CN108363689A - Secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud and system - Google Patents
Secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud and system Download PDFInfo
- Publication number
- CN108363689A CN108363689A CN201810122376.8A CN201810122376A CN108363689A CN 108363689 A CN108363689 A CN 108363689A CN 201810122376 A CN201810122376 A CN 201810122376A CN 108363689 A CN108363689 A CN 108363689A
- Authority
- CN
- China
- Prior art keywords
- document
- vector
- keyword
- retrieval
- cloud server
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/93—Document management systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/04—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks
- H04L63/0428—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks wherein the data content is protected, e.g. by encrypting or encapsulating the payload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/06—Protocols specially adapted for file transfer, e.g. file transfer protocol [FTP]
Abstract
The invention discloses a kind of secret protection multi-key word Top k cipher text retrieval methods and system towards mixed cloud.Present invention mainly solves the low problems of recall precision.Its scheme is:Data provide end and generate keyword dictionary sequence by clustering strategies using the correlativity between keyword;The document filter vectors of the document vector sum low-dimensional of each document structure tree higher-dimension, are then outsourced to incredible publicly-owned Cloud Server by ciphertext document and encrypted document vector, by the document filter vectors storage of plaintext to believable privately owned Cloud Server.When retrieval, candidate documents are calculated by privately owned Cloud Server first, retrieval result Top k document calculations are then realized by publicly-owned Cloud Server.Associative key aggregation properties improve the effect of privately owned Cloud Server filtering in keyword dictionary sequence, have compressed the size of candidate documents.Flow of the present invention is simple, safe, it is easy to accomplish, it can realize that efficient multi-key word searching ciphertext is handled by less computing cost in mixing cloud environment.
Description
Technical field
The present invention relates to user data secret protection more particularly to a kind of secret protection multi-key words towards mixed cloud
Top-k cipher text retrieval methods and system.
Background technology
The thought of IT resource services becomes increasingly popular, and becoming for " all are all serviced " (X as a Service, XaaS) is presented
Gesture, " service " become the key concept of cloud computing.However while cloud computing flourishes, cloud security also becomes to be closed extensively
The problem of note.In cloud environment, it is placed in remote cloud server (Cloud Server, CS) since user can not directly control
Data, worry that the outer bag data of oneself is illegally obtained or abused by cloud service provider, especially for privacy require compared with
High sensitive data, such as electronic health record, bank transaction data, user mail etc..Although cloud service provider declares their meetings
Some Security Countermeasures are provided to cope with privacy leakage problem, such as access control technology, firewall technology and Intrusion Detection Technique
Deng, but user is unquestionably the main problem for restricting cloud computing and further developing to the worry of problem of data safety.
Protection data-privacy a kind of common practice be outsourced to publicly-owned Cloud Server again after handling data encryption, but
It is the use for seriously constraining outer bag data in this way.In Research into information retrieval field, existing keywords-based retrieval is mainly
Towards clear data, searching ciphertext field can not be applied directly to.And by all encryption datas from high in the clouds it is locally downloading into
Row decryption is clearly a kind of unrealistic and the wasting of resources processing method.Therefore, it studies and solves have privacy in cloud environment
The ciphertext data retrieval mechanism of defencive function is a challenge, this also has become cloud computing research field concern in recent years
One of hot issue.
Big multi-method acquiescence is all to use public cloud service in the prior art, based on public cloud according to " semi-honesty model " side
Formula, which provides, services this hypothesis, multi-key word searching ciphertext processing method in a series of encryption cloud environments of proposition, but these
One or more problems such as that there are recall precisions is low for method, retrieval result is inaccurate, index tree structure complexity.
For these problems, the Chinese invention patent application of application number 201710181664.6 discloses in a kind of cloud computing
The quick multi-key word Semantic Ranking searching method for protecting data-privacy, by the way that privately owned Cloud Server is added, for each document
While creating document vector, corresponding mark vector is created, encrypted document vector is outsourced to publicly-owned Cloud Server, it will
The storage of clear identification vector realizes the preliminary filter operation of document sets to privately owned Cloud Server, by privately owned Cloud Server, reduce with
The document vector number that vector's correlation degree score calculates is retrieved, reduces retrieval computing cost, but this method is due to key word character
The distribution of keyword is random in allusion quotation, and privately owned Cloud Server filter effect is bad, causes to need to calculate a large amount of texts in public cloud
Relevance score between shelves vector sum retrieval vector.Therefore privately owned Cloud Server filter effect how is improved, it is mixed for improving
Closing in cloud supports the multi-key word searching ciphertext efficiency of secret protection to play an important role.
Invention content
Goal of the invention:In view of the problems of the existing technology the present invention, it is more to provide a kind of secret protection towards mixed cloud
Keyword Top-k cipher text retrieval methods and system, the present invention can effectively realize the secret protection of user data, promote close more
The efficiency of keyword searching ciphertext realizes that accurately Top-k is retrieved.
Technical solution:Secret protection multi-key word Top-k cipher text retrieval method packets of the present invention towards mixed cloud
It includes:
(1) data provide end and extract keyword set from the document sets of offer, and generate keyword by the piecemeal that clusters
Dictionary sequence;It is one corresponding plain text document of each document structure tree vector in document sets further according to keyword dictionary sequence, and according to
According to the piecemeal situation of keyword dictionary sequence by plain text document vector piecemeal, document filter vectors are formed;Later by plain text document
Vector encryption forms encrypted document vector, and each file encryption in document sets is formed encrypted document collection;Finally by document mistake
Filter vector is transmitted to privately owned Cloud Server, and encrypted document vector sum encrypted document collection is transmitted to publicly-owned Cloud Server;
(2) data retrieval end generates retrieval vector according to multiple keywords that user provides, using peace after being normalized
Full algorithm generates retrieval trapdoor, needs the document number k retrieved to be transmitted to publicly-owned Cloud Server together with user;Again according to keyword
The piecemeal situation of keyword in dictionary sequence, multiple keywords that user is provided generate retrieval filter vectors, and are transmitted to private
There is Cloud Server;
(3) privately owned Cloud Server carries out the retrieval filter vectors received and the document filter vectors of each document respectively
With operation, if all positions of vector that operation obtains are not all 0, corresponding document code, which is added to candidate documents, to be concentrated, and
Candidate documents are transmitted to publicly-owned Cloud Server;
(4) publicly-owned Cloud Server is counted respectively according to the candidate documents, retrieval trapdoor and the search file number k that receive
Calculating candidate documents concentrates the corresponding encrypted document vector sum of each document to retrieve the safe inner product between trapdoor, according to safe inner product
It chooses candidate documents and the maximally related k ciphertext document of keyword provided with user is provided, which is back to number
According to retrieval end;
(5) k ciphertext document of reception is decrypted in data retrieval end, obtains maximally related k plain text document.
