WO2021031801A1

WO2021031801A1 - Indexing system and method for protecting user privacy in social network

Info

Publication number: WO2021031801A1
Application number: PCT/CN2020/105048
Authority: WO
Inventors: 张明武; 陈誉
Original assignee: 湖北工业大学
Priority date: 2019-08-16
Filing date: 2020-07-28
Publication date: 2021-02-25
Also published as: CN110569413B; CN110569413A

Abstract

Disclosed are an indexing system and method for protecting a user privacy in a social network. Said system comprises a server and several users; and a network connectivity diagram is constructed according to a social relationship among the users, taking user information as a vertex, and taking a social state among the users subjected to abstract mapping as a weight value: the more intimate, the smaller the weight value, and the more alienated, the larger the weight value; the user information comprises a unique identity identifier, and a name, contact information, user groups and grouping levels which are encrypted, the groups including a family group, a friend group, a companion group, a classmate group and an acquaintance group, and level values being 1, 2, 3, 4 and 5 respectively; and it is assumed that the users of an address book are all included in specific groups. Said method comprises initializing the system, constructing an encrypted social network, generating an optimal index path, and acquiring optimal index information by means of oblivious transfer. The present invention provides high privacy protection security, fast calculation speed and high processing efficiency.

Description

Index system and method for protecting member privacy in social network

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on August 16, 2019, the application number is 201910759501.0, and the invention title is "An Index System and Method for Protecting Member Privacy in Social Networks". The reference is incorporated in this application.

Technical field

The invention belongs to the technical field of information security, and relates to an indexing system and method for user data privacy protection, and in particular to a system and method for member indexing in a social network aimed at user privacy protection requirements.

Background technique

With the efficient development of cloud storage, people's various private information will be uploaded to the cloud, such as backup data such as address book and mobile phone photos. The server side claims that it will not leak these private data, but as the data leaks out after the server is breached, people will gradually not give full trust to the server, but will encrypt the original data and upload it. Therefore, compared to operations on plaintext, processing data on ciphertext to achieve corresponding goals has attracted more and more attention.

Nowadays, the process of entering a higher education and job hunting requires the introduction of celebrities, but it takes a lot of time and energy to find the contact information of this target member in a large number of people. Since everyone has a certain social relationship, a social network can be constructed to index target members through the address book. However, this information is private information and cannot be stored directly on the server to avoid leakage. Therefore, there is an urgent need in the art to develop a method or system that can find an optimal indexing result according to the intimacy of social relationships during the indexing process.

Summary of the invention

The purpose of the present invention is to provide an indexing system and method for protecting the privacy of members in a social network, so as to quickly and accurately find the best index result according to the intimate state of the social relationship during the indexing process.

In order to achieve the above objective, the present invention provides the following solutions:

An index system to protect the privacy of members in social networks, including: a server and several members;

The server constructs a network connection graph according to the social relations between the members; the vertices of the network connection graph are member information; the edges of the network connection graph are the relationships between the members; The weight of the edge is the social status of each member after the abstract mapping; the member information includes the unique identity, the encrypted name, the encrypted contact information, the member group and the group level; the member group includes: family group , Friend group, partner group, classmate group and infrequent relationship group; the numerical values of the grouping levels are 1, 2, 3, 4, and 5 respectively;

The closer the social state is, the smaller the corresponding weight is, and the more distant the social state is, the greater the corresponding weight; and each member in the address book is set to be included in a specific group.

A method for protecting member privacy in a social network is applied to the aforementioned index system for protecting member privacy in a social network; the method includes:

System initialization, including:

The server selects the security parameter κ, and determines the public key N=pq,g and the private key (λ,μ) of the Paillier encryption algorithm; where p and q are two large prime numbers and satisfy |p|=|q|=κ, g is

A generator of, Z is the set of integers, Z ^* is the set of Z except 0, N=pq;

The member determines the identity and the member key corresponding to the member according to the random number generated during encryption by the Paillier encryption algorithm and the public key;

Constructing an encrypted network connection graph according to the identity identifier and the member key; constructing the encrypted network connection graph specifically includes:

Construct vertex information and weight information;

Construct a graph G=(V,E) according to the vertex information and the weight information, where V represents a vertex set, E represents a weight set; the graph G=(V,E) is the encrypted Network connectivity diagram;

