CN106803784B

CN106803784B - Lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage

Info

Publication number: CN106803784B
Application number: CN201710200885.3A
Authority: CN
Inventors: 杨旸; 张煜超; 郑相涵; 叶少珍; 邹剑; 董晨
Original assignee: Fuzhou University
Current assignee: Fuzhou University
Priority date: 2017-03-30
Filing date: 2017-03-30
Publication date: 2020-11-27
Anticipated expiration: 2037-03-30
Also published as: CN106803784A

Abstract

The invention relates to a lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage. Private semantic search of encrypted data is achieved by using semantic information. The search keywords selected by the user do not need to be completely the same as the keywords in the outsourced encrypted multimedia data; aiming at the defect that the existing scheme only supports single-user application, the scheme introduces a searchable broadcast encryption method; encrypted multimedia files can be shared by a group of users without sharing their respective private keys; each authorized user can use a unique private key to generate a key word trapdoor of the authorized user; the user may also upload encrypted multimedia documents using the group public key as a data sender. The scheme introduces the concept of post quantum security, designs a searchable broadcast encryption scheme supporting semantic keyword search, and is constructed by utilizing a lattice-based password technology and a lattice-based agent method; based on the difficulty of the LWE (error learning) problem, the scheme achieves security against quantum attacks.

Description

Lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage

Technical Field

The invention relates to a lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage.

Background

Cloud computing has generated an increasing interest from individuals, businesses, and governments by being able to leverage shared resources to provide high quality services. The local information can be outsourced and stored to the cloud, so that huge expenditure caused by purchasing and maintaining private storage equipment is avoided, and the management cost is effectively reduced. This new computing model enables enterprises to focus on core business rather than infrastructure. Through the internet, consumers can obtain computing and storage services from the cloud anytime and anywhere.

Because the multimedia data stored in the Multimedia Cloud (MC) is stored in the clear, the internal staff of the MC service provider may sell the user's sensitive data (e.g., financial negotiation video, electronic health image, business contract photo, etc.) to a third party for benefit. User concerns about privacy of personal information become a bottleneck in the development of MC. Traditional encryption methods can guarantee confidentiality of sensitive multimedia files, but do not support information retrieval queries. It is important to support search queries in encrypted multimedia data so that a user can find documents related to certain topics from a large amount of encrypted data. Conventional encryption algorithms have been unable to meet these requirements.

Among the many challenges facing MC, the most pressing issue that prevents widespread deployment of MC is security and privacy issues. Kang et al achieve multimedia compression by a privacy-preserving compressed sensing method, and transmit compressed multimedia files using sparse coding and transmission sensing methods. Troncoso et al propose a secure signal processing technique to implement MC processing that supports privacy protection. Diaz-Sanchez et al devised a new multimedia gateway to provide privacy protection for interoperability. Want et al introduced secure sharing and digital watermarking algorithms to achieve secure protection of user data, and combined with reed-shannon coding to reduce transmission errors. Li et al utilize a semantic-based access control mechanism to implement the security services of the MC. It has also been proposed to protect the MC using a cross-secure multimedia model (2 SBM). Yang et al propose a smart card based security authentication protocol to ensure privacy of users and data.

Searchable Encryption (SE) is an important mechanism for protecting data privacy while supporting information retrieval from encrypted documents. A user of the searchable encryption system specifies keywords that are used to describe the file and encrypts the keywords into ciphertext. During information retrieval, a user submits a search request to the MC server, the request being generated from a keyword and a private key of the user. After receiving the request, the MC server searches the matched multimedia information and then sends the multimedia information to the user. In the whole process, sensitive multimedia plaintext and specified keywords must be kept secret from the server in order to protect the privacy of the user. To date, most existing SE schemes have been constructed based on number theory challenges. However, it has been demonstrated that: polynomial quantum algorithms can solve these difficult assumptions based on number theory. This means that SE schemes based on these assumptions can be broken in quantum age.