Further, step (1) specifically includes:
(1-1) data provide end and extract keyword from the document sets DS of offer, obtain keyword set { w1,w2,…,
wn};
Keyword root in keyword set is carried out the operation that clusters by (1-2) according to correlativity, obtains several cluster
Cluster { c1,c2,…,ct};
(1-3) using each submanifold as a piecemeal, to obtain t piecemeal, respectively b1,b2,…,bt, further according to point
Block generates keyword dictionary sequence W={ w (b1,1),w(b1,2),…,w(b2,1),w(b2,2),…,w(bt,1),w(bt,
2) ... }, wherein w (bj, x) and it indicates to belong to piecemeal bjIn x-th of keyword, each keyword in the block is unordered;Piecemeal bj=
{w(bj,x)|0<x≤|bj|};
(1-4) is according to the position of keyword in keyword dictionary sequence using TF-IDF algorithms and vector space model
Document sets DS={ D1,D2,…,DmIn each document DiGenerate a corresponding plain text document vector Vi, and place is normalized
Reason;Wherein, ViDimension be n, every value is the corresponding keyword of this in document DiIn word frequency TF values;
(1-5) according to the piecemeal situation of keyword dictionary sequence, by plain text document vector ViIt is divided into t piecemeal, piecemeal side
Boundary is identical with the piecemeal boundary of keyword dictionary sequence, obtains each document DiDocument filter vectors DFi={ b1,b2,…,
bt};Wherein, if ViMiddle piecemeal bjPosition value where corresponding all keywords is all 0, then bjThe value of block is 0, otherwise bj
The value of block is 1, DFiIt is the vector that every value of t dimensions is 0/1;
(1-6) generates encryption key SK (S, M1,M2,kf);Wherein, S is the random vector that every value is 0/1, M1
And M2It is two n × n invertible matrix, n is the length of keyword dictionary sequence, kfIt is document encryption key;
(1-7) is by safe KNN technologies using the encryption key generated to each plain text document vector ViEncryption, obtains pair
The encrypted document vector answeredWherein, as j-th of element S [j]=0 in random vector S, Vi′+Vi″
=Vi, as S [j]=1, Vi'=Vi"=Vi;
(1-8) obtains encrypted document collection ES={ e by each document in symmetric encipherment algorithm encrypted document collection DS1,
e2,…,em};
Document filter vectors are transmitted to privately owned Cloud Server and stored by (1-9), and encrypted document vector sum is encrypted text
Shelves collection is transmitted to publicly-owned Cloud Server and is stored.
Further, step (2) specifically includes:
Multiple keyword { w that (2-1) data retrieval end is provided according to user1,w2,…,wx, using TF-IDF algorithms and
Vector space model generates retrieval vector Q, and is normalized;Wherein, j-th of element Q [j] of Q is jth position keyword wj
Data provide the inverse document frequency IDF values in the document sets DS that end provides;
(2-2) is based on retrieval vector Q, using safe KNN algorithms, generates retrieval trapdoorWherein,
Wherein when data provide jth position S [j]=0 of the random vector in the encryption key that end generates, Q ' [j]=Q " [j]=Q
[j], as S [j]=1, Q ' [j]+Q " [j]=Q [j];
(2-3) generates retrieval filter vectors QF, QF={ b according to the piecemeal situation of keyword in keyword dictionary sequence1,
b2,…,bt, QF is the vector that t ties up every value 0/1, if piecemeal bjCorresponding all keywords are corresponding in retrieving vector Q
All values are all 0, then QF [j]=0, otherwise QF [j]=1;
(2-4) will retrieve trapdoor and user needs the document number k retrieved to be transmitted to publicly-owned Cloud Server, and retrieval is filtered
Vector is transmitted to privately owned Cloud Server.
Further, step (3) specifically includes:
(3-1) privately owned Cloud Server is by the document filter vectors DF of the retrieval filter vectors QF received and each documenti
Progress and operation respectively, if QF&DFiThe obtained all positions of vector of operation be not all 0, then by DFiCorresponding document code Didi
It is added to candidate documents concentration, obtains candidate documents CDS={ d1,d2,…};
Candidate documents CDS is sent to publicly-owned Cloud Server by (3-2).
Secret protection multi-key word Top-k searching ciphertext systems of the present invention towards mixed cloud, including data carry
For end, data retrieval end, privately owned Cloud Server and publicly-owned Cloud Server, wherein:
Data provide end and generate keyword for extracting keyword set from the document sets of offer, and by the piecemeal that clusters
Dictionary sequence;And it is vectorial for one corresponding plain text document of each document structure tree in document sets according to keyword dictionary sequence, and
Plain text document vector piecemeal is formed document filter vectors by the piecemeal situation according to keyword dictionary sequence;And it will be literary in plain text
Shelves vector encryption forms encrypted document vector, and each file encryption in document sets is formed encrypted document collection;And by document
Filter vectors are transmitted to privately owned Cloud Server, and encrypted document vector sum encrypted document collection is transmitted to publicly-owned Cloud Server;
Data retrieval end is used to generate retrieval vector according to multiple keywords that user provides, using peace after being normalized
Full algorithm generates retrieval trapdoor, needs the document number k retrieved to be transmitted to publicly-owned Cloud Server together with user;And according to key
The piecemeal situation of keyword in word dictionary sequence, multiple keywords that user is provided generate retrieval filter vectors, and are transmitted to
Privately owned Cloud Server;
The document filter vectors of retrieval filter vectors and each document that privately owned Cloud Server is used to receive respectively into
Row and operation, if all positions of vector that operation obtains are not all 0, corresponding document code, which is added to candidate documents, to be concentrated,
And candidate documents are transmitted to publicly-owned Cloud Server;
Publicly-owned Cloud Server is used to, according to the candidate documents, retrieval trapdoor and the search file number k that receive, count respectively
Calculating candidate documents concentrates the corresponding encrypted document vector sum of each document to retrieve the safe inner product between trapdoor, according to safe inner product
It chooses candidate documents and the maximally related k ciphertext document of keyword provided with user is provided, which is back to number
According to retrieval end;
Data retrieval end is additionally operable to that k ciphertext document of reception is decrypted, and obtains maximally related k plain text document.
Further, the data provide end and specifically include:
Keyword extracting module obtains keyword set { w for extracting keyword from the document sets DS of offer1,
w2,…,wn};
Cluster module, for the keyword root in keyword set to be carried out the operation that clusters according to correlativity, obtains several
A cluster submanifold { c1,c2,…,ct};
Keyword dictionary generation module is used for using each submanifold as a piecemeal, to obtain t piecemeal, respectively
b1,b2,…,bt, keyword dictionary sequence W={ w (b are generated further according to piecemeal1,1),w(b1,2),…,w(b2,1),w(b2,
2),…,w(bt,1),w(bt,2),…};Wherein w (bj, x) and it indicates to belong to piecemeal bjIn x-th of keyword, it is each in the block
Keyword is unordered;Piecemeal bj={ w (bj,x)|0<x≤|bj|};
Plain text document vector generation module, for using TF-IDF algorithms and vector space model, according to keyword dictionary
The position of keyword in sequence is document sets DS={ D1,D2,…,DmIn each document DiGenerate a corresponding plain text document
Vectorial Vi, and be normalized;Wherein, ViDimension be n, every value is the corresponding keyword of this in document DiIn
Word frequency TF values;
Document filter vectors generation module, for the piecemeal situation according to keyword dictionary sequence, by plain text document vector
ViIt is divided into t piecemeal, piecemeal boundary is identical with the piecemeal boundary of keyword dictionary sequence, obtains each document DiDocument filtering
Vectorial DFi={ b1,b2,…,bt};Wherein, if ViMiddle piecemeal bjPosition value where corresponding all keywords is all 0, then
bjThe value of block is 0, otherwise bjThe value of block is 1, DFiIt is the vector that every value of t dimensions is 0/1;
Key production module, for generating encryption key SK (S, M1,M2,kf);Wherein, it is 0/1 that S, which is every value,
Random vector, M1And M2It is two n × n invertible matrix, n is the length of keyword dictionary sequence, kfIt is document encryption key;
Document vector encrypting module, for using the encryption key generated to each plain text document by safe KNN technologies
Vectorial ViEncryption obtains corresponding encrypted document vectorWherein, when j-th of element S in random vector S
When [j]=0, Vi′+Vi"=Vi, as S [j]=1, Vi'=Vi"=Vi;
File encryption module, for by each document in symmetric encipherment algorithm encrypted document collection DS, obtaining encrypted document
Collect ES={ e1,e2,…,em};
Transmission module is stored for document filter vectors to be transmitted to privately owned Cloud Server, by encrypted document vector
Publicly-owned Cloud Server is transmitted to encrypted document collection to be stored.
