IRJET- Review on Privacy Preserving on Multi Keyword Search over Encrypted Data in Cloud

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 06 | June 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 2030
Review on Privacy Preserving on Multi Keyword Search over Encrypted
Data in Cloud
Ms. Apurva S. Solanke
Student, Department of Computer Science, MSS CET, Jalna, Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - As cloud computingbecomeswidespread, more and
more sensitive information are being centralized into the
cloud for great convenience and reduced cost in data
management. Cloud information proprietors liketooutsource
records in on encoded form for the capacity of secrecy
protecting. User convey data using internet which can be
affected by malicious attacks. So to preserve the data privacy
of document data should be outsourced in encrypted format
which perform traditional data utilization based on plaintext
keyword search. As a result allowing an encrypted cloud data
search service is of supreme significance. In view of the large
number of data users and document in the cloud, it is essential
to permit several keywords in the search demand and return
document in the order of their appropriate to these keywords.
In this proposed system top k query is used to retrieve top
matching result from huge encrypted answer set relevant to
user interest. In retrieve the document of user interest multi-
keyword query has to be fired which will result in accessing
the top k ranked document. Lucene indexing technique has
been developed to build on index of keywords from document.
Index will store hashed values of keyword. Blowfish algorithm
is used for encryption and decryption of data.
Key Words: Lucene indexing, Multikeyword, Blowfish,top
k, Privacy.
1. INTRODUCTION
Cloud computing can be defined as a model for enabling
ubiquitous, convenient and on demand network access to
shared pool of configurable computing resourcesthatcanbe
rapidly provisioned and releasedwithminimal management
effort from the user side and minimal service provider
interaction. Cloud computing is considered the evolution of
variety of technologies that come together to change an
organizations approach for building their IT infrastructure.
Cloud is not simply the latest term for internet though the
internet is necessary foundation for the cloud, the cloud
something more than internet. Cloud is where you go to use
technology when you need it, you need not install anything
on your desktop and you do not need to pay for the
technology when you are not usingit.Cloudcomputingrelies
on sharing resourcesto ensureconsistencyandeconomiesof
scale. Placing our confidential data in hand of cloud storage
is critical because of the term data confidentiality has
outstanding importance. Our original plain data must be
accessible by only trusted parties. Data must be encrypted
and must not be accessible byuntrustedthirdpartiessuchas
cloud provider, intermediaries and internet. Therefore to
ensure data confidentiality and unauthorized access to the
data placed on cloud owners have to adopt encryption
process on data while placing confidential data on cloud.
User will assemble encrypted data on cloud. So, traditional
data searching process will not be effective and efficient for
encrypted cloud storage. To retrieve efficient result on
encrypted data multi-keyword query should be structured
and fired. This multi-keyword query helps to find out top
related data belonging to user concern. Existing system
works on single query searching procedures or on single
user with Boolean searchprocedures.TheExistingsearching
method for data retrieval are only restricted for single
keyword queries. These searching method fetch all the
related data correspondingtothespecifiedkeywordwithout
ranking the data of user interest. Indexing is common
operation in web search engines. Large amountofdocument
demands the cloud server instead of returning
undifferentiated result only relevant document can retrieve
using ranking.Rankingcanalsoeliminateunwantednetwork
traffic for protection purpose the information related to
keyword should not be leak while use the ranking scheme.
As we step into the big data era terabyte of data are
produced worldwide per data enterprises and users who
own large amount of data usually choose to outsource their
precious data to cloud facility in order to reduce data
management cost and storage facility spending. As a result
data volume in cloud storage facilities is experiencing
dramatic increase.
Search over encrypted data is technique of greatinterest
in the cloud computing because many believe that sensitive
data has to be encrypted before outsourcing to the cloud
server in order to ensure user data privacy.
2 LITERATURE SURVEY
In some existing methods, for generation of queries user
must have to gather all necessary informationaboutall valid
keywords along with its positions. Cloud storage acts as
database repository as it stores all information and
documents of users. One of the cryptographic primitive is
searchable encryption which allows private keyword based
searching over encrypted database. Existing system has
some of the disadvantages like: only single user can search,
searching possible for Single or Boolean keyword without
ranking, index creation happens very rarely.

