Enhanced Algorithm for Efficient Retrieval of Data from a Secure Cloud

Transcription

1 OPEN JOURNAL OF MOBILE COMPUTING AND CLOUD COMPUTING Volume 1, Number 2, November 2014 OPEN JOURNAL OF MOBILE COMPUTING AND CLOUD COMPUTING Enhanced Algorithm for Efficient Retrieval of Data from a Secure Cloud NISHA T. M. 1*, LIJO V. P. 2 1 Information Technology Department, MES College of Engineering, Kuttippuram, Malappuram, India. 2 Computer Science & Engineering, MES College of Engineering, Kuttippuram, Malappuram, India. *Corresponding author: nishunni@gmail.com Abstract: In this technical era of computing, one of the most important words used is Cloud Computing. It is growing day by day due to its rich features of services. Cloud computing uses internet and central remote servers to maintain data and applications. The data in the cloud are stored in off premises as encrypted form. One of the most popular ways is selectively retrieve files through keyword-based search instead of retrieving all the encrypted files back. There are so many challenges in cloud computing. One of the main challenges is to increase the efficiency of data retrieval and searching. There are several keyword search mechanisms for data retrieval. Currently many researches are focusing on the fuzzy keyword search, which searches and retrieves the data in most secure and privacy preserving manner. The existing system uses single fuzzy keyword search mechanism, which struggling with inefficiency of search and that may be overloaded with resultant files of irrelevant data. In the proposed system a conjunctive keyword search is introducing so that more efficient and relevant data files can be retrieved. The conjunctive keyword search automatically generate a ranked result so that the searching relevancy and efficiency will be improved. This will reduce the searching time and improve the efficiency by reducing time for index generation. Keywords: Cloud Computing; Fuzzy Keyword; Trie-tree; Advanced Trie-tree;Conjunctive Keyword; Encryption 1. INTRODUCTION In the 1960s, J.C.R.Licklider introduced the term intergalactic computer network [1] at the Advanced Research Projects Agency, introduced the most important concept Internet. The term Cloud originates from the telecommunications world of 1990s. It is the collection of servers, where all the applications or data are stored. All the servers are not installed in same location. They are distributed geographically desperate locations. It provides a scalable service delivery platform. Resources are being shared at various levels and they offer different services. The main modes of services are Software as a service (SaaS), Platform as a service (PaaS), and Infrastructure as a service (IaaS). There are so many security issues. The main security issues arise in data storage, data retrieval, searching etc. There are so many issues in storing the data securely in the cloud because most of the 50

2 Enhanced Algorithm for Efficient Retrieval of Data from a Secure Cloud sensitive information is centralized in to clouds. Examples are personal health records, different types of medicines, government documents, private videos and photos, company finance data etc. But the most important aspect arises in data retrieval part. The data owner stores their data in the cloud and any authorized person can access those files. If an unauthorized person enters, on retrieval they can do modification, insertion and deletion in the original files and can store back in clouds. Here the original data can be interrupted, which may cause security problems. So these sensitive data are encrypted before outsourcing. One of the most popular ways or techniques is to selectively retrieve files through keyword-based search instead of retrieving all the encrypted files back. The data encryption also demands the protection of keyword privacy since keywords usually contain important information related to the data files. The existing searchable encryption techniques do not suit for cloud computing scenario because they support only exact keyword search. This significant drawback of existing schemes signifies the important need for new techniques that support searching edibility, tolerating both minor typos and format inconsistencies. A secure fuzzy search [2] capability is demanded for achieving enhanced system usability in cloud computing. The main problem is how efficiently searching the data and retrieves the results in most secure and privacy preserving manner. So it is clear that one of the main challenges in cloud computing is the data retrieval and searching. For retrieving the data in a most secure and privacy preserving manner the keyword searching technique is used. To search the data in more efficient manner, the fuzzy keyword search is introduced. When the files are retrieved in an efficient manner, most relevant data can be retrieved. That is the fuzzy keyword search will provide a privacy preserving keyword search mechanism. In the existing system it is mainly focusing on the fuzzy keyword search method. The data that is outsourced is encrypted, constructs fuzzy sets based on both wild card technique and gram based technique, and has introduced a symbol-based trie-traverse search scheme, where a multi-way tree has constructed for storing the fuzzy keyword set and finally retrieving the data. All fuzzy words in the trie tree can be found by a depth first search. 2. RELATED WORKS The main challenge in cloud computing is its security. Security in data storage, efficient data retrieval from the centralized cloud storage, relevant data retrieval by keyword searching, etc. The security in searching is important because, each keyword should contain the meaning of the underlying information. So the keyword search should retrieve the most relevant data in an efficient manner. As cloud computing is in its growing stage, so many researches are going on in this area. D. Boneh, proposed a method [3], which deals with the problem of searching data that is encrypted using a public key system. The mechanism is known as Public Key Encryption with keyword Search. S. Ji, proposed a new computing paradigm, called interactive, fuzzy search [4]. It has two unique features such as Interactive and Fuzzy. It gives the idea about the queries with a single keyword, and presents an incremental algorithm for computing keyword prefixes [4]. It also gives an idea about various techniques for computing the intersection of the inverted lists of query keywords. J. Li, proposed another fuzzy keyword search method which includes Wild card [2] based method and Gram based [2] method for constructing fuzzy keyword sets, a symbol-based trie-traverse search scheme where a multi-way tree was constructed for storing the fuzzy keyword set and finally retrieving the data. This greatly reduces the storage and representation overheads. It also exploits Edit distance [2] to quantify keywords similarity, to build storage-efficient fuzzy keyword sets to facilitate the searching process. In order to enrich the search functionalities, introducing a conjunctive [5, 6] fuzzy keyword search over encrypted data. Conjunctive keyword search returns all-or-nothing, which means it only returns those 51