The generation of the optimal index path according to the network connectivity graph specifically includes:

Obtaining vertices and edges in the network connected graph; constructing a vertex set according to the obtained vertices, and constructing an edge set according to the obtained edges;

Taking a vertex in the vertex set as the current vertex, traversing the vertices adjacent to the current vertex, and determining the vertex with the smallest weight among the adjacent vertices;

Add the vertex with the smallest weight to the set of vertices, and use the vertex with the smallest weight as the new current vertex, and return to step "Take a vertex in the vertex set as the current vertex, traverse and The vertex adjacent to the current vertex, determine the vertex with the smallest weight among the adjacent vertices" until all the vertices in the network connectivity graph are added to the vertex set;

Obtain the identity of the target member;

The member determines the optimal index path according to the identity of the target member; the process of determining the optimal index path specifically includes:

The member determines the vertex index path according to the identity, and locates the vertex corresponding to the target member according to the uniqueness of the identity;

The server sends the precursor tuple to the applicant, and the member performs a recursive query based on the precursor vertex of the vertex corresponding to the target member to obtain information about the optimal index path.

Preferably, the member determines the identity corresponding to the member according to the random number generated during encryption by the Paillier encryption algorithm and the public key, which specifically includes:

_{The member generates an encrypted value C i} according to the random number generated during encryption by the Paillier encryption algorithm and the public key, and sends the encrypted value C _i to the server;

The server obtains a random integer i as the first identity of the member, and decrypts the encrypted value C _i according to the public key to obtain the decrypted value of _{C i;}

The server sends the sum of the first identity and the _{decrypted value of C i} to the member, and the member determines according to the received sum of the first identity and the decrypted value of _{C i} Obtain the identity corresponding to the member.

Preferably, the construction process of the member key specifically includes:

The member uses a random number generated during encryption with the Paillier encryption algorithm as its own member key;

The member encrypts its own member key and sends it to other members through the neighboring member group;

The neighboring members multiply their own encrypted member keys by homomorphism and send the products to the server;

The server decrypts the product according to the private key to obtain the upload key of the member.

Preferably, the process of constructing vertex information in the network connectivity graph specifically includes:

The member encrypts its own identity and the identity of the neighboring member with the system public key, and sends the encrypted identity of the first member and the identity of the neighboring member to the server;

The member uses the system public key and his own member key to encrypt and upload his own information;

The server decrypts the information uploaded by all members and makes a network map based on the uploaded member ID;

Obtain the vertex information according to the network graph; the vertex information includes the identity of the member, the identity information and the contact information that have been confused by the member key.

Preferably, before determining the vertex with the smallest weight among the adjacent vertices, the method further includes:

The member counts the number of short messages sent and the number of calls with neighboring members;

Determine the number of communications according to the number of short messages sent and the number of calls, and determine the weight between members according to the number of communications;

The member uses the member key and the system public key to encrypt the weight between the neighboring members, and the server decrypts the weight to obtain the confused weight of the member key.

Preferably, determining the vertex with the smallest weight among the adjacent vertices specifically includes:

For vertices in the network connectivity graph that are not in the vertex set and edges are not in the edge set, the server recursively finds the predecessor vertices of the vertex in turn until the predecessor vertex is the current vertex, and records the number of predecessors;

Determine the weight according to the number of precursors;

The members compare the ownership values and determine the vertex corresponding to the smallest weight value.

Preferably, the update process of the precursor vertex specifically includes:

Determine whether all the vertices that are not in the vertex set in the network connected graph are reachable with the current vertex. If they are not reachable, the precursor vertex is not updated. If the edge of the current vertex does not belong to the edge set, and all Vertices that are not in the set of vertices belong to the set of edges, then the current vertex is updated as a precursor vertex, and if all vertices not in the set of vertices are reachable with the current vertex, the current vertex Is the precursor vertex, and calculate the weight of the current vertex and the weight of the initial precursor vertex;

The weight value of the current vertex is compared with the weight value of the initial predecessor vertex to obtain a comparison result; the server will compare the vertex with the smaller weight value in the comparison result as the new predecessor vertex.