Baek et al propose a public key SE scheme without a secure channel based on bilinear pairing. In 2011, Zhang et al proposed a public key SE scheme supporting join key search, but did not give a complete security proof. Xu et al introduced a fuzzy keyword search into a public key SE system. Cao et al propose a multi-keyword ranking search and design a public key SE scheme supporting a join keyword search using a secure KNN algorithm, which are all constructed on the mathematical basis of bilinear mapping. Hwang proposes secure channel public key SE algorithm and supports join keyword search. In 2005, Regev et al proposed a difficult hypothesis against quantum attacks: error learning assumptions (LWE) and the first LWE-based cryptosystem to provide proof of security. Gentry designed an identity-based encryption scheme based on LWE. Cash proposes a method of lattice-based brokering that can generate new short bases in a secure manner using a given lattice of short bases. Since then, the two techniques described above have become the basis of many lattice-based cryptosystems, such as public key encryption schemes, hierarchical identity based encryption schemes (HIBE) and other cryptographic schemes. In 2012, Zhang et al proposed a lattice-based searchable encryption scheme, which, however, did not generate public and private keys for system users, which is a fatal disadvantage. Gu and Hou designed two lattice-searchable encryption schemes, respectively, but both schemes were designed by the scheme of Gentry, so they are very similar. Furthermore, both schemes are only applicable to single-user application scenarios. If the data owner wishes to share keyword query permissions with other users, the data owner must share its private key with authorized users. Therefore, they are not ideal for supporting multi-user systems.

Aiming at the problems that in the prior searchable encryption scheme of multimedia cloud storage, the query efficiency is low, the scheme is only suitable for a single user, quantum attack cannot be resisted and the like, the invention provides a novel security data retrieval scheme which can improve the query efficiency, is suitable for multiple users and resists quantum attack.

Disclosure of Invention

The invention aims to provide a lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage, and the method

In order to achieve the purpose, the technical scheme of the invention is as follows: a lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage comprises four entities: the method comprises the following steps that a key generation center KGC, a multimedia data sending party, a user group and a cloud server are specifically realized:

s1, when the system is established, KGC executes a KeyGen algorithm to generate a public key pk for the user group in the multimedia cloud, and generates a private key for each user group member, which is specifically expressed as: KeyGen (N) → (pk, { sk)₁,...,sk_N}), where N is the total number of users, sk₁,...,sk_NPrivate keys of N users respectively; the identity of user i is denoted as ID_i；

S2, in the encryption stage, the multimedia data sender will extract a keyword to describe the multimedia file, and at the same time, define a user group set, and the users in the set can access the multimedia file; then, the multimedia data sender executes a BEKS algorithm to generate an encrypted file;

s3, inputting a keyword to be searched by a user, generating a corresponding trapdoor by using a private key, and sending the trapdoor serving as a search request to a cloud server;

s4, after receiving the information retrieval request, the cloud server executes a Test algorithm to search a matched multimedia file; the search results will then be returned to the user.

In an embodiment of the present invention, the step S1 is implemented as follows,

taking the total number N of users as input;

s11, selecting random

S12, lattice according to TrapGen algorithm

Generating

And

s13, calculating

And

s14, lattice

Short radical of (A) is B_i←SampleBasis(A_i',B₀,S＝{1},L(1))；

S15, returning a public key for the user group: pk ═ a₀V), returning private keys for user group members: (sk)₁,...sk_N)＝(B₁,...B_N)。

In an embodiment of the present invention, the step S2 is implemented as follows,

searching keyword KW E (0, 1) by using user group public key pk^*And user set

(k is less than or equal to N) as input;

s21, searching word Net library by BEKS algorithm to create dictionary sequence synonym set of KW keywords_KW；

S22, calculating

Wherein

S23, calculating

S24, randomly selecting vectors

And σ ∈ {0,1 };

s25, calculating

Wherein x₁←χ^3km,x₂←χ；

S26, outputting ciphertext CT ═ p, c, Q_S,σ)。

In an embodiment of the present invention, the step S3 is implemented as follows,

adopting Trapdoor algorithm and using private key sk of user i_i＝B_iAnd keyword KW as input;

s31, expanding the keyword KW by using WordNet to obtain the synonym set thereof_KW；