Further, the data retrieval end specifically includes:
Vector generation module is retrieved, multiple keyword { w for being provided according to user1,w2,…,wx, using TF-IDF
Algorithm and vector space model generate retrieval vector Q, and are normalized;Wherein, j-th of element Q [j] of Q is that jth position is crucial
Word wjInverse document frequency IDF values in data provide the document sets DS that end provides;
Trapdoor generation module is retrieved, for generating retrieval trapdoor using safe KNN algorithms based on retrieval vector QWherein, wherein when data provide end generate encryption key in random vector jth position S [j]=
When 0, Q ' [j]=Q " [j]=Q [j], as S [j]=1, Q ' [j]+Q " [j]=Q [j];
Filter vectors generation module is retrieved, for the piecemeal situation according to keyword in keyword dictionary sequence, generates inspection
Rope filter vectors QF, QF={ b1,b2,…,bt, QF is the vector that t ties up every value 0/1, if piecemeal bjCorresponding all keys
Word corresponding all values in retrieving vector Q are all 0, then QF [j]=0, otherwise QF [j]=1;
Transmission module needs the document number k retrieved to be transmitted to publicly-owned Cloud Server for will retrieve trapdoor and user, will
Retrieval filter vectors are transmitted to privately owned Cloud Server.
Further, the privately owned Cloud Server specifically includes:
With computing module, the document filter vectors DF of retrieval filter vectors QF and each document for that will receiveiPoint
It Jin Hang not be with operation, if QF&DFiThe obtained all positions of vector of operation be not all 0, then by DFiCorresponding document code DidiAdd
It is added to candidate documents concentration, obtains candidate documents CDS={ d1,d2,…};
Transmission module, for candidate documents CDS to be sent to publicly-owned Cloud Server.
Advantageous effect:Compared with prior art, the present invention its remarkable advantage is:
1, safe
The present invention is accomplished that in insincere publicly-owned cloud environment, is realized and is carried out searching ciphertext according to multi-key word, is led to
Cross safe KNN technologies and realize that safe inner product calculates, may be implemented the inner product value between two encryption vectors be equal to two plaintexts to
Inner product value between amount, in publicly-owned cloud environment need not to retrieval trapdoor be decrypted, also need not to encrypted document to
Operation is decrypted in amount, more operation need not be decrypted to encrypted document.In public cloud part, whole process is all under ciphertext
Carry out operation, finally obtained Top-k results.Therefore, the utilization of safe KNN technologies can be realized carries out according to multi-key word
While Top-k retrieval results calculate, the data-privacy of data owner is protected.Safe KNN technologies have been widely used for more
Keyword searching ciphertext field.
2, accuracy is high
The secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud that the present invention provides, in data retrieval
When the end interested multi-key word of offer is retrieved, it is divided into what two steps carried out, Cloud Server privately owned first, which generates, waits
Document sets CDS is selected, is then tied with the maximally related Top-k of interested multi-key word in public cloud server search candidate documents
Fruit.When private clound generates candidate documents, for any document D in whole document sets DSiAs long as including one or more data
The interested multi-key word that user provides can all be added to candidate documents, therefore there is no the Top-k for the condition that meets
Document is not the candidate documents the case where;And when publicly-owned Cloud Server obtains the candidate documents that privately owned Cloud Server sends over
When, it is that the encrypted document vector sum for each document strictly concentrated according to candidate documents retrieves the inner product result of calculation between trapdoor
Obtained Top-k is as a result, therefore, the mode of the privately owned Cloud Server of the present invention and public cloud server collaboration can be to retrieval result
It is accurately sorted and returns to Top-k documents and give data retrieval end as retrieval result.
3, recall precision is high
The secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud that the present invention provides, for current peace
The inefficient problem of encryption method is can search for based on the Technology applications such as full KNN calculating, TF-IDF and vector space model, is led to
It crosses and adds credible privately owned Cloud Server, give the method that document vector piecemeal generates document filter vectors, document is filtered
Vector uploads to privately owned Cloud Server, because the dimension of document filter vectors is smaller, privately owned Cloud Server can make according to data
The retrieval filter vectors that user provides, can be obtained by candidate documents, fast filtering falls largely by less computing overhead
Irrelevant document (these documents filtered out are unlikely to be final Top-k results), candidate documents are relative to original document collection
For be much smaller, therefore publicly-owned Cloud Server only need to carry out the inner product between a small amount of encryption vector calculating, can save significantly
Save the computing cost of publicly-owned Cloud Server.In addition to this, often relevant in view of interested multi-key word input by user
The fact that, in order to improve the filter effect of privately owned Cloud Server, further compressed candidature document sets, in keyword dictionary sequence
The position of keyword is not to put at random, but clustered according to keyword correlativity, then obtains multiple submanifolds,
Keyword is all located at the document filter vectors in identical piece, generated in this way in keyword dictionary sequence and retrieval in each submanifold
Filter vectors are more advantageous to the size of compressed candidature document sets, and document number tails off in candidate documents, then publicly-owned Cloud Server
It is middle that the number for calculating the inner product between encryption vector is needed just to become smaller, computing cost is reduced naturally, and it is close to improve multi-key word
Literary effectiveness of retrieval.
Description of the drawings
Fig. 1 is the framework of the secret protection multi-key word Top-k cipher text retrieval methods provided by the invention towards mixed cloud
Figure;
Fig. 2 is the flow of the secret protection multi-key word Top-k cipher text retrieval methods provided by the invention towards mixed cloud
Schematic diagram;
Fig. 3 be keyword cluster construction keyword dictionary sequence schematic diagram, it is illustrated that middle keyword set clusters into 10 small
Cluster, corresponding keyword dictionary sequence are 10 blocks, include that keyword number is indefinite, and corresponds to keyword number in tuftlet in each block
It measures identical;
Fig. 4 is the document vector sum document filter vectors and retrieval vector sum retrieval filtering before not doing normalized
The schematic diagram of vector;
Fig. 5 is retrieval flow schematic diagram, wherein Cloud Server privately owned first by document filter vectors and retrieval filtering to
Between amount and operation obtains candidate documents, and candidate documents are then sent to publicly-owned Cloud Server;Publicly-owned Cloud Server
It is concentrated in candidate documents, Top-k documents is obtained by calculating the relevance score between document vector sum retrieval vector.In order to draw
Figure terseness, document vector sum retrieval vector is not normalized and encryption.