Preferably, the server sends the predecessor tuple to the applicant, and the member performs a recursive query based on the predecessor vertex of the vertex corresponding to the target member to obtain the information of the optimal index path, which specifically includes:

The server encrypts the identities of all members with public keys, and generates a two-tuple based on the encrypted identity and the upload key, and then sends the two-tuple to the applicant;

The member encrypts the two-tuple to form a new two-tuple;

The server uses the key to decrypt the new two-tuple;

The member determines a random number according to the first value in the decrypted two-tuple, and obtains the identity of the target member according to the second value in the decrypted two-tuple and the random number;

The member obtains an index path according to the identity of the target member, and locates the vertex of the target member in the network connectivity graph by the uniqueness of the identity;

Determine the precursor vertex corresponding to the vertex according to the vertex; use the precursor vertex to perform a recursive query to obtain an index path; the index path is the optimal index path.

Preferably, the member encrypts the two-tuple to form a new two-tuple, which specifically includes:

The member shuffles the order in the two-tuple, and converts the identity information of the target member into the code unit _E , encrypts the code unit _E with a public key, and the first in the two-tuple The product of the two values is used as the new first value in the two-tuple; multiple random numbers are selected, and the random numbers are respectively encrypted and multiplied by the second value of the two-tuple as the two-tuple Group the new second value to form a new two-tuple.

Compared with the prior art, the method of the present invention has the following advantages and beneficial effects:

The present invention has high security. All encryption processes are implemented by using Paillier homomorphic encryption algorithm. Due to the semantic security of this algorithm, even if valid ciphertexts are published, such as weight information, criminals cannot obtain any private information. In addition to knowing the size relationship, the applicant does not know the additional information when updating the weight. In the end, he only knows the referral path and does not know the information of other tutors; the server does not know the member key information, and the member’s private information is still not known after decryption. Therefore, the present invention has high privacy protection security.

The present invention considers the effective domain [0, N] and the cipher text area [0, N ² ] of the paillier encryption system, and distinguishes the cipher text added with the normal homomorphism from the unreachable data and considers them separately.

Using the inadvertent transmission method, the server does not know the specific identity of the target member that the applicant wants, and the transmission parameters can also allow the applicant to obtain valid information to verify whether it has been cheated and prevent tampering.

The present invention proposes an index method for protecting member privacy in a social network. Under the premise of ensuring security, server operation response speed is fast, processing is efficient, and parameters in the process can be generated in advance.

Description and drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some of the present invention. Embodiments, for those of ordinary skill in the art, without creative work, other drawings can be obtained from these drawings.

Figure 1 is a system architecture diagram of an embodiment of the present invention;

Figure 2 is a flowchart of a method according to an embodiment of the present invention;

Figure 3 is a flowchart of member registration in a method according to an embodiment of the present invention;

Figure 4 is a flowchart of constructing a network diagram in a method according to an embodiment of the present invention;

FIG. 5 is a flowchart of generating an optimal index path in the method of an embodiment of the present invention;

Fig. 6 is a flowchart of inadvertent transmission in a method according to an embodiment of the present invention.

detailed description

The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

The purpose of the present invention is to provide an indexing system and method for protecting member privacy in a social network, so as to quickly and accurately find the best indexing result according to the intimate state of the social relationship during the indexing process.

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

Example one

FIG. 1 is a system architecture diagram of an embodiment of the present invention. As shown in FIG. 1, an index system for protecting member privacy in a social network includes: a server and several members.

The server constructs a network connection graph according to the social relationship between the members. The vertices of the network connected graph are member information. The edges of the network connectivity graph are the relationships between the members. The weights of the edges of the network connection graph are the social states between members after abstract mapping. The member information includes a unique identification, an encrypted name, an encrypted contact method, a member group, and a group level. The member groups include: family group, friend group, buddy group, classmate group and infrequent relationship group. The numerical values of the grouping levels are 1, 2, 3, 4, and 5, respectively.

The closer the social state, the smaller the corresponding weight, and the more distant the social state, the larger the corresponding weight. And each member in the set address book has been included in a specific group.

FIG. 2 is a flowchart of a method according to an embodiment of the present invention. As shown in FIG. 2, a method for protecting member privacy in a social network is applied to the aforementioned indexing system for protecting member privacy in a social network. The method includes:

System initialization, including:

The server selects the security parameter κ, and determines the public key N=pq, g and the private key (λ, μ) of the Paillier encryption algorithm. Where p and q are two large prime numbers, and satisfy |p|=|q|=κ, g is

A generator of, Z is a set of integers, {0,1,2,...,N-1}, Z ^* is a set other than 0 in Z, N is a number generated by security parameters, N=pq.