3 OPEN JOURNAL OF MOBILE COMPUTING AND CLOUD COMPUTING Figure 1. Architecture of Fuzzy Keyword Search documents in which all the keywords specified by the search query appear. In the previous work [7] it is proposed the conjunction of fuzzy keyword search mechanism. Here it is introducing a new algorithm for this conjunction of fuzzy keyword search and it increases the efficiency and reduces the complexity. 3. SEARCH AND RETRIEVAL BY CONJUNCTION OF FUZZY KEYWORDS 3.1 Single Fuzzy Keyword Search The existing fuzzy keyword search using single fuzzy keyword is described here. A cloud data system consisting of data owner, data user and cloud server. The architecture of fuzzy keyword search [2] is shown in the Figure 1. The working principle of fuzzy keyword search [2, 8] is given here. Initially given n encrypted data, C = (F 1,F 2,...,F n ) which is stored in the cloud server. Along with that an encrypted predefined set of distinct keywords W = {w 1,w 2,...,w p }are also stored in the cloud server. Cloud server provides the authorization to the users who want to access the encrypted files. Only the authorized person can access the stored data. The authorized user types the request that they want to search. The cloud server maps the request to the data files, which is indexed by a file ID and is linked to a set of keywords. The execution of fuzzy keyword search returns these set of file IDs whose corresponding data files possibly contain the word w, denoted as FID w, if w = w i 2 W, then return FID wi. Fuzzy keyword search will return the results by keeping the following two rules. 1. If the user s searching input exactly matches the predefined keyword, then the server is expected to return the files containing that keyword. 2. If there is no exact match or some inconstancies in the searching input, the server will return the closest possible results based on pre-defined similarity semantics. 52

4 Enhanced Algorithm for Efficient Retrieval of Data from a Secure Cloud 3.2 Fuzzy Keyword Set Construction In single fuzzy keyword search method, different keyword set construction mechanisms is discussed. From that it is cleared that wildcard method is the best keyword set construction method in the security aspect. So in the current method we are using wildcard method for fuzzy keyword set construction. So it will generate all the variants of the given keyword. When an encrypted file is outsourcing in to the cloud server, fuzzy set for all the keywords in the file should be generated by wild card method and stored in the cloud server. An authorized user wants to search for a conjunction of request (AND, OR, NOT, BOTH), its fuzzy set also want to generate by the same wild card technique. 3.3 Efficiency of Data Retrieval In the proposed work, the data owner stores the conjunction of keywords and their fuzzy set in the advanced trie-tree data structure. Here the Conjunction of keywords can be stored in the same advanced trie-tree. This is because the fuzzy set s offsets for each keyword with different edit distance (d = 1; d = 2...) can store in the fuzzy pointers. (FUZZY 1; FUZZY 2... ) in the same data structure. In the same method the user can also generate the search request. In general for a keyword of average length m, the number of variants constructed are 2* m So for conjunction of keywords having n number of keywords n (2 * m ) variants are constructed. For gram based method for the key word SIMPLE with the pre-set edit distance 1, the total number of variants constructed is m + 1. So for conjunction of keywords having n number of keywords n (m + 1) variants are constructed. The complexity of searching will reduce from O (mˆ2*n) in to O (m* n). The space complexity is reduced from O (mˆ2* n* p) in to O (m* n* p). So the search complexity is O (m)*o (n). That is the order is reduced from O (mˆ2* n) in to O (m*n). One m is eliminated due to the fuzzy pointer in the advanced trie-tree (Fuzzy set list). The space complexity is O (m* n* p). The space complexity is O (mˆ2*n *p) in to O (m* n* p). So the search complexity is O (m) *O (n). That is the order is reduced from O (mˆ2* n) in to O (m* n). One m is eliminated due to the fuzzy pointer in the advanced trie-tree (Fuzzy set list). The space complexity is O (m* n* p). The space complexity is also reduced. 3.4 Implementation When a single keyword and its fuzzy set is wanted to insert in the advanced trie- tree, the main processes are: Insertion 1. Initially each character in the given keyword is added in to the node. 2. The exact index field is is set to -1 until reaches at the end of the word. 3. When the word s end is reached, the exact index field stores the offset of the file where the word is stored or stores the file ID (FID). 4. The fuzzy pointer will stores the offsets of the fuzzy keywords set. Each fuzzy pointer can store the fuzzy set with different edit distance. 5. Then the next word reads and the process goes on until all the keyword sets are inserted in the trie-tree. The node defined in an advanced trie tree typdef struct TrieNode 53