Specific implementation mode
Embodiment 1
Present embodiments provide a kind of secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud, such as Fig. 1
Shown in Fig. 2, include the following steps:
(1) data provide end and extract keyword set from the document sets of offer, and generate keyword by the piecemeal that clusters
Dictionary sequence;It is one corresponding plain text document of each document structure tree vector in document sets further according to keyword dictionary sequence, and according to
According to the piecemeal situation of keyword dictionary sequence by plain text document vector piecemeal, document filter vectors are formed;Later by plain text document
Vector encryption forms encrypted document vector, and each file encryption in document sets is formed encrypted document collection;Finally by document mistake
Filter vector is transmitted to privately owned Cloud Server, and encrypted document vector sum encrypted document collection is transmitted to publicly-owned Cloud Server.
The step specifically includes:
(1-1) data provide end and extract keyword from the document sets DS of offer, obtain keyword set { w1,w2,…,
wn};
Keyword root in keyword set is carried out the operation that clusters by (1-2) according to correlativity, obtains several cluster
Cluster { c1,c2,…,ct};
(1-3) using each submanifold as a piecemeal, to obtain t piecemeal, respectively b1,b2,…,bt, further according to point
Block generates keyword dictionary sequence W={ w (b1,1),w(b1,2),…,w(b2,1),w(b2,2),…,w(bt,1),w(bt,
2) ... }, wherein w (bj, x) and it indicates to belong to piecemeal bjIn x-th of keyword, each keyword in the block is unordered;Piecemeal bj=
{w(bj,x)|0<x≤|bj|};The characteristic because keyword clusters, the strong keyword of correlation is gathered in keyword dictionary sequence
In same piece.Such as in Fig. 3, keyword set clusters into altogether 10 small natural cluster, and the keyword quantity in each cluster is not
Fixed, then keyword dictionary includes 10 keyword blocks, the keyword quantity for including in each keyword block and includes in corresponding cluster
Keyword quantity is identical, then according to these blocks, generates keyword dictionary sequence;
(1-4) is according to the position of keyword in keyword dictionary sequence using TF-IDF algorithms and vector space model
Document sets DS={ D1,D2,…,DmIn each document DiGenerate a corresponding plain text document vector Vi, and place is normalized
Reason;Wherein, ViDimension be n, every value is the corresponding keyword of this in document DiIn word frequency TF values;
(1-5) according to the piecemeal situation of keyword dictionary sequence, by plain text document vector ViIt is divided into t piecemeal, piecemeal side
Boundary is identical with the piecemeal boundary of keyword dictionary sequence, obtains each document DiDocument filter vectors DFi={ b1,b2,…,
bt};Wherein, if ViMiddle piecemeal bjPosition value where corresponding all keywords is all 0, then bjThe value of block is 0, otherwise bj
The value of block is 1, DFiIt is the vector that every value of t dimensions is 0/1;Such as in Fig. 4, gives document vector sum and correspond to document
The specific example of filter vectors, it is contemplated that picture terseness is not normalized document vector, document Di, right
It is V to answer document vectori, according to the position of block boundary in keyword dictionary sequence, the document filter vectors DF of compositioniSuch as institute in Fig. 4
Show;
(1-6) generates encryption key SK (S, M1,M2,kf);Wherein, S be every value be 0/1 n tie up random column to
Amount, M1And M2It is two n × n invertible matrix, n is the length of keyword dictionary sequence, kfIt is document encryption key;SK is only provided
DO, DU are used, to CS secrets.
(1-7) is by safe KNN technologies using the encryption key generated to each plain text document vector ViEncryption, obtains pair
The encrypted document vector answeredWherein, as j-th of element S [j]=0 in random vector S, Vi′+Vi″
=Vi, as S [j]=1, Vi'=Vi"=Vi;
(1-8) obtains encrypted document collection ES={ e by each document in symmetric encipherment algorithm encrypted document collection DS1,
e2,…,em};
Document filter vectors are transmitted to privately owned Cloud Server and stored by (1-9), and encrypted document vector sum is encrypted text
Shelves collection is transmitted to publicly-owned Cloud Server and is stored.
(2) data retrieval end generates retrieval vector according to multiple keywords that user provides, using peace after being normalized
Full algorithm generates retrieval trapdoor, needs the document number k retrieved to be transmitted to publicly-owned Cloud Server together with user;Again according to keyword
The piecemeal situation of keyword in dictionary sequence, multiple keywords that user is provided generate retrieval filter vectors, and are transmitted to private
There is Cloud Server.
The step specifically includes:
Multiple keyword { w that (2-1) data retrieval end is provided according to user1,w2,…,wx, using TF-IDF algorithms and
Vector space model generates retrieval vector Q, and is normalized;Wherein, j-th of element Q [j] of Q is jth position keyword wj
Data provide the inverse document frequency IDF values in the document sets DS that end provides;
(2-2) is based on retrieval vector Q, using safe KNN algorithms, generates retrieval trapdoorWherein,
Wherein when data provide jth position S [j]=0 of the random vector in the encryption key that end generates, Q ' [j]=Q " [j]=Q
[j], as S [j]=1, Q ' [j]+Q " [j]=Q [j];
(2-3) generates retrieval filter vectors QF, QF={ b according to the piecemeal situation of keyword in keyword dictionary sequence1,
b2,…,bt, QF is the vector that t ties up every value 0/1, if piecemeal bjCorresponding all keywords are corresponding in retrieving vector Q
All values are all 0, then QF [j]=0, otherwise QF [j]=1;
(2-4) will retrieve trapdoor and user needs the document number k retrieved to be transmitted to publicly-owned Cloud Server, and retrieval is filtered
Vector is transmitted to privately owned Cloud Server.
(3) privately owned Cloud Server carries out the retrieval filter vectors received and the document filter vectors of each document respectively
With operation, if all positions of vector that operation obtains are not all 0, corresponding document code, which is added to candidate documents, to be concentrated, and
Candidate documents are transmitted to publicly-owned Cloud Server.
The step specifically includes:
(3-1) privately owned Cloud Server is by the document filter vectors DF of the retrieval filter vectors QF received and each documenti
Progress and operation respectively, if QF&DFiThe obtained all positions of vector of operation be not all 0, then by DFiCorresponding document code Didi
It is added to candidate documents concentration, obtains candidate documents CDS={ d1,d2,…};
Candidate documents CDS is sent to publicly-owned Cloud Server by (3-2).Fig. 5 provides a specific query example, privately owned
Cloud Server is by retrieval filter vectors and document filter vectors are done and operation, finding the document mistake that operation result is not all 0
The corresponding document code of filter vector, to obtain candidate documents CDS={ Did1,Did5,Did6, CDS is then sent to public affairs
There is Cloud Server.
(4) publicly-owned Cloud Server is counted respectively according to the candidate documents, retrieval trapdoor and the search file number k that receive
Calculating candidate documents concentrates the corresponding encrypted document vector sum of each document to retrieve the safe inner product between trapdoor, according to safe inner product
It chooses candidate documents and the maximally related k ciphertext document of keyword provided with user is provided, which is back to number
According to retrieval end.