The member determines the identity and the member key corresponding to the member according to the random number generated during encryption by the Paillier encryption algorithm and the public key.

Wherein, the member determines the identity corresponding to the member according to the random number generated during encryption by the Paillier encryption algorithm and the public key, which specifically includes:

_{The member generates an encrypted value C i} according to the random number generated during encryption by the Paillier encryption algorithm and the public key, and sends the encrypted value C _i to the server.

The server obtains a random integer i as the first identity of the member, and decrypts the encrypted value C _i according to the public key to obtain the decrypted value of _{C i.}

The construction process of the member key specifically includes:

The member uses a random number generated during encryption with the Paillier encryption algorithm as its own member key.

The member encrypts its own member key and sends it to other members through the neighboring member group.

The neighboring member multiplies the encrypted member key by the homomorphic property and sends the product to the server.

An encrypted network connection graph is constructed according to the identity identifier and the member key. The construction of the encrypted network connection graph specifically includes:

Construct vertex information and weight information.

The process of constructing vertex information in the network connectivity graph specifically includes:

The member uses the system public key to encrypt its own identity and the neighboring member's identity, and sends the encrypted first member's identity and neighboring member's identity to the server.

The member uses the system public key and his own member key to encrypt and upload his own information.

The server decrypts the information uploaded by all members and makes a network map based on the uploaded member ID.

Acquire the vertex information according to the network graph. The vertex information includes: the identity of the member and the identity information and contact information that have been confused by the member key.

Construct a graph G=(V, E) according to the vertex information and the weight information, where V represents a vertex set, and E represents a weight set. The graph G=(V, E) is the encrypted network connectivity graph.

Obtain vertices and edges in the network connected graph. Construct a vertex set based on the acquired vertices, and construct an edge set based on the acquired edges.

Taking a vertex in the vertex set as the current vertex, traversing the vertices adjacent to the current vertex, and determining the vertex with the smallest weight among the adjacent vertices.

Wherein, determining the vertex with the smallest weight among the adjacent vertices specifically includes:

For vertices in the network connectivity graph that are not in the vertex set and edges are not in the edge set, the server recursively finds the predecessor vertices of the vertex in turn until the predecessor vertex is the current vertex, and records the number of predecessors.

The weight is determined according to the number of precursors.

Add the vertex with the smallest weight to the set of vertices, and use the vertex with the smallest weight as the new current vertex, and return to step "Take a vertex in the vertex set as the current vertex, traverse and For the vertex adjacent to the current vertex, determine the vertex with the smallest weight among the adjacent vertices" until all the vertices in the network connectivity graph are added to the vertex set.

Get the identity of the target member.

The member determines the best index path according to the identity of the target member. The process of determining the optimal index path specifically includes:

The member determines the vertex index path according to the identity identifier, and locates the vertex corresponding to the target member according to the uniqueness of the identity identifier.

Wherein, the server sends the precursor tuple to the applicant, and the member performs a recursive query based on the precursor vertices of the vertices corresponding to the target member to obtain information on the optimal index path, which specifically includes:

The server encrypts the identities of all members with public keys, and generates a two-tuple according to the encrypted identity and upload key, and then sends the two-tuple to the applicant.

The member encrypts the two-tuple to form a new two-tuple, which specifically includes:

The member shuffles the order in the two-tuple, and converts the identity information of the target member into the code unit _E , encrypts the code unit _E with a public key, and the first in the two-tuple The product of these values is used as the new first value in the two-tuple. Select a plurality of random numbers, and multiply the encrypted random numbers with the second value of the two-tuple as the new second value of the two-tuple to form a new two-tuple.

The server uses the key to decrypt the new two-tuple.

The member determines a random number according to the first value in the decrypted two-tuple, and obtains the identity of the target member according to the second value in the decrypted two-tuple and the random number.

The member obtains the index path according to the identity of the target member, and locates the vertex of the target member in the network connectivity graph by the uniqueness of the identity.

The precursor vertex corresponding to the vertex is determined according to the vertex. The predecessor vertices are used for recursive query to obtain the index path. The index path is the optimal index path.