Deepali D. Rane et.al, proposed implementation of the
encryption and decryption, secure index construction is
successfully completed with desirable performance. After
index construction it will get compressed and will be stored
in .cfs file format. After firing single-keyword query,userwill
get all documents that contain the specified keyword. The
advantages are protects data privacy by encrypting
documents before outsourcing, rank based retrieval of the
documents, To easily access the encrypted data by multi
keyword rank search using keyword index.
A single keyword searchable encryption schemes
usually builds an encrypted searchable index such that, it’s
content is hidden to the server, unless it is givenappropriate
trapdoors generated via secret key(s). Early work solves
secure ranked keyword search which utilizes keyword
frequency to rank results instead of returning
undifferentiated results. However, it only supports single
keyword search. Where anyone with public key can write to
the data stored on server, but only authorized users with
private key can search. Traditional single keyword
searchable encryption schemes are usually built in a way by
creating an encrypted searchable index. Such indexes
content will be hidden to the server. The information will be
revealed only when the server gives the correct trapdoors
that are generated via a secret key(s). The main drawback of
single keyword-based search is that it is not comfortable
enough to express complex information needs.
2.1 Comparison of Algorithm
Symmetric encryption is also called as secret key
encryption which uses a single key for both encryption and
decryption processes. Same key is utilized toconveydata for
encryption process and in decryption process. Conveying
key using Internet can be affected by malicious attacks. So if
both parties are aware about the keys then symmetric
encryption technique is more appropriate for
communication to transfer secrete data over internet.
Asymmetric encryption is also called as public key private
key encryption which utilizes two distinct keys for
encryption and decryption processes, public key and the
private key respectively. The key can be transformedintoan
array of bytes for the intention of storing and transferring
secretes data. This process is called as wrapping. When key
is required for decryption purpose then array of bytes is
again converted in key. This process iscalledasunwrapping.
There are various algorithms for data encryption using
symmetric and asymmetric encryptionkeys.Fewofthem are
stated and compared below:
2.2 DES:
DES (Data Encryption Standard) is the very earliest
encryption standard developed by NIST (National Institute
of Standards and Technology) and commonly used in the
commercial, military, and other domains. It was emerged in
1947 by IBM and accepted as a national standard in 1997.
DES standard is public & the design criteria used are
classified. DES is a 64-bit block cipher which utilizes 56-bit
key. This algorithm contains permutation of sixteen rounds
block cipher. DES algorithm carried out permutation and
substitution procedures on every block of plaintext data
which is afterwards XORed with the given input. The same
process is carried out 16 times with different sub keys. It is
vulnerable to security attacks and secrete data can be easily
retrieved.
Fig. DES Algorithm
2.3 AES:
AES (Advanced Encryption Standard) was discovered by
scientists Joan and Vincent Rijmen in the year 2000. It
replaced DES as the official standard of US National Institute
for Standards and Technology (NIST).AESworksonvariable
key sizes and variable block sizes of 128, 192 or 256 bits.
AES uses Rijndael block cipher and Rijndael key. Due to the
number of permutations and combinations in AES
contributing to its higher complexity, AES has a higher
security as compared to DES002E.

Fig. AES algorithm
3. PROPOSED SYSTEM
3.1 System Development:
Major objective of the proposed systemistoextractthe
top relevant data similar to users queryaswell astoenhance
experience of users while searching. Single keyword
searching technique generates unwanted traffic of data
which is not that much related to users query. Hence we
need to develop robust system which will provide efficient
results with multi-keyword searching. As cloud storage is
not trusted to keep our plaintext data, we need to save data
in encrypted form to ensure privacy. User will register on
cloud storage by registration process. Key generationcenter
(KGC) will provide login credentials to user. After successful
login user can store their data on cloud. User can upload his
confidential data and documents on cloud. Next time
whenever user will ask for uploaded document, it will
retrieve in encrypted form. So that only user of data is able
to decrypt the encrypted data with key. Once document get
uploaded on cloud server, proposed system extracts
attributes of files such as file name, file date, author name,
title of paper, etc,. Results are ranked and retrieved using
indexing techniques. Lucene indexing is efficient among
other indexing techniques and hence used in the proposed
system. Top relevant results are also retrieved matching
with user queries using top-k query.
Proposed system will solve the problem of searching of
information from encrypted data stored on cloud with the
help of multi-keyword search query which also preserves
system-wide privacy in cloud computing. Proposed system
used Keyword relevance technique as an intermediate
similarity measure which extracts keywords from search
query to match with keywords stored on cloud. Ranking
system is developed which provides the effective retrieval of
results using lucene indexing. Top matching results are
retrieved using top-k querysothatunwantednetwork traffic
gets avoided. Results relevant to user’sinterestare retrieved
efficiently using multiple keyword search procedure. Single
Keyword search query will yield too much unnecessary
results so that there is necessity of supporting multi-
keyword query.
3.2 Lucene Indexing:
When person wants to upload his official or personal
data on the cloud server, keywords are get abstracted from
that documents and index will be generated for all
abstracted keywords [1], [4]. Hashed valuesofkeywordsare
stored as index. When documentwill getaddedandremoved
from cloud storage, index will get restructured accordingly.
In proposed method, for generating index of keywords
extracted from documents, Lucene indexing technique is
used. Lucene possesses full-text search engine architecture
which delivers complete query and indexing engine. Lucene