For example, in Fig. 5, publicly-owned Cloud Server receives the candidate documents CDS=CDS=that privately owned Cloud Server is sent
{Did1,Did5,Did6, the search space of Top-k documents has not been complete or collected works DS={ D at this time1,D2,…,D10, and become
Candidate documents { D1,D5,D6, search space becomes 3 present documents by 10 original documents, therefore only needs to count
The inner product operation between 3 vectors is calculated, at this time by calculating the corresponding encrypted document of each document one by one in candidate documents concentration
Dot product between vector sum encryption retrieval vector selects maximum k corresponding encryption texts to obtain 3 relevance scores
Shelves, return to data retrieval end.
(5) k ciphertext document of reception is decrypted using symmetric key for data retrieval end, obtains maximally related k
Plain text document.
Embodiment 2
A kind of secret protection multi-key word Top-k searching ciphertext systems towards mixed cloud are present embodiments provided, including
Data provide end, data retrieval end, privately owned Cloud Server and publicly-owned Cloud Server, wherein:
Data provide end and generate keyword for extracting keyword set from the document sets of offer, and by the piecemeal that clusters
Dictionary sequence;And it is vectorial for one corresponding plain text document of each document structure tree in document sets according to keyword dictionary sequence, and
Plain text document vector piecemeal is formed document filter vectors by the piecemeal situation according to keyword dictionary sequence;And it will be literary in plain text
Shelves vector encryption forms encrypted document vector, and each file encryption in document sets is formed encrypted document collection;And by document
Filter vectors are transmitted to privately owned Cloud Server, and encrypted document vector sum encrypted document collection is transmitted to publicly-owned Cloud Server;
Data retrieval end is used to generate retrieval vector according to multiple keywords that user provides, using peace after being normalized
Full algorithm generates retrieval trapdoor, needs the document number k retrieved to be transmitted to publicly-owned Cloud Server together with user;And according to key
The piecemeal situation of keyword in word dictionary sequence, multiple keywords that user is provided generate retrieval filter vectors, and are transmitted to
Privately owned Cloud Server;
The document filter vectors of retrieval filter vectors and each document that privately owned Cloud Server is used to receive respectively into
Row and operation, if all positions of vector that operation obtains are not all 0, corresponding document code, which is added to candidate documents, to be concentrated,
And candidate documents are transmitted to publicly-owned Cloud Server;
Publicly-owned Cloud Server is used to, according to the candidate documents, retrieval trapdoor and the search file number k that receive, count respectively
Calculating candidate documents concentrates the corresponding encrypted document vector sum of each document to retrieve the safe inner product between trapdoor, according to safe inner product
It chooses candidate documents and the maximally related k ciphertext document of keyword provided with user is provided, which is back to number
According to retrieval end;
Data retrieval end is additionally operable to that k ciphertext document of reception is decrypted, and obtains maximally related k plain text document.
Further, the data provide end and specifically include:
Keyword extracting module obtains keyword set { w for extracting keyword from the document sets DS of offer1,
w2,…,wn};
Cluster module, for the keyword root in keyword set to be carried out the operation that clusters according to correlativity, obtains several
A cluster submanifold { c1,c2,…,ct};
Keyword dictionary generation module is used for using each submanifold as a piecemeal, to obtain t piecemeal, respectively
b1,b2,…,bt, keyword dictionary sequence W={ w (b are generated further according to piecemeal1,1),w(b1,2),…,w(b2,1),w(b2,
2),…,w(bt,1),w(bt,2),…};Wherein w (bj, x) and it indicates to belong to piecemeal bjIn x-th of keyword, it is each in the block
Keyword is unordered;Piecemeal bj={ w (bj,x)|0<x≤|bj|};
Plain text document vector generation module, for using TF-IDF algorithms and vector space model, according to keyword dictionary
The position of keyword in sequence is document sets DS={ D1,D2,…,DmIn each document DiGenerate a corresponding plain text document
Vectorial Vi, and be normalized;Wherein, ViDimension be n, every value is the corresponding keyword of this in document DiIn
Word frequency TF values;
Document filter vectors generation module, for the piecemeal situation according to keyword dictionary sequence, by plain text document vector
ViIt is divided into t piecemeal, piecemeal boundary is identical with the piecemeal boundary of keyword dictionary sequence, obtains each document DiDocument filtering
Vectorial DFi={ b1,b2,…,bt};Wherein, if ViMiddle piecemeal bjPosition value where corresponding all keywords is all 0, then
bjThe value of block is 0, otherwise bjThe value of block is 1, DFiIt is the vector that every value of t dimensions is 0/1;
Key production module, for generating encryption key SK (S, M1,M2,kf);Wherein, it is 0/1 that S, which is every value,
Random vector, M1And M2It is two n × n invertible matrix, n is the length of keyword dictionary sequence, kfIt is document encryption key;
Document vector encrypting module, for using the encryption key generated to each plain text document by safe KNN technologies
Vectorial ViEncryption obtains corresponding encrypted document vectorWherein, when j-th of element S in random vector S
When [j]=0, Vi′+Vi"=Vi, as S [j]=1, Vi'=Vi"=Vi;
File encryption module, for by each document in symmetric encipherment algorithm encrypted document collection DS, obtaining encrypted document
Collect ES={ e1,e2,…,em};
Transmission module is stored for document filter vectors to be transmitted to privately owned Cloud Server, by encrypted document vector
Publicly-owned Cloud Server is transmitted to encrypted document collection to be stored.
Further, the data retrieval end specifically includes:
Vector generation module is retrieved, multiple keyword { w for being provided according to user1,w2,…,wx, using TF-IDF
Algorithm and vector space model generate retrieval vector Q, and are normalized;Wherein, j-th of element Q [j] of Q is that jth position is crucial
Word wjInverse document frequency IDF values in data provide the document sets DS that end provides;
Trapdoor generation module is retrieved, for generating retrieval trapdoor using safe KNN algorithms based on retrieval vector QWherein, wherein when data provide end generate encryption key in random vector jth position S [j]=
When 0, Q ' [j]=Q " [j]=Q [j], as S [j]=1, Q ' [j]+Q " [j]=Q [j];
Filter vectors generation module is retrieved, for the piecemeal situation according to keyword in keyword dictionary sequence, generates inspection
Rope filter vectors QF, QF={ b1,b2,…,bt, QF is the vector that t ties up every value 0/1, if piecemeal bjCorresponding all keys
Word corresponding all values in retrieving vector Q are all 0, then QF [j]=0, otherwise QF [j]=1;
Transmission module needs the document number k retrieved to be transmitted to publicly-owned Cloud Server for will retrieve trapdoor and user, will
Retrieval filter vectors are transmitted to privately owned Cloud Server.
Further, the privately owned Cloud Server specifically includes:
With computing module, the document filter vectors DF of retrieval filter vectors QF and each document for that will receiveiPoint
It Jin Hang not be with operation, if QF&DFiThe obtained all positions of vector of operation be not all 0, then by DFiCorresponding document code DidiAdd
It is added to candidate documents concentration, obtains candidate documents CDS={ d1,d2,…};
Transmission module, for candidate documents CDS to be sent to publicly-owned Cloud Server.
The method of this system and embodiment 1 corresponds, and other parts repeat no more, reference implementation example 1.
Above disclosed is only presently preferred embodiments of the present invention, and the right model of the present invention cannot be limited with this
It encloses, therefore equivalent changes made in accordance with the claims of the present invention, is still within the scope of the present invention.