5 OPEN JOURNAL OF MOBILE COMPUTING AND CLOUD COMPUTING Figure 2. Example for Keyword Set Construction { char data; //for storing character struct TrieNode*child[26] ; int exactindex; pfuzzynode head; }TrieNode; *ptrienode; Searching The searching is carried out in a depth First search manner. The algorithm for conjunction of keyword search is given in the Algorithm 1. Example An example is shown in Figure 2. Suppose a small file calls FILE, that want to store in the cloud server by the data owner contains the following keywords, and file ID FID DATA, OS, BASE, MOBILE, DESIGN, MATHS, GRAPH, MINING, MODEL. These keywords are stored in the advanced trie-tree data structure. For each keyword contains an exact index and fuzzy pointers. The fuzzy pointer FUZZY1 contains the offsets of the fuzzy set generated by applying wild card method with edit distance 1. FUZZY2 contains the offsets of fuzzy set generated by wildcard method with edit distance 2. Exact field contains the FID of corresponding file. Now the user search for a conjunction of search request that is the input query. DATA AND BASE AND DESIGN NOT MINING The searching request is shown in the Figure 3. The first keyword DATA is searched in advanced trietree by using its hash value and then its fuzzy set is also searched. After that it checks for the conjunction keyword. If it is AND, checking the next keyword BASE and so on. If it is the conjunction keyword NOT neglects the next keyword MINING. 54

6 Enhanced Algorithm for Efficient Retrieval of Data from a Secure Cloud Figure 3. Representation of Input Request 4. RESULTS AND DISCUSSION It contains a cloud database, which consists of almost one thousand of data files, and has an Administrator 55

7 OPEN JOURNAL OF MOBILE COMPUTING AND CLOUD COMPUTING (owner) and user. The administrator can add, view and delete the data. When a data or file is added it should be encrypted and stored in the database that is the data entered will be stored as encrypted files. AES encryption algorithm is used for encryption. The administrator can also view the encrypted data and delete the data. This encryption provides the security of the stored data. When a user wants to access the data by keyword search mechanism, first he wants to get the authorization. That is the authorized person can only retrieve the data. The authorization is provided by a key which is randomly generated. That key is unique for each user. The user should remember this key throughout the searching process. The user can enter the key words which is the conjunction of single keywords. That is AND, OR, BOTH and he get a search result which is very efficient and also in a ranked order. The efficiency is increased by introducing an advanced trie-tree data structure instead of normal trie-tree data structure. That is for normal trie-tree method, for each single keyword request, the search cost only O(m) at the server side, where m is the length of hash value. That is for each character a hash value is calculated. By using this hash value it is directly find the next character. In order to pointing in to the child node each 56