indexing technique delivers distinct advantages such as
indexing is like file format and application platform
independent. It also delivers searching of data over
documents [9]. A Lucene technique provides continuous
updations into indexes of fetched keywords when they are
get added and removed from storage space.
Lucene is very wealthy and potentfull-textsearchlibrary
written in Java. Lucene indexing is utilized to deliver fulltext
indexing over both database objects and documents in
various formats. Lucene indexing API is independent from
other file formats. Textual data from various resources such
as pdf, HTML, Microsoft word, openDocument is get
extracted first. In Lucene indexing, to construct an index,
whole document gets scan first so as to produce list of
postings. Occurrence of word in a document is get described
by postings [10]. Posting generally consist of word,
document identifier and frequency of the word within that
document. Indexing process is one of the main functionality
delivered by Lucene. Figure shown below describes the
overall indexing procedure carried out by lucene and use of
each class. IndexWriter is the most significant and main
module of the indexing process.
Fig. Lucene Indxing Architecture
Supporting full-text search using Lucene indexing
executes two steps:
(1) Generating a lucene index on the documents and on
database objects and
(2) Parsing the user query and looking up the prebuiltindex
to answer the query.
Above figure describes complete working procedure of
Lucene. As soon as the documentwill getsuploadedontothe
cloud server index segment will getgenerated.Inaccordance
with merge factor, index segments will get merged. Index
contains hashed values of keywords.
3.3 Top-K Query:
Information retrieval systems utilize diverse
approaches to rank query answers. Users are more
apprehensive for the most significant i.e., top-k query
response from large answer space. Some emerging
applications claim for support of top-k queries. Incaseof the
Web, top-k query can be used to provide effectiveness and
efficiency of Meta search engines. Distinct applications are
also presents in fields of information retrieval and data
mining. Goal of this top-k query algorithm is to fetch top
matching results from large record set. Top-k query helpsto
search more precise answers from specified record set that
equivalent with filtering keyword, and assemble relevant
answers in accordance with their scores. Steiner tree is
constructed with each result set which is basically an XML
tree [3]. The constructed Steiner tree will holds the list of all
records which are already assembled in accordance with
their scores. Top-K query algorithm utilizes scores
documents against keywords. This algorithm is used torank
“Tuple Units”. A tuple unit consists of a set of similar tuples
which again consists of query keywords. These tuple units
are used to construct tuple set. When two tuples are directly
associated with each other then they are merges into one
single tuple. Direct scoring and indirect scoring techniques
are used in top-k approach to rank every tuple. Top-K query
scores every tuple according to two scoring methods,
namely, Direct Scoring and Indirect Scoring. Direct Scoring
depends on TF-IDF algorithm. TF-IDF stands for “Term
Frequency, Inverse Document Frequency”.
TF-IDE presents a way to score the words depending
on how repeatedly they come into sight across various
documents. Scores are termed according to the following
criteria: 1) If a word appears frequently in a tuples, it is
termed as important and is scored high. 2) If a word appears
in many tuples, it is termed as a unique identifier and is
scored low. Therefore, common words like "the" and "for",
which appear in many tuples, will be neglected. Words that
appear frequently in a single tuple will be scale up. The
second type of scoring method is Indirect Scoring. Indirect
Scoring scores a tuple unit based on the keyword that is
indirectly present in the tuple unit. The keyword is
scrutinized against each tuple unit for indirect similarity.
3.5 Blowfish Algorithm:
Encryption approaches are categorized in two parts as
symmetric key encryptions and public key encryptions.
Blowfish falls under Symmetric algorithms. Blowfish
algorithm is developed by Bruce Schneider and also known
as block cipher. As name in indicated, it dividesmessageinto