Claims (8)
1. a kind of secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud, it is characterised in that this method includes:
(1) data provide end and extract keyword set from the document sets of offer, and generate keyword dictionary by the piecemeal that clusters
Sequence;It is one corresponding plain text document of each document structure tree vector in document sets further according to keyword dictionary sequence, and according to pass
Plain text document vector piecemeal is formed document filter vectors by the piecemeal situation of keyword dictionary sequence;Later by plain text document vector
Encryption forms encrypted document vector, and each file encryption in document sets is formed encrypted document collection;Finally by document filter to
Amount is transmitted to privately owned Cloud Server, and encrypted document vector sum encrypted document collection is transmitted to publicly-owned Cloud Server;
(2) data retrieval end generates retrieval vector according to multiple keywords that user provides, and is calculated using safety after being normalized
Method generates retrieval trapdoor, needs the document number k retrieved to be transmitted to publicly-owned Cloud Server together with user;Again according to keyword dictionary
The piecemeal situation of keyword in sequence, multiple keywords that user is provided generate retrieval filter vectors, and are transmitted to private clound
Server;
(3) the retrieval filter vectors received and the document filter vectors of each document are carried out and are transported by privately owned Cloud Server respectively
It calculates, if all positions of vector that operation obtains are not all 0, corresponding document code, which is added to candidate documents, to be concentrated, and will be waited
Document sets are selected to be transmitted to publicly-owned Cloud Server;
(4) publicly-owned Cloud Server calculates separately time according to the candidate documents, retrieval trapdoor and the search file number k that receive
Selection shelves concentrate the safe inner product between the corresponding encrypted document vector sum retrieval trapdoor of each document, are chosen according to safe inner product
Candidate documents concentrate the maximally related k ciphertext document of keyword provided with user, which is back to data inspection
Bitter end;
(5) k ciphertext document of reception is decrypted in data retrieval end, obtains maximally related k plain text document.
2. the secret protection multi-key word Top-k cipher text retrieval methods according to claim 1 towards mixed cloud, feature
It is:Step (1) specifically includes:
(1-1) data provide end and extract keyword from the document sets DS of offer, obtain keyword set { w1,w2,…,wn};
Keyword root in keyword set is carried out the operation that clusters by (1-2) according to correlativity, obtains several cluster submanifolds { c1,
c2,…,ct};
(1-3) using each submanifold as a piecemeal, to obtain t piecemeal, respectively b1,b2,…,bt, given birth to further according to piecemeal
At keyword dictionary sequence W={ w (b1,1),w(b1,2),…,w(b2,1),w(b2,2),…,w(bt,1),w(bt, 2) ... },
Wherein w (bj, x) and it indicates to belong to piecemeal bjIn x-th of keyword, each keyword in the block is unordered;Piecemeal bj={ w (bj,x)
|0<x≤|bj|};
(1-4) uses TF-IDF algorithms and vector space model, is document according to the position of keyword in keyword dictionary sequence
Collect DS={ D1,D2,…,DmIn each document DiGenerate a corresponding plain text document vector Vi, and be normalized;Its
In, ViDimension be n, every value is the corresponding keyword of this in document DiIn word frequency TF values;
(1-5) according to the piecemeal situation of keyword dictionary sequence, by plain text document vector ViIt is divided into t piecemeal, piecemeal boundary and pass
The piecemeal boundary of keyword dictionary sequence is identical, obtains each document DiDocument filter vectors DFi={ b1,b2,…,bt};Wherein,
If ViMiddle piecemeal bjPosition value where corresponding all keywords is all 0, then bjThe value of block is 0, otherwise bjThe value of block
It is 1, DFiIt is the vector that every value of t dimensions is 0/1;
(1-6) generates encryption key SK (S, M1,M2,kf);Wherein, S is the random vector that every value is 0/1, M1And M2
It is two n × n invertible matrix, n is the length of keyword dictionary sequence, kfIt is document encryption key;
(1-7) is by safe KNN technologies using the encryption key generated to each plain text document vector ViEncryption, obtains corresponding
Encrypted document vectorWherein, as j-th of element S [j]=0 in random vector S, V 'i+V″i=Vi,
As S [j]=1, V 'i=V "i=Vi;
(1-8) obtains encrypted document collection ES={ e by each document in symmetric encipherment algorithm encrypted document collection DS1,e2,…,
em};
Document filter vectors are transmitted to privately owned Cloud Server and stored by (1-9), by encrypted document vector sum encrypted document collection
Publicly-owned Cloud Server is transmitted to be stored.
3. the secret protection multi-key word Top-k cipher text retrieval methods according to claim 1 towards mixed cloud, feature
It is:Step (2) specifically includes:
Multiple keyword { w that (2-1) data retrieval end is provided according to user1,w2,…,wx, using TF-IDF algorithms and space
Vector model generates retrieval vector Q, and is normalized;Wherein, j-th of element Q [j] of Q is jth position keyword wjIn data
Inverse document frequency IDF values in the document sets DS that end provides are provided;
(2-2) is based on retrieval vector Q, using safe KNN algorithms, generates retrieval trapdoorWherein, wherein
When data provide jth position S [j]=0 of the random vector in the encryption key that end generates, Q ' [j]=Q " [j]=Q [j] works as S
When [j]=1, Q ' [j]+Q " [j]=Q [j];
(2-3) generates retrieval filter vectors QF, QF={ b according to the piecemeal situation of keyword in keyword dictionary sequence1,
b2,…,bt, QF is the vector that t ties up every value 0/1, if piecemeal bjCorresponding all keywords are corresponding in retrieving vector Q
All values are all 0, then QF [j]=0, otherwise QF [j]=1;
(2-4) will retrieve trapdoor and user needs the document number k retrieved to be transmitted to publicly-owned Cloud Server, will retrieve filter vectors
It is transmitted to privately owned Cloud Server.
4. the secret protection multi-key word Top-k cipher text retrieval methods according to claim 1 towards mixed cloud, feature
It is:Step (3) specifically includes:
(3-1) privately owned Cloud Server is by the document filter vectors DF of the retrieval filter vectors QF received and each documentiRespectively into
Row and operation, if QF&DFiThe obtained all positions of vector of operation be not all 0, then by DFiCorresponding document code DidiIt is added to
Candidate documents are concentrated, and candidate documents CDS={ d are obtained1,d2,…};
Candidate documents CDS is sent to publicly-owned Cloud Server by (3-2).