The update process of the precursor vertex specifically includes:

Determine whether all the vertices that are not in the vertex set in the network connected graph are reachable with the current vertex. If they are not reachable, the precursor vertex is not updated. If the edge of the current vertex does not belong to the edge set, and all Vertices that are not in the set of vertices belong to the set of edges, then the current vertex is updated as a precursor vertex, and if all vertices not in the set of vertices are reachable with the current vertex, the current vertex Is the predecessor vertex, and calculate the weight of the current vertex and the weight of the initial predecessor vertex.

The weight of the current vertex is compared with the weight of the initial predecessor vertex to obtain a comparison result. The server will compare the vertex with the smaller weight value as the new predecessor vertex.

As a preferred embodiment of the present invention, before determining the vertex with the smallest weight among the adjacent vertices, the method further includes:

The member counts the number of short messages sent and the number of calls with neighboring members.

The number of communications is determined according to the number of short messages sent and the number of calls, and the weight between members is determined according to the number of communications.

Example two

As shown in Figure 1, the present invention provides an indexing system for protecting member privacy in a social network. The system includes: servers and members. Assuming that there are n members in the system, a member User _S applies for a referral from the target member User _E , but he does not directly know this member, and he needs to obtain a way to contact the target member through the address book members he knows. Each member has a certain social relationship. The information is encrypted and uploaded to the server to form a network connection graph. The vertex is the member's unique identification ID _i (i=1, 2,...,n) and the encrypted Identity information (work unit, name), contact information, where these identities are mapped to [1,n]. Assuming that each member has been included in a specific group, the grouping levels include: family, friends, partners, classmates, and uncommon relationships, and the level values are (1,2,3,4,5). The weight of the edge is the social state between members after abstraction: according to the grouping level and the number of calls and SMS, the state mapping is calculated as a positive integer (the closer the weight is, the smaller the weight is, the greater the distance is). After the entire calculation process is over, User _S obtains an optimal index path for target members: the total weight of the current path is the smallest, which means the higher the feasibility of the referral, and at the same time, it does not know anything other than the member information on the path Member's social information. The server also does not know the specific identity of the target member for which _{User S applies for indexing.}

Please refer to Figure 2. The present invention provides an indexing method for protecting member privacy in social networks, which includes the following steps:

Step 1: System initialization.

Step 1.1: The server selects the security parameter κ, and calculates the public key (N=pq, g) and private key (λ, μ) of the Paillier encryption algorithm. Where p and q are two large prime numbers and satisfy |p|=|q|=κ, g is

A generator of.

Step 1.2: Member registration identity. The membership registration process involves Figure 3.

Step 1.2.1: The member User _i (i=1, 2,...,n) selects a random number r _i and sends it to the server after encryption:

Among them, all r in the encryption process are random numbers generated by the Paillier system during encryption.

Step 1.2.2: The server randomly selects a unique integer from the integer sequence [i=1, 2,...,n] as the current member’s identity ID _i , adds it to the decrypted _{value of C i} , and returns the result to User _i .

Step 1.2.3: User _i subtracting the values of the received self-generated random number r _i acquires its own identity ID _i.

Step 1.2.4: User _i sends his identity ID _i to neighboring members (members who have communication links).

Step 1.3: System members construct member keys.

Step 1.3.1: The member User _i (i=1, 2,...,n) chooses a random number as its member key sk _i .

Step 1.3.2: User _S encrypts his own member key sk _S and sends it to all members of the system through his neighboring member group.

Step 1.3.3: The member User _j (j≠i) is sent to the server after being homomorphic multiplied by the encrypted member key.

Step 1.3.4: The server uses the Paillier decryption key (λ, μ) and decryption algorithm to decrypt n-1 ciphertexts to obtain the upload key:

SK _j =sk _S +sk _j (j≠i)

Step 2: Build an encrypted social network. The construction process of encrypted social network involves Figure 4.

Step 2.1: Construct secure vertex information.

Step 2.1.1: User _i (i=1, 2,...,n) uses the system public key to encrypt his ID _i and the identities of neighboring members and send them to the server.

Step 2.1.2: User _i uses the system public key and his own member key sk _{i to} encrypt and upload his own information: identity information and contact information. The encryption process first converts the characters into integer types unit _i and tel _i through ASCII encoding, and then encrypts the upload (beyond the plaintext space size can be grouped):

Step 2.1.3: The server decrypts the information uploaded by all members:

Unit _i =unit _i +sk _i Tel _i =tel _i +sk _i .