S32, calculating

S33, lattice

Short base T of_KW,iThrough T_KW,i←SampleBasis(Q_i,B_iAnd S is {1}, L (1));

s34, Return T_KW,iTrapdoors as keywords KW, where | | T_KW,i||≤L(1)。

In an embodiment of the present invention, the step S4 is implemented as follows,

adopting a Test algorithm, and taking CT and S as input;

s41, if the user ID_iE ← GenSamplePre (Q) is calculated for S_S,Q_i,T_KW,iV, r (k +1)), where e obeys

Distributing;

s42, calculating eta ═ c-e^Tp∈Z_q；

S43, if η is close to 0, the degree is more than close

Degree of (d), μ ═ 0; otherwise, mu is 1;

s44, if μ and σ are equal, return to 1 indicates that CT includes KW, and if 0 is output, CT does not include KW.

Compared with the prior art, the invention has the following beneficial effects:

(1) searching semantic keywords: according to the scheme, the semantic keyword search of the encrypted data is realized through semantic information; the search keywords selected by the user do not need to be completely the same as the keywords in the outsourced encrypted multimedia data; if semantic relation exists between the search keywords and the encrypted keywords of the outsourced file, related multimedia documents can be retrieved;

(2) searchable broadcast encryption: in the scheme, the encrypted multimedia file can be shared by group users without sharing a private key; each authorized user can use a respective private key to generate a key word trapdoor of the authorized user; the user can also be used as a data sender to upload encrypted multimedia documents by using the group public key;

(3) the rear quantum is safe: the scheme realizes a searchable broadcast encryption scheme which is resistant to quantum attack and supports semantic keyword search; the specific algorithm utilizes a lattice-based agent method, and realizes the function of resisting quantum attack based on the difficulty of the LWE problem.

Drawings

FIG. 1 is a system framework of the method of the present invention.

Fig. 2 shows a key generation process used in the method of the present invention.

Fig. 3 is a ciphertext generation process employed by the method of the present invention.

FIG. 4 is a trapdoor creation process employed by the method of the present invention.

FIG. 5 is a document retrieval process employed by the method of the present invention.

Detailed Description

The technical scheme of the invention is specifically explained below with reference to the accompanying drawings.

As shown in fig. 1 to 5, the lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage according to the present invention includes four entities: the method comprises the following steps that a key generation center KGC, a multimedia data sending party, a user group and a cloud server are specifically realized:

Taking the total number N of users as input;

s11, selecting random

S12, lattice according to TrapGen algorithm

Generating

And

(||B₀||≤L)；

s13, calculating

And

s14, lattice

Short radical of (A) is B_i←SampleBasis(A_i',B₀,S＝{1},L(1))；

S22, calculating

Wherein

S23, calculating

S24, randomly selecting vectors

And σ ∈ {0,1 };

s25, calculating

Wherein x₁←χ^3km,x₂←χ；

S26, outputting ciphertext CT ═ p, c, Q_S,σ)。

S32, calculating

S33, lattice

Short base T of_KW,iThrough T_KW,i←SampleBasis(Q_i,B_iAnd S is {1}, L (1));

s34, Return T_KW,iTrapdoors as keywords KW, where | | T_KW,i||≤L(1)。

S4, after receiving the information retrieval request, the cloud server executes a Test algorithm to search a matched multimedia file; then, the search result is returned to the user;

adopting a Test algorithm, and taking CT and S as input;

Distributing;

s42, calculating eta ═ c-e^Tp∈Z_q；

S43, if η is close to 0, the degree is more than close

Degree of (d), μ ═ 0; otherwise, mu is 1;

The following is a specific implementation of the present invention.