5. a kind of secret protection multi-key word Top-k searching ciphertext systems towards mixed cloud, it is characterised in that the system includes
Data provide end, data retrieval end, privately owned Cloud Server and publicly-owned Cloud Server, wherein:
Data provide end and generate keyword dictionary for extracting keyword set from the document sets of offer, and by the piecemeal that clusters
Sequence;And it is vectorial for one corresponding plain text document of each document structure tree in document sets according to keyword dictionary sequence, and foundation
Plain text document vector piecemeal is formed document filter vectors by the piecemeal situation of keyword dictionary sequence;And by plain text document to
Amount encryption forms encrypted document vector, and each file encryption in document sets is formed encrypted document collection;And document is filtered
Vector is transmitted to privately owned Cloud Server, and encrypted document vector sum encrypted document collection is transmitted to publicly-owned Cloud Server;
Data retrieval end is used to generate retrieval vector according to multiple keywords that user provides, and is calculated using safety after being normalized
Method generates retrieval trapdoor, needs the document number k retrieved to be transmitted to publicly-owned Cloud Server together with user;And according to keyword word
The piecemeal situation of keyword in canonical ordering row, multiple keywords that user is provided generate retrieval filter vectors, and are transmitted to privately owned
Cloud Server;
The document filter vectors of retrieval filter vectors and each document that privately owned Cloud Server is used to receive carry out respectively and
Operation, if all positions of vector that operation obtains are not all 0, corresponding document code, which is added to candidate documents, to be concentrated, and will
Candidate documents are transmitted to publicly-owned Cloud Server;
Publicly-owned Cloud Server is used to, according to the candidate documents, retrieval trapdoor and the search file number k that receive, calculate separately time
Selection shelves concentrate the safe inner product between the corresponding encrypted document vector sum retrieval trapdoor of each document, are chosen according to safe inner product
Candidate documents concentrate the maximally related k ciphertext document of keyword provided with user, which is back to data inspection
Bitter end;
Data retrieval end is additionally operable to that k ciphertext document of reception is decrypted, and obtains maximally related k plain text document.
6. the secret protection multi-key word Top-k searching ciphertext systems according to claim 5 towards mixed cloud, feature
It is:The data provide end and specifically include:
Keyword extracting module obtains keyword set { w for extracting keyword from the document sets DS of offer1,w2,…,
wn};
Cluster module, and for the keyword root in keyword set to be carried out the operation that clusters according to correlativity, it is poly- to obtain several
Class submanifold { c1,c2,…,ct};
Keyword dictionary generation module is used for using each submanifold as a piecemeal, to obtain t piecemeal, respectively b1,
b2,…,bt, keyword dictionary sequence W={ w (b are generated further according to piecemeal1,1),w(b1,2),…,w(b2,1),w(b2,2),…,
w(bt,1),w(bt,2),…};Wherein w (bj, x) and it indicates to belong to piecemeal bjIn x-th of keyword, each keyword in the block
It is unordered;Piecemeal bj={ w (bj,x)|0<x≤|bj|};
Plain text document vector generation module, for using TF-IDF algorithms and vector space model, according to keyword dictionary sequence
The position of middle keyword is document sets DS={ D1,D2,…,DmIn each document DiGenerate a corresponding plain text document vector
Vi, and be normalized;Wherein, ViDimension be n, every value is the corresponding keyword of this in document DiIn word
Frequency TF values;
Document filter vectors generation module, for the piecemeal situation according to keyword dictionary sequence, by plain text document vector ViIt is divided into
The piecemeal boundary of t piecemeal, piecemeal boundary and keyword dictionary sequence is identical, obtains each document DiDocument filter vectors DFi
={ b1,b2,…,bt};Wherein, if ViMiddle piecemeal bjPosition value where corresponding all keywords is all 0, then bjBlock takes
Value is 0, otherwise bjThe value of block is 1, DFiIt is the vector that every value of t dimensions is 0/1;
Key production module, for generating encryption key SK (S, M1,M2,kf);Wherein, S be every value be 0/1 with
Machine vector, M1And M2It is two n × n invertible matrix, n is the length of keyword dictionary sequence, kfIt is document encryption key;
Document vector encrypting module, for using the encryption key generated to each plain text document vector V by safe KNN technologiesi
Encryption obtains corresponding encrypted document vectorWherein, when j-th of element S [j] in random vector S=
When 0, V 'i+V″i=Vi, as S [j]=1, V 'i=V "i=Vi;
File encryption module, for by each document in symmetric encipherment algorithm encrypted document collection DS, obtaining encrypted document collection ES
={ e1,e2,…,em};
Transmission module stores for document filter vectors to be transmitted to privately owned Cloud Server, encrypted document vector sum is added
Confidential document collection is transmitted to publicly-owned Cloud Server and is stored.
7. the secret protection multi-key word Top-k searching ciphertext systems according to claim 5 towards mixed cloud, feature
It is:The data retrieval end specifically includes:
Vector generation module is retrieved, multiple keyword { w for being provided according to user1,w2,…,wx, using TF-IDF algorithms
Retrieval vector Q is generated with vector space model, and is normalized;Wherein, j-th of element Q [j] of Q is jth position keyword wj
Inverse document frequency IDF values in data provide the document sets DS that end provides;
Trapdoor generation module is retrieved, for generating retrieval trapdoor using safe KNN algorithms based on retrieval vector QWherein, wherein when data provide end generate encryption key in random vector jth position S [j]=
When 0, Q ' [j]=Q " [j]=Q [j], as S [j]=1, Q ' [j]+Q " [j]=Q [j];
Filter vectors generation module is retrieved, for the piecemeal situation according to keyword in keyword dictionary sequence, generation was retrieved
Filter vector QF, QF={ b1,b2,…,bt, QF is the vector that t ties up every value 0/1, if piecemeal bjCorresponding all keywords exist
It is all 0 to retrieve corresponding all values in vector Q, then QF [j]=0, otherwise QF [j]=1;
Transmission module needing the document number k retrieved to be transmitted to publicly-owned Cloud Server, will retrieve for will retrieve trapdoor and user
Filter vectors are transmitted to privately owned Cloud Server.
