has certain confidentiality and corresponding access rights.

Security sandbox: divides the isolated area on the terminal operating system, and controls the operations performed by the processes in it.

Access control: Control users when they operate on files according to predefined policies, including allowing and blocking their operations.

Attribute-based encryption: encrypt files based on user attributes and environment attributes.

The steps of the attribute-based file dynamic access control and encryption method are as follows:

S1: Necessary initialization of the entire system, and related operations such as key generation and attribute definition related to encryption and decryption.

S101: The system authorization center performs key initialization, and generates a system master key MSK and a system public key PK.

This embodiment is realized by using the related functions provided in the library of Pairing-Based Cryptography (PBC) based on bilinear pairing, and the class A prime order elliptic curve y ² =x ³ +x is selected, and the corresponding generation parameters are shown in the table 1. The calculated system master key MSK is 128 bytes.

Table 1 Type A prime order curve parameters

Define the complete set of attributes in the system, including attributes such as user groups and file groups.

This example is given a set of attributes:

Admin: user group, administrator;

UserGroup: user group, common user;

SuFile: file group, the highest level file.

Perform SHA-256 hash calculation on different attributes, map to the points on the group through the element_from_hash function, and append to the system public key PK. Administrators can add corresponding members to different attribute groups.

S102: The user performs a registration operation by providing identity information, and the identity information provided to the authorization center needs to contain an ID that uniquely identifies the user, and the value of this specific ID provided by each user is different.

S2: The user logs in using the supported authentication methods on the terminal, and establishes a secure sandbox environment for the user after obtaining authorization and authentication. Figure 1 is a schematic diagram of the process.

After the user performs authorization and authentication on the terminal, the server generates an authentication credential and sends it to the user terminal, and the terminal uses the credential as the name to establish a sandbox environment. Note that a credential is only bound to one user at a time. Realize the dynamic binding of sandbox dynamics and user identities to prevent counterfeiting sandboxes to obtain permissions.

S3: Establish a sandbox environment on the user terminal and isolate it from the host by capturing file operations.

A method for capturing and redirecting file operations in Windows provided by the present invention is shown in FIG. 2 . This embodiment provides basic capabilities for file access control and encryption and decryption protection, and realizes its own logic by operating the Hook of the original API on the file. The method for capturing file operations in the invention is different from capturing file operations directly through the driver program, and is a way of capturing in user mode. The process consists of the following steps:

S301: DLL injection is performed when the process in the sandbox starts, and Inline Hook is performed on the file operation API, including NtCreateFile, NtOpenFile, NtDeleteFile, NtReadFile, NtWriteFile, etc.

S302: Create a Windows service, realize data collection, file encryption, access control and other judgments in the service, and create named pipes for different function points to realize synchronous or asynchronous remote procedure calls.

S303: After successfully hooking the common file API, send the logic that needs to be executed to the server through the named pipe communication before the original operation. You can choose to wait for the return result of the remote function call in a synchronous manner, or execute the origin in a non-blocking manner downwards. There is logic.

The advantage of this method is to extract the logic out of the kernel layer and the code in the Hook as much as possible. On the one hand, this solution only uses the driver layer to ensure the success of DLL injection, and the logic is implemented in the ring3 layer; on the other hand, try to reduce operations in the post-hook function, and separate the key logic from the original process space through named pipes and system services . Through this solution, the problem of unexpected execution of the original program caused by problems in the Hook code part and the blue screen crash of the system caused by triggering PatchGuard can be greatly reduced.

S4: When a file operation occurs, determine whether it is a file operation in the sandbox. If the file belongs to the current sandbox, proceed to step 6; if the file does not belong to the current sandbox, proceed to step 5

In the following scenarios, it is judged as a file operation in the sandbox:

1) Files accessed by processes in the sandbox;

2) The file path is located in the sandbox cache path.

S5: Identify the file and determine whether it is a protected file, as shown in FIG. 3 .

The beneficial effect of this step on the file double-layer protection mechanism of the present invention is reflected in the fact that the user will have file operation behaviors outside the sandbox, such as file copy, file download outside the sandbox, etc., so there will be files that are not defined in the attribute set. Content similarity judgment. The process consists of the following steps:

S501: Do not perform Dll injection for the sandboxed external program, and intercept file operations through the MiniFilter filter driver.

S502: When triggering the PreOperation callback function, perform a filter, filter by file format, such as docx, txt and other document format files.

The file format matching method is file header feature matching.

(1) For a specific file format such as docx, png or wav, etc., compare it with the feature library by parsing the file header data of the file;

(2) For files without a specific format, such as txt or files with no format suffix, the first 32 bytes of the code are detected,

Checks for known encoding sets to determine whether a file contains readable text information.

S503: Perform content matching. Classify the filtering results of S302: image format, audio format, text format, etc., and match them with the file feature library in the system. The invention provides a programmable interface for the content matching algorithm scheme, and the user can provide an optional file content similar algorithm.

In particular, the present invention provides an offline text similarity comparison method using a local sensitive hash algorithm.