fixed length blocks while encryptionanddecryption process.
Blowfish algorithm can be used asquick alternativeforsome
existing encryption processes and as it delivers robustness,
none of the attack becomes successful against it. Blowfish
algorithm consists of key expansion a technique which
makes it more complex to break. It also takes less time for
encryption process as compared to other encryption
algorithms as AES, DES.
Blowfish contains 16 rounds. Each round holds XOR
operation and a function. Each round consists of key
expansion method and data encryption process [6]. Key
expansion generally used for generating initial contents of
one array and data encryption uses a 16 round Feistal net-
work. Foundation of Blowfish Algorithm is Feistel Network
which iterates simple encryption function 16 times [12].
Proposed system employ Blowfish encryption algorithm for
encryption of data. Homomorphic encryption is
implemented along with Blowfish algorithm. Homomorphic
encryption permits computations on cipher text directly. So
that, generated encrypted outcome when again gets
decrypted will exactly be similar because all operations are
carried out on plaintext. Homomorphic encryption provides
facility to perform complex mathematical computations on
encrypted data without negotiating the encryption data
quality. Blowfish algorithm consists of 64-bit block cipher
and variable length key. This algorithm requires less
memory space and yields high speed as compared to other
algorithms. This algorithm needs 32 bit microprocessor at a
rate of one byte for every 26 clock cycles. Itemploysvariable
length key block cipher up to 448 bits. Blowfish contains
total 16 rounds. Plain text and key are the inputs provided
for Blowfish algorithms.
4. CONCLUSION
Proposed systemfocusedonhowtheusercaneffectively
accumulate their private documents on cloud while
preserving privacy of their documents and whenever and
wherever necessary they can retrieve them by sending a
query consisting of multiple keywords. Privacy will be
achieved by encrypting the queries. In response to the users
query, system will match the keywords from query to the
documents using “keyword-matchingprinciple”.Topranked
documents will get fetched which will consist of keywords
specified by user in query. Checking therank oftheretrieved
document can be done by calculating how many docs
contains the specified keywords how many times.
5. REFERENCES:
[1] Miss Deepali D. Rane, Dr. V. R. Ghorpade , “Multi-User
Multi keywordPrivacyPreservingRankedBasedSearchover
Encryted Cloud Data”,International ConferenceonPervasive
Computing (ICPC).
[2] Prachi Jain, Prof.Shubhangi Kharche Effectuation of
Blowfish Algorithm using Java Cryptography”, International
Journal of Scientific & Engineering Research Volume4,Issue
5, May-2013.
[3] Ning Cao, Cong Wang, Ming Li, kuiren, Wenjing Lou ,
“Privacy Preserving Multi-Keyword Ranked Search over
Encrypted Cloud Data” IEEE Transactions on parallel and
Distributed Systems, Vol. 25,January 2014
[4] Ayad Ibrahim, Hai Jin, Ali A.Yassin, deqingzou, “Secure
Rank-Ordered Search of Multi-Keyword Trapdoor over
Encrypted Cloud Data” IEEEAsia-PacificServicesComputing
Conference 2012
[5] Yanjiang Yang, “Towards Multi-User Private Keyword
Search for Cloud Computing” IEEE 4th International
Conference on Cloud Computing. 2011.
[6] Qin Liu, Guojun Wag, Jie Wu, “An Efficient Privacy
Preserving Keyword Search Scheme in Cloud Computing”
IEEE International ConferenceonComputational Science and
Engineering, 2009.
[7] Twofish : A 128 Bit Block Cipher by Bruce Schneier, John
Kelseyy, Doug Whitingz David, Wagnerx Chris Hall . [8]
“Analysis of AES and Twofish Encryption Schemes” IEEE
Transaction 2011.
[8] Yong Zhang, Jian-lin Li, “Research and Improvement of
Search Engine Based on Lucene” IEEE Transaction 2009.
[10] QIAN Liping, WANG Lidong, “An Evaluation of Lucene
for Keywords Search in Large-scaleShortTextStorage”IEEE
Transaction 2010.

[9] Govind S.Pole, Madhuri Potey, “A Highly Efficient
Distributed Indexing system based on large cluster of
commodity machines” IEEE Transaction 2012.
[10] Ms Neha Khatri – Valmik, Prof. V. K Kshirsagar,
“Blowfish Algorithm”,IOSR Journal ofComputerEngineering
(IOSR-JCE) e-ISSN: 2278- 0661, p- ISSN: 2278-8727Volume
16, Issue 2, Ver. X (Mar-Apr. 2014), PP 80-83.

IRJET- Review on Privacy Preserving on Multi Keyword Search over Encrypted Data in Cloud

More Related Content

What's hot (20)

Similar to IRJET- Review on Privacy Preserving on Multi Keyword Search over Encrypted Data in Cloud (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Review on Privacy Preserving on Multi Keyword Search over Encrypted Data in Cloud