The server makes a network map according to the uploaded member identification. Wherein the vertex information and the identity of a member of the member key obfuscated identity information, contact information, i.e., V _i vertex information _{_{comprising: ID i, Tel i, Unit}} i, the link between the side members: a member User _i The sent neighbor member ID is determined.

Step 2.2: Construct safe weight information.

Step 2.2.1: Member User _i counts the number of short messages sent and the number of calls with neighboring member User _j , so as to find the number of communications: num _ij is the sum of the number of short messages and calls (it is assumed that the number of communications does not exceed 1000) . Therefore, the weight is:

e(i,j)=rank _ij +1000-num _ij .

That is, the closer the connection (frequent calls, the higher the group level), the smaller the weight between members.

Step 2.2.2: The member User _i uses the member key sk _i and the system public key to encrypt the weight e(i,j) between the _{neighboring member User j:}

Step 2.2.3: After decryption, the server obtains the obfuscated weight of the member key:

E(i,j)=e(i,j)+sk _i .

Step 2.3: Construct a graph G=(V, E), where V represents a set of vertices, and E represents a set of weights.

Step 3: Generation of the best index path. Vertex selection and weight update refer to Figure 5.

Step 3.1: The server constructs a set S, a set D and a tuple P, and sets the current vertex.

Step 3.1.1: The server constructs a vertex set: S={v _S }, and sets the current vertex to v _t =v _S.

Step 3.1.2: If there is no direct connection between all two points in graph G, that is, they are unreachable, add the edge to D.

Step 3.1.3: All precursors are set to vertices _{_{v s: P (v i)}} = v s, v i ∈V.

Step 3.2: Pick the best vertex and join the set S.

_{Step 3.2.1: For vertices v j} where all points in graph G are not in S and edges are not in D: the server recursively finds its predecessor vertices v _k until the predecessor vertex is v _S , records the number of predecessors as count, calculates the weight E(s,j) is the sum of index weights of all predecessors. If v _j is not _{the adjacent vertex of v S} _{, then SK k} must be subtracted in this process. Send the weight E(s,j) to the applicant User _S , and User _S adds (count-2)×sk _{S to} this weight.

Step 3.2.2: User _S compares the ownership values, and sends the vertex v _{j'corresponding} to the smallest weight value to the server.

Step 3.2.3: server vertices v _{j 'is} added to the _{S: S = S∪ {v j'} }, and set the current vertex v _t = v _{j '.}

Step 3.3: Update the predecessor vertices of all vertices.

Step 3.3.1: For all the vertices v _j not in S in the graph G: If the current vertex is unreachable from it, that is, E(t,j)∈D, it means that the weight will only be greater after the update, so it is not updated. If

Indicates that E(s,j) will change from unreachable to reachable after the weight is updated, then the precursor vertex is updated to P(v _j )=v _t , and D is updated. If both sides are reachable, the same as in step 3.2.1, the applicant judges that _{the weight calculated by assuming that the precursor vertex is v t} is compared with the weight calculated by the original precursor vertex v _k and the comparison result is sent to the server.

Step 3.3.2: The server updates the vertex with a smaller weight to be the new precursor vertex.

Step 3.4: Repeat steps 3.2 and 3.3 until S contains all the vertices in the graph G.

Step 4: Inadvertently obtain the best index information. The specific form of inadvertent transmission refers to Figure 6.

Step 4.1: Obtain the identity ID _{E of the} _{target member User E.}

Step 4.1.1: The server _{encrypts the ID i of} all members with the public key, and multiplies their identity information by the upload key SK _{i to} form a two-tuple y _i , and sends n two-tuples to the application By.

Step 4.1.2: User _S randomly shuffles the order of the two-tuples first, then converts the identity information of the _{target member User E} _{into unit E} _{, encrypts sk S} -unit _E with the public key, and multiplies it in the first of all the two-tuples Value. Finally, select n random numbers (r ₁ ,...,r _n ), encrypt the random numbers and multiply them on the second value of the _{two-tuple to form a new z j} , and send the n two-tuples to the server .