(1) When creating and modifying text files in the sandbox, mark the files that need to be protected, and add the hashed results to the hash table chain corresponding to the sandbox according to the local sensitive hash algorithm.

(2) The preset similarity sensitivity value, that is, the corresponding Hamming distance, the present invention uses 3 as the preset value. If the text Hamming distance is less than or equal to the preset value, it is considered that the operating text file is highly similar to the protected text.

If the external files of the sandbox are highly similar to the protected files, the external files will be encrypted according to the corresponding attributes of the protected files, and given the same permissions as the protected files, for access by programs in the sandbox.

S6: Collect the current operating environment, specifically, including time, file operation path, user identity information, file keywords and file permission fields, and obtain whether the file operation is allowed through the access control policy engine, as shown in Figure 4.

Attribute-based access control is introduced to describe roles and permissions uniformly, and the PERM meta-model is used for policy storage, which enhances the flexibility of the access control model. The beneficial effect of this embodiment on the present invention is to classify the file attributes, distinguish the attributes in the attribute set into static attributes and dynamic attributes, improve the flexibility of the access control strategy, and realize real-time access control when the server strategy is updated. The policy is issued and applied.

The process consists of the following steps:

S601: Define and issue an attribute access control policy.

S601-1: Attribute access control policy definition. Administrators define access control policies on the server side.

First, model design is carried out according to the PERM meta-model, including request definition, policy definition, and matching rule definition. The invention divides file access request attribute into file attribute, environment attribute and file operation.

File attributes include: file name, file format, and file protected attributes;

The environment attributes include: computer time, computer identification (will be specified in S402), user credentials, and sandbox environment status, etc.;

File operations include: file read operations, file write operations.

The following is a defined access control model.

[request_definition]

r=file,env,act

[policy_definition]

p=file,env,act

[policy_effect]

e=some(where(p.eft==allow))

[matchers]

m=eval(p.file)&&eval(p.env)&&r.act==p.act

After the model is determined, the administrator needs to set the policy according to the policy definition. A specific implementation strategy is given by the above definition:

p,

r.file.name=="secret.txt"&&r.file.format=="asciitext"&&r.file.protected==true,

r.env.box_name=="DefaultBox"&&r.env.Hour>=8&&r.env.Hour<=22&&

r.env.Day>＝1&&r.env.Day<＝5&&r.env.image_name＝＝"notepad.exe"&&r.env.id＝＝"650fe3266b5b152137e362627c581902"&&r.env.user＝＝"eyJhbGbioiJIUzj1NiJ9…",

write

According to the policy description, when a file request occurs, the following matches are performed:

(1) The file name is "secret.txt"

(2) The file format is asciitext

(3) The file protection bit is protected

(4) The current sandbox name is "DefaultBox"

(5) The current visit time is within the range of 8:00 to 22:00 from Monday to Friday

(6) The current request process is notepad.txt

(7) The current terminal ID is "650fe3266b5b152137e362627c581902"

(8) The current user ID is "eyJhbGbioiJIUzj1NiJ9..."

(9) The current operation is a write operation

When a file request is successfully matched, its access can be allowed, otherwise access is denied.

S601-2: Issuing the attribute access control policy. Since the present invention performs access control on the terminal side, it needs to be synchronized with the access control policy of the server in time. The present invention issues strategies based on ZeroMQ.

The server maintains the policy version number, and when the administrator changes the policy, the version number is updated.

After passing the identity authentication, the terminal user actively applies to the server for the current policy. The server returns the response request, returns the policy version, and compares it with the local version. If the local version number is smaller than that of the server, request policy synchronization from the server.

During the use of the end user, the client and the server use the ZMQ publish-subscribe mode to realize a long-term connection. When the server changes the policy, it broadcasts the new version number to all clients in the subscriber queue to notify the change. When the client is inconsistent with the local version, it requests the server to perform policy synchronization.

S602: Realize at the terminal side of the attribute access control. The invention proposes a method for realizing attribute access control on the terminal side.

(1) Start the local service process, create a new named pipe "\\.\pipe\abacpipe" and listen to this named pipe to receive information related to client access control.

(2) Inline hook the file operation-related API by injecting DLL into the program running in the sandbox. When the corresponding file operation occurs, the current sandbox environment parameters, file attributes, and file operations are first sent to the service process through the named pipe. When the service process receives the operation related to the file request, it completes the missing environment parameters in the policy. The specific missing parameter completion methods are listed below:

Computer ID: Determined by the system hardware, a string is generated by concatenating CPUID, BIOS UUID, and system serial number, which is hashed and recorded as a computer ID;

System time: The client and server perform NTP time calibration to obtain the corrected system time.

(3) In the local service process, the policy engine performs dynamic matching according to the policy model, and the execution results are "passed" and "denied", which are returned to the client. After receiving the execution result, the client decides whether to continue to execute the original system API,

And record user operations.

S7: Perform file encryption and decryption operations according to the captured file operation type.

This step provides the present invention with a multi-user key sharing file transparent encryption method, which belongs to the second-layer encryption protection in the file double-layer protection mechanism. The beneficial effect of the present invention is that there is no need to memorize the file encryption key, and multiple users can decrypt the ciphertext at the same time, which improves the security of file sharing. As shown in Figure 5.