8. the secret protection multi-key word Top-k searching ciphertext systems according to claim 5 towards mixed cloud, feature
It is:The privately owned Cloud Server specifically includes:
With computing module, the document filter vectors DF of retrieval filter vectors QF and each document for that will receiveiIt carries out respectively
With operation, if QF&DFiThe obtained all positions of vector of operation be not all 0, then by DFiCorresponding document code DidiIt is added to time
Selection shelves are concentrated, and candidate documents CDS={ d are obtained1,d2,…};
Transmission module, for candidate documents CDS to be sent to publicly-owned Cloud Server.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810122376.8A CN108363689B (en) | 2018-02-07 | 2018-02-07 | Privacy protection multi-keyword Top-k ciphertext retrieval method and system facing hybrid cloud |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810122376.8A CN108363689B (en) | 2018-02-07 | 2018-02-07 | Privacy protection multi-keyword Top-k ciphertext retrieval method and system facing hybrid cloud |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108363689A true CN108363689A (en) | 2018-08-03 |
CN108363689B CN108363689B (en) | 2021-03-19 |
Family
ID=63005057
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810122376.8A Active CN108363689B (en) | 2018-02-07 | 2018-02-07 | Privacy protection multi-keyword Top-k ciphertext retrieval method and system facing hybrid cloud |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108363689B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109194666A (en) * | 2018-09-18 | 2019-01-11 | 东北大学 | A kind of safe kNN querying method based on LBS |
CN109271485A (en) * | 2018-09-19 | 2019-01-25 | 南京邮电大学 | It is a kind of to support semantic cloud environment encrypted document ordering searching method |
CN109739945A (en) * | 2018-12-13 | 2019-05-10 | 南京邮电大学 | A kind of multi-key word ciphertext ordering searching method based on hybrid index |
CN110727951A (en) * | 2019-10-14 | 2020-01-24 | 桂林电子科技大学 | Lightweight outsourcing file multi-keyword retrieval method and system with privacy protection function |
CN112597268A (en) * | 2020-12-22 | 2021-04-02 | 南京邮电大学 | Retrieval filtering threshold value selection method for cloud environment ciphertext retrieval efficiency optimization |
WO2021103708A1 (en) * | 2019-11-26 | 2021-06-03 | 支付宝(杭州)信息技术有限公司 | Data query method, apparatus, device and system based on privacy information protection |
CN114189391A (en) * | 2022-02-14 | 2022-03-15 | 浙江易天云网信息科技有限公司 | Privacy data control and management method suitable for hybrid cloud |
CN116521743A (en) * | 2023-06-27 | 2023-08-01 | 北京中科江南信息技术股份有限公司 | Ciphertext retrieval method and device, storage medium and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120054485A1 (en) * | 2010-08-25 | 2012-03-01 | Sony Corporation | Terminal device, server, data processing system, data processing method, and program |
CN104765848A (en) * | 2015-04-17 | 2015-07-08 | 中国人民解放军空军航空大学 | Symmetrical searchable encryption method for supporting result high-efficiency sequencing in hybrid cloud storage |
CN105681280A (en) * | 2015-12-29 | 2016-06-15 | 西安电子科技大学 | Searchable encryption method based on Chinese in cloud environment |
CN106815350A (en) * | 2017-01-19 | 2017-06-09 | 安徽大学 | Dynamic ciphertext multi-key word searches for method generally in a kind of cloud environment |
CN107634829A (en) * | 2017-09-12 | 2018-01-26 | 南京理工大学 | Encrypted electronic medical records system and encryption method can search for based on attribute |
-
2018
- 2018-02-07 CN CN201810122376.8A patent/CN108363689B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120054485A1 (en) * | 2010-08-25 | 2012-03-01 | Sony Corporation | Terminal device, server, data processing system, data processing method, and program |
CN104765848A (en) * | 2015-04-17 | 2015-07-08 | 中国人民解放军空军航空大学 | Symmetrical searchable encryption method for supporting result high-efficiency sequencing in hybrid cloud storage |
CN105681280A (en) * | 2015-12-29 | 2016-06-15 | 西安电子科技大学 | Searchable encryption method based on Chinese in cloud environment |
CN106815350A (en) * | 2017-01-19 | 2017-06-09 | 安徽大学 | Dynamic ciphertext multi-key word searches for method generally in a kind of cloud environment |
CN107634829A (en) * | 2017-09-12 | 2018-01-26 | 南京理工大学 | Encrypted electronic medical records system and encryption method can search for based on attribute |
Non-Patent Citations (2)
Title |
---|
戴华 等: "面向云环境的多关键词密文排序检索研究综述", 《计算机科学》 * |
李晖 等: "公共云存储服务数据安全及隐私保护技术综述", 《计算机研究与发展》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109194666A (en) * | 2018-09-18 | 2019-01-11 | 东北大学 | A kind of safe kNN querying method based on LBS |
CN109194666B (en) * | 2018-09-18 | 2021-06-01 | 东北大学 | LBS-based security kNN query method |
CN109271485A (en) * | 2018-09-19 | 2019-01-25 | 南京邮电大学 | It is a kind of to support semantic cloud environment encrypted document ordering searching method |
CN109739945A (en) * | 2018-12-13 | 2019-05-10 | 南京邮电大学 | A kind of multi-key word ciphertext ordering searching method based on hybrid index |
CN109739945B (en) * | 2018-12-13 | 2022-11-08 | 南京邮电大学 | Multi-keyword ciphertext sorting and searching method based on mixed index |
CN110727951A (en) * | 2019-10-14 | 2020-01-24 | 桂林电子科技大学 | Lightweight outsourcing file multi-keyword retrieval method and system with privacy protection function |
WO2021103708A1 (en) * | 2019-11-26 | 2021-06-03 | 支付宝(杭州)信息技术有限公司 | Data query method, apparatus, device and system based on privacy information protection |
CN112597268A (en) * | 2020-12-22 | 2021-04-02 | 南京邮电大学 | Retrieval filtering threshold value selection method for cloud environment ciphertext retrieval efficiency optimization |
CN112597268B (en) * | 2020-12-22 | 2022-09-20 | 南京邮电大学 | Retrieval filtering threshold value selection method for cloud environment ciphertext retrieval efficiency optimization |
CN114189391A (en) * | 2022-02-14 | 2022-03-15 | 浙江易天云网信息科技有限公司 | Privacy data control and management method suitable for hybrid cloud |
CN116521743A (en) * | 2023-06-27 | 2023-08-01 | 北京中科江南信息技术股份有限公司 | Ciphertext retrieval method and device, storage medium and electronic equipment |
Also Published As
Publication number | Publication date |
---|---|
CN108363689B (en) | 2021-03-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108363689A (en) | Secret protection multi-key word Top-k cipher text retrieval methods towards mixed cloud and system | |
Zhang et al. | PIC: Enable large-scale privacy preserving content-based image search on cloud | |
US10013574B2 (en) | Method and apparatus for secure storage and retrieval of encrypted files in public cloud-computing platforms | |
CN104408177B (en) | Cipher text retrieval method based on cloud document system | |
US9313232B2 (en) | System and method for data mining and security policy management | |
Chuah et al. | Privacy-aware bedtree based solution for fuzzy multi-keyword search over encrypted data | |
Sadeghian et al. | SQL injection is still alive: a study on SQL injection signature evasion techniques | |
CN106407447A (en) | Simhash-based fuzzy sequencing searching method for encrypted cloud data | |
US11595435B2 (en) | Methods and systems for detecting phishing emails using feature extraction and machine learning | |
CN106326360A (en) | Fuzzy multi-keyword retrieval method of encrypted data in cloud environment | |
Su et al. | Privacy-preserving top-k spatial keyword queries in untrusted cloud environments | |
CN108959567A (en) | It is suitable for the safe retrieving method of large-scale image under a kind of cloud environment | |
CN106972927A (en) | A kind of encryption method and system for different safety class | |
CN115314295B (en) | Block chain-based searchable encryption technical method | |
US20130246338A1 (en) | System and method for indexing a capture system | |
Boucenna et al. | Secure inverted index based search over encrypted cloud data with user access rights management | |
CN109739945A (en) | A kind of multi-key word ciphertext ordering searching method based on hybrid index | |
CN109740378B (en) | Security pair index structure resisting keyword privacy disclosure and retrieval method thereof | |
Wang et al. | A modified homomorphic encryption method for multiple keywords retrieval | |
Abduljabbar et al. | Secure biometric image retrieval in IoT-cloud | |
Gupta et al. | A learning oriented DLP system based on classification model | |
CN108829714A (en) | A kind of ciphertext data multi-key word searches for method generally | |
CN112966086A (en) | Verifiable fuzzy search method based on position sensitive hash function | |
Hussain et al. | A novel method for preserving privacy in big-data mining | |
Yao et al. | Topic-based rank search with verifiable social data outsourcing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20180803 Assignee: NUPT INSTITUTE OF BIG DATA RESEARCH AT YANCHENG Assignor: NANJING University OF POSTS AND TELECOMMUNICATIONS Contract record no.: X2021980013920 Denomination of invention: Hybrid cloud oriented privacy protection multi keyword Top-k ciphertext retrieval method and system Granted publication date: 20210319 License type: Common License Record date: 20211202 |
|
EE01 | Entry into force of recordation of patent licensing contract |