Step 4.1.3: The server uses the key to decrypt all the first values of the n-tuple, and sets the sequence subscript l of the two-tuple with a unique value of 0 (assuming that the positioning information is not repeated) and the second value of the current tuple. The values form a new two-tuple and are sent to User _S. Among them, only the plaintext of the first value in the two-tuple is unit _E , and the ciphertext appended by User _S can be homomorphically added with a modulus of 0.

Step 4.1.4: User _S _{finds the corresponding random number r l} by subscripting the first value of the tuple, and then subtracts r _l from the second value to obtain the identity mapped by the target member User _E _{: ID E} = ( ID _E +r _l )-r _l .

Step 4.2: Get the best index.

Step 4.2.1: User _S searches the vertex index path obtained in step 3 according to ID _E _{, and locates the vertex v E} corresponding to the target member by the uniqueness of the identity.

Step 4.2.2: The server sends the predecessor tuple P to the applicant, and User _S queries the predecessor vertex of vertex v _E recursively to obtain the information of the entire optimal index path.

The present invention implements an indexing scheme for protecting member privacy in social networks based on the discrete logarithm problem and homomorphic encryption algorithm. The scheme realizes resistance to external attacks and internal attacks, and can also detect server fraud.

The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other.

Specific examples are used in this article to illustrate the principles and implementation of the present invention. The description of the above examples is only used to help understand the method and core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the present invention There will be changes in the specific implementation and scope of application. In summary, the content of this specification should not be construed as limiting the present invention.

Claims

An index system for protecting the privacy of members in a social network, which is characterized by comprising: a server and a number of members;

The server constructs a network connection graph according to the social relations between the members; the vertices of the network connection graph are member information; the edges of the network connection graph are the relationships between the members; The weight of the edge is the social status of each member after the abstract mapping; the member information includes the unique identity, the encrypted name, the encrypted contact information, the member group and the group level; the member group includes: family group , Friend group, partner group, classmate group and infrequent relationship group; the numerical values of the grouping levels are 1, 2, 3, 4, and 5 respectively;

The closer the social state is, the smaller the corresponding weight is, and the more distant the social state is, the greater the corresponding weight; and each member in the address book is set to be included in a specific group.
A method for protecting member privacy in a social network, characterized in that it is applied to an index system for protecting member privacy in a social network according to claim 1; the method includes:

System initialization, including:

The server selects the security parameter κ, and determines the public key N=pq,g and the private key (λ,μ) of the Paillier encryption algorithm; where p and q are two large prime numbers and satisfy |p|=|q|=κ, g is
A generator of, Z is the set of integers, Z * is the set of Z except 0, N=pq;

The member determines the identity and the member key corresponding to the member according to the random number generated during encryption by the Paillier encryption algorithm and the public key;

Constructing an encrypted network connection graph according to the identity identifier and the member key; constructing the encrypted network connection graph specifically includes:

Construct vertex information and weight information;

Construct a graph G=(V,E) according to the vertex information and the weight information, where V represents a vertex set, E represents a weight set; the graph G=(V,E) is the encrypted Network connectivity diagram;

The generation of the optimal index path according to the network connectivity graph specifically includes:

Obtaining vertices and edges in the network connected graph; constructing a vertex set according to the obtained vertices, and constructing an edge set according to the obtained edges;

Taking a vertex in the vertex set as the current vertex, traversing the vertices adjacent to the current vertex, and determining the vertex with the smallest weight among the adjacent vertices;

Add the vertex with the smallest weight to the set of vertices, and use the vertex with the smallest weight as the new current vertex, and return to step "Take a vertex in the vertex set as the current vertex, traverse and The vertex adjacent to the current vertex, determine the vertex with the smallest weight among the adjacent vertices" until all the vertices in the network connectivity graph are added to the vertex set;

Obtain the identity of the target member;

The member determines the optimal index path according to the identity of the target member; the process of determining the optimal index path specifically includes:

The member determines the vertex index path according to the identity, and locates the vertex corresponding to the target member according to the uniqueness of the identity;

The server sends the precursor tuple to the applicant, and the member performs a recursive query based on the precursor vertex of the vertex corresponding to the target member to obtain information about the optimal index path.
The method for protecting member privacy in a social network according to claim 2, wherein the member determines the identity corresponding to the member according to the random number generated during encryption by the Paillier encryption algorithm and the public key, Specifically:

The member generates an encrypted value C i according to the random number generated during encryption by the Paillier encryption algorithm and the public key, and sends the encrypted value C i to the server;