FIG. 1 is a system framework of the present invention, comprising four entities: the invention discloses a Key Generation Center (KGC), a multimedia data sender, a user group and a cloud server, and particularly discloses a lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage, which is realized as follows:

(1) once the system is established, the KGC executes a KeyGen algorithm to generate a public key pk for the user group of the multimedia cloud and a private key sk for each group member. Fig. 2 shows the key generation process of the present invention. KeyGen (N) → (pk, { sk)₁,...,sk_N}): the maximum value N of the user set is taken as input. The identity of user i is denoted as ID_i。

1. Selecting random

2. TrapGen algorithm according to lem 1

Generating

And

(||B₀||≤L)。

3. computing

And

4. calculation grid

Short radical B of_i←SampleBasis(A_i',B₀,S＝{1},L(1))。

5. Returning public key p for a groupk＝(A₀V) return private key (sk) for user₁,...sk_N)＝(B₁,...B_N)。

(2) During the encryption phase, the data sender will extract a keyword to describe the multimedia file, and also define a set of users who can access the file. Then, the data sender executes the BEKS algorithm to generate an encrypted file. These encrypted files will be outsourced to the multimedia cloud for storage and processing. Fig. 3 shows the process of generating the ciphertext according to the present invention. Beeks (pk, KW, S) → CT: with the public key pk, the keyword KW belongs to {0,1}^*And user group set

(k.ltoreq.N) as input.

BEKS algorithm searches WordNet library to construct dictionary order synonym set of keyword KW_KW。

2. Computing

And

3, definition of

4. Randomly selecting a vector

And σ ∈ {0,1 }.

5. Computing

Wherein

x₂←χ。

6. Output ciphertext CT ═ p, c, Q_S,σ)。

(3) Group members define the keywords they will search and then use privacyAnd generating a corresponding trapdoor by using the key, and sending the trapdoor as a search request to the cloud server. Fig. 4 shows the creation process of the trapdoor of the present invention. Trapdoor (sk)_i,i,KW)→T_KW,i: with the private key sk of user i_i＝B_iAnd keyword KW as input.

1. Method for expanding keyword KW by using WordNet library to obtain synonym set thereof_KW。

2. Computing

3. Through T_KW,i←SampleBasis(Q_i,B_iS ═ 1, L (1)) generation lattice

Short base T of_KW,i。

4. Algorithm returns T_KW,iAs a trapdoor for keyword KW. According to theorem 1, | | T can be known_KW,i||≤L(1)。

(4) The multimedia cloud server executes a test algorithm by receiving the information retrieval request, and searches for a matched multimedia file. The search results are then sent to the group users. FIG. 5 is a process of document retrieval. Test (pk, CT, T)_KW,iI, S) → 1or 0: CT and S are used as inputs.

1. If the user ID_iE ← GenSamplePre (Q) is calculated for S_S,Q_i,T_KW,iV, r (k + 1)). Obey 2, e

And (4) distribution.

2. Calculating eta ═ c-e^Tp∈Z_q。

3. If η is closer to 0 than to

Degree of (c), μ ═ 0. Otherwise, μ ═ 1.

4. If μ and σ are equal, the algorithm returns 1, indicating that CT contains KW. If 0 is output, CT does not contain KW.

In the method, the explanation of WordNet, lattice and related theorem and lemma is as follows:

1、WordNet：

WordNet is an english vocabulary database. It divides synonyms of the english vocabulary into different subsets. Two words are considered synonyms if they have at least one common meaning. A synonym set of words is called a "synonym set". Therefore, we can also consider WordNet as an integration of english dictionary and synonyms. WordNet constructs a synonym set for nouns, verbs, adjectives and adverbs, but omits prepositions and qualifiers. Here we will construct a synonym set for the keyword KW using WordNet: consisting of a keyword KW and its synonyms. The collection is then rearranged to dictionary order S' ═ S (S)₁,...,s_n) So as to facilitate the use. Use of synonym sets for KW_KWAnd (KW, S').