Since the files in the sandbox may maliciously bypass the access control engine or be leaked by physically reading the hard disk, the files need to be encrypted. Aiming at the problems of single key and complex ciphertext sharing method in the traditional encryption method, the present invention proposes a file transparent encryption method for multi-user key sharing under the background of attribute access control. The process consists of the following steps:

S701: Generate a user key. The user key is issued by the server after the user passes the authentication. It is the same as the policy delivery mechanism described in S601-2. The user key belongs to runtime information and can be modified by the administrator in real time. User key construction method:

(1) Definition of user attributes. The administrator defines user attributes on the server side, and the user attributes need to be included in the user attribute set

(2) Generate user key. Based on CP-ABE, the server generates user key SK through user attributes and system master key MSK, and sends it to the user terminal in real time through ZMQ.

S702: File transparent encryption. When the file writing operation in the sandbox is released, the file transparent encryption operation will be performed. First move the file to be encrypted into the plaintext buffer protected by the sandbox, and encrypt the file through attribute encryption. The encrypted file is moved into the original path of the file, so as to realize the transparent encryption of the file, as shown in Figure 6.

The specific encryption method is as follows:

(1) Construct user access structure

The access structure is constructed according to the attribute group or single attribute in the global attribute set defined by the system. Divided into request object attributes and target object attributes, request object attributes are user groups or specific users, and response object attributes are file groups or specific files. An example would be:

SuFile∩Admin

SuFile and Admin are defined with the attribute set in S101, which means that only the access object and the members of the administrator group who have access to the SuFile group authority can decrypt the file.

On the user terminal, for a file with an original access structure, the original access structure is inherited by default during encryption; for ordinary files, the access structure is constructed with a specific login user by default, such as Bob is the current login user, the constructed access structure is Bob; for Files that need to expand the scope of access can apply to the server, and the access structure will be issued after approval.

(2) File transparent encryption

Start the local service process, create a new named pipe "\\.\pipe\encrypt" and listen to this named pipe, which is used to receive the encrypted file path of the client.

Inline hook CloseFile by injecting DLL into programs running in the sandbox. When the corresponding file operation occurs, the current file path is first sent to the service process through the named pipe. When the service process receives the file path, it analyzes the current file access structure to determine the specific access structure used in attribute encryption. Write the original path of the file after attribute encryption.

S703: Transparently decrypt the file. When the file read operation in the sandbox is released, the file transparent decryption operation is performed, as shown in Figure 7.

The specific decryption method is as follows:

Start the local service process, create a new named pipe "\\.\pipe\decrypt" and listen to this named pipe to receive the client decrypted file path.

Inline hook OpenFile by injecting DLL into programs running in the sandbox. When the corresponding file operation occurs, the current file path is first sent to the service process through the named pipe. When the service process receives the file path, it judges whether the current file exists in the file buffer. If the same file exists in the file buffer, it means that the file has been decrypted, and the file redirection operation will return the file handle in the buffer; if there is no identical file in the file buffer, it will be decrypted according to the current user's key. If the decryption is successful, it will be added to the file buffer, and the file handle in the buffer will be returned; if the decryption fails, access will be denied.

Claims (1)

1. A method for controlling and encrypting dynamic file access in a security sandbox based on attributes is characterized by comprising the following steps:

step S1: carrying out system initialization, including generation of a system master key and a public key and user identity registration;

step S2: a user logs in at a terminal by using a supported authentication mode, and establishes a safe sandbox environment for the user after obtaining authorized authentication;

the supported authentication methods include: password authentication, USB-Key authentication and certificate authentication;

and step S3: intercepting file operation in a terminal security sandbox;

and step S4: judging whether a file corresponding to one file operation belongs to a current safety sandbox or not;

if the file belongs to the safe sandbox, performing operation in step 6;

if the file does not belong to the safe sandbox, performing the operation of the step 5;

step S5: in the file self-identification process, analyzing the file, and matching based on the file name, the file format and the file content; if the corresponding file exists in the attribute library, encrypting the file through the attribute in the attribute library, and moving the file into a sandbox; if the corresponding file does not exist, the file operation is released;

step S6: collecting the current operating environment, specifically comprising time, a file operating path, user identity information, file keywords and a file authority field, and obtaining whether the file operation is allowed or not through an access control strategy engine;

if the result of the policy engine is that the access is refused, discarding the file operation;

if the result of the strategy engine is that the access is allowed, distinguishing the file read-write operation, and entering a transparent encryption and decryption process of the file;

step S7: performing encryption and decryption operation according to the file operation type;

for file reading operation, reading a file encryption zone bit;

if the file is an encrypted file, judging whether a decryption cache exists or not, directly returning a plaintext file handle to the file with the cache, if the cache does not exist, decrypting by using the current user attribute, and returning the plaintext file handle and adding the plaintext file handle to the cache if the decryption is successful;

if the file is a non-encrypted file, returning a plaintext file handle of the user;

for file write operations, the plaintext is updated in the cache, and the file is encrypted according to the corresponding attributes.

Images