The server obtains a random integer i as the first identity of the member, and decrypts the encrypted value C i according to the public key to obtain the decrypted value of C i;

The server sends the sum of the first identity and the decrypted value of C i to the member, and the member determines according to the received sum of the first identity and the decrypted value of C i Obtain the identity corresponding to the member.
The method for protecting member privacy in a social network according to claim 2, wherein the process of constructing the member key specifically includes:

The member uses a random number generated during encryption with the Paillier encryption algorithm as its own member key;

The member encrypts its own member key and sends it to other members through the neighboring member group;

The neighboring members multiply their own encrypted member keys by homomorphism and send the products to the server;

The server decrypts the product according to the private key to obtain the upload key of the member.
The method for protecting member privacy in a social network according to claim 2, wherein the process of constructing vertex information in the network connectivity graph specifically includes:

The member encrypts its own identity and the identity of the neighboring member with the system public key, and sends the encrypted identity of the first member and the identity of the neighboring member to the server;

The member uses the system public key and his own member key to encrypt and upload his own information;

The server decrypts the information uploaded by all members and makes a network map based on the uploaded member ID;

Acquire the vertex information according to the network graph; the vertex information includes: the identity of the member, the identity information and contact information that have been confused by the member key.
The method for protecting member privacy in a social network according to claim 5, characterized in that before determining the vertex with the smallest weight among the adjacent vertices, the method further comprises:

The member counts the number of short messages sent and the number of calls with neighboring members;

Determine the number of communications according to the number of short messages sent and the number of calls, and determine the weight between members according to the number of communications;

The member uses the member key and the system public key to encrypt the weight between the neighboring members, and the server decrypts the weight to obtain the confused weight of the member key.
The method for protecting member privacy in a social network according to claim 1, wherein determining the vertex with the smallest weight among the adjacent vertices specifically comprises:

For vertices in the network connectivity graph that are not in the vertex set and edges are not in the edge set, the server recursively finds the predecessor vertices of the vertex in turn until the predecessor vertex is the current vertex, and records the number of predecessors;

Determine the weight according to the number of precursors;

The members compare the ownership values and determine the vertex corresponding to the smallest weight value.
The method for protecting member privacy in a social network according to claim 7, wherein the update process of the precursor vertex specifically includes:

Determine whether all the vertices that are not in the vertex set in the network connected graph are reachable with the current vertex. If they are not reachable, the precursor vertex is not updated. If the edge of the current vertex does not belong to the edge set, and all Vertices that are not in the set of vertices belong to the set of edges, then the current vertex is updated as a precursor vertex, and if all vertices not in the set of vertices are reachable with the current vertex, the current vertex Is the precursor vertex, and calculate the weight of the current vertex and the weight of the initial precursor vertex;

The weight value of the current vertex is compared with the weight value of the initial predecessor vertex to obtain a comparison result; the server will compare the vertex with the smaller weight value in the comparison result as the new predecessor vertex.
The method for protecting member privacy in a social network according to claim 8, wherein the server sends a precursor tuple to the applicant, and the member performs a recursive query based on the precursor vertex of the vertex corresponding to the target member, Get information about the best index path, including:

The server encrypts the identities of all members with public keys, and generates a two-tuple based on the encrypted identity and the upload key, and then sends the two-tuple to the applicant;

The member encrypts the two-tuple to form a new two-tuple;

The server uses the key to decrypt the new two-tuple;

The member determines a random number according to the first value in the decrypted two-tuple, and obtains the identity of the target member according to the second value in the decrypted two-tuple and the random number;

The member obtains an index path according to the identity of the target member, and locates the vertex of the target member in the network connectivity graph by the uniqueness of the identity;

Determine the precursor vertex corresponding to the vertex according to the vertex; use the precursor vertex to perform a recursive query to obtain an index path; the index path is the optimal index path.
The method for protecting member privacy in a social network according to claim 9, wherein the member encrypts the two-tuple to form a new two-tuple, which specifically includes:

The member shuffles the order in the two-tuple, and converts the identity information of the target member into the code unit E , encrypts the code unit E with a public key, and the first in the two-tuple The product of the two values is used as the new first value in the two-tuple; multiple random numbers are selected, and the random numbers are respectively encrypted and multiplied by the second value of the two-tuple as the two-tuple Group the new second value to form a new two-tuple.