2. Grid:

A＝[a₁,...,a_n]consisting of n linearly independent vectors, with an n-dimensional lattice generated by the matrix a, denoted as Λ ═ Ac ═ Σ_{i∈{1,...,n}}c_ia_i,c_iE.g. z), two patterns are defined

And

wherein q is a prime number and

| a | | represents the norm of the longest column of matrix a,

Gram-Schmidt orthogonal matrix representing matrix a.

RⁿThe upper n-dimensional gaussian function centered around c is: rho_σ,c(x)＝exp(-π||x-c||²/σ²) And ρ_σ,_c(Λ)＝∑_xρ_σ,c(x) (σ > 0). Of ΛDiscrete Gaussian distribution of D_Λ,σ(x)＝ρ_σ,c(x)/ρ_σ,c(Λ)。

Definition 1: let n, m, q be positive integers. X is

Is distributed from

S is randomly extracted. Random decimation

And x ∈ χ, A_s,χObey { A, A^Ts + x (modq) }. Deterministic LWE (error learning) problem should be A_s,χAnd

are distinguished from random samples of (a).

Introduction 1: a probability polynomial time algorithm TrapGen exists, positive integers n, m and q (q is more than or equal to 2, m is more than or equal to 5nlgq) are taken as input and output

Such that A is statistically

In a uniform distribution, B is a lattice

A base of (A) and

has a probability of n^ω(1)。

Definition matrix

And a ═ a₁,...A_k]Wherein

For the

S＝{i₁,...,i_j}, definition of A_STo represent

The following theorem can be used with lattices

Middle generation lattice

Theorem 1: assuming n, q, m, k (n, q ≧ 2, m ≧ 2nlgq) is a positive integer, the input matrix

Collection

Group B of_SAnd an integer

There is a probabilistic polynomial time algorithm samplebalis output B ← samplebalis (a, B)_SS, L) in which B is

A base of (a) and

theorem 2: assuming n, q, m, k (n, q ≧ 2, m ≧ 2nlgq) is a positive integer, the input matrix

Collection

Group B of_SVector of motion

And integer of

There is one probability polynomial time algorithm, GenSamplePre output e ← GenSamplePre (a, B)_SS, y, r), wherein the distribution of e obeys

The application of the invention is as follows: by utilizing the multimedia cloud, a user can store own multimedia files in the cloud so as to reduce inconvenience brought by local storage and maintenance of the multimedia files. In order to guarantee data security and personal privacy, a user can encrypt part of sensitive data and then store the encrypted data in a cloud server. When the data needs to be used, the user can use the method and the system to perform keyword retrieval on the cloud data. Semantically related documents may also be matched when an authorized user wishes to search for documents that are semantically related to the query keyword, or fails to enter an accurate keyword for various reasons. Meanwhile, the invention supports a multi-user broadcast encryption mechanism without a shared key and can resist quantum attack.

The above are preferred embodiments of the present invention, and all changes made according to the technical scheme of the present invention that produce functional effects do not exceed the scope of the technical scheme of the present invention belong to the protection scope of the present invention.

Claims

1. A lattice-based multi-user fuzzy searchable encryption method in secure multimedia cloud storage is characterized by comprising the following steps: the method comprises four entities: the method comprises the following steps that a key generation center KGC, a multimedia data sending party, a user group and a cloud server are specifically realized:

s1, when the system is established, the KGC executes a KeyGen algorithm to be the multimedia cloudThe user group in (1) generates a public key pk, and generates a private key for each user group member, which is specifically represented as: KeyGen (N) → (pk, { sk)₁,...,sk_N}), where N is the total number of users, sk₁,...,sk_NPrivate keys of N users respectively; the identity of user i is denoted as ID_i；

s4, after receiving the information retrieval request, the cloud server executes a Test algorithm to search a matched multimedia file; the search results will then be returned to the user;

the specific implementation process of the keygen (n) algorithm in S1 is as follows,

taking the total number N of users as input;

s11, selecting random