8 Enhanced Algorithm for Efficient Retrieval of Data from a Secure Cloud hash value is added with key value, which is specifically defined for each levels in the advanced trie-tree. Space complexity is of O (m* s), s is the width of the tree. So by introducing conjunction of keyword in normal trie-tree data structure, for n conjunction of keywords having average keyword length l keeping the edit distance 1 is of O (l*n)* O(m). That is O (l * m * n). Space complexity is O (l * m * n * s). Here each fuzzy set is storing separately in the normal trie-tree. So conjunction of keyword will produce the combination of different small normal trie-tree joint together by the root nodes containing AND, OR, NOT etc: But our proposed work the storage complexity and searching complexity is reduced. Which is of the order of O (m* n * s) and O(m) * O(n) respectively. The relevancy is improved by introducing the ranking method. A ranking rules performance should be based on the total ranking it imposes on the collection with respect to the query. The two important measures of effectiveness for retrieving the data are Recall and Precision. The most common way to describe retrieval performance is to calculate how many of the relevant documents have been retrieved and how early in the ranking they were listed. The Precision Pr of a ranking method for some cutoff point r is the fraction of the top r ranked documents that are relevant to the query. Pr = (Number retrieved that are relevant) / (Total number retrieved) The recall Rr of a method at some value are r is the proportion of the total number of relevant documents that were retrieved in the top r. Rr = (Number relevant that are retrieved) / (Total number relevant) Suppose the searching query contains n keywords, from this m keywords are available in the file, where m * n, that is n m 0. Then the ranking ratio h = m/n, where 0 0 apple h apple 1. If the value of h = 1, that means the searching result contains all the keywords in the input query. This file will come first in the search result. The search result will display the files according to the decreasing order of the ranking ratio. Figure 4. Relevancy of Data (h = m/n) Suppose a searching query contains the sequence of 10 keywords. If all the searching keywords contain in the file, that is m = n, then the value of h = 1. The file with highest ranking ratio comes first in the search result. A graphical representation is given in the Figure 4. Where the straight line represents the value h =1. The curve represents the value of h < 1, that is 0 apple h < 1. C. Wang, in his paper Secure ranked key word search over encrypted cloud data, analytically proved that if the ranking ratio is greater than 0.6, the data will be relevant. As the value increases above 0.6 the relevancy also increases. 57

9 OPEN JOURNAL OF MOBILE COMPUTING AND CLOUD COMPUTING 5. CONCLUSIONS Searching and retrieving the relevant data from the cloud is one of the main challenges in the cloud computing scenario. Privacy-preserving fuzzy search is the search technique for achieving effective utilization of remotely stored encrypted data in Cloud Computing. Here it is introduced an enhanced algorithm for searching and retrieving the data by conjunction/sequence of fuzzy-keyword search. Conjunction of keywords automatically generates ranking according to the relevancy of the retrieved files. It is ensured that the proposed system retrieves highly relevant data by more efficient search, without compromising privacy and security. References [1] L. M. Kaufman, Data Security in the World of Cloud Computing, IEEE Security and Privacy, vol. 7, no. 4, pp , [2] J. Li, Q. Wang, C. Wang, N. Cao, K. Ren, and W. Lou, Fuzzy keyword search over encrypted data in cloud computing, in INFOCOM10 Mini- Conference, 2010 Proceedings IEEE, pp. 1 5, IEEE, [3] D. Boneh, G. Di Crescenzo, R. Ostrovsky, and G. Persiano, Public key encryption with keyword search, in Proc. of EUROCRYP04, [4] S. Ji, G. Li, C. Li, and J. Feng, Efficient interactive fuzzy keyword search, in Proceedings of the 18th International Conference on World Wide Web, pp , ACM, [5] P. Golle, J. Staddon, and B. Waters, Secure Conjunctive Keyword Search Over Encrypted Data, in Applied Cryptography and Network Security, pp , Springer, [6] L. Ballard, S. Kamara, and F. Monrose, Achieving efficient conjunctive keyword searches over encrypted data, in Information and Communications Security, pp , Springer, [7] T. Nisha and V. Lijo, Improving the Efficiency of Data Retrieval in Secure Cloud, International Journal of Computer Applications, vol. ICACT Nov , [8] S. Ji, G. Li, C. Li, and J. Feng, Efficient interactive fuzzy keyword search, in Proceedings of the 18th International Conference on World Wide Web, pp , ACM,

10 About This Journal MCCC is an open access journal published by Scientific Online Publishing. This journal focus on the following scopes (but not limited to): Autonomic Business Process and Workflow Management in Clouds Cloud Composition, Federation, Bridging and Bursting Cloud Computing Consulting Cloud Configuration, Performance and Capacity Management Cloud DevOps Cloud Game Design Cloud Migration Cloud Programming Models and Paradigms Cloud Provisioning Orchestration Cloud Quality Management and Service Level Agreement (SLA) Cloud Resource Virtualization and Composition Cloud Software Patch and License Management Cloud Workload Profiling and Deployment Control Cloud Video and Audio Applications Economic, Business and ROI Models for Cloud Computing Green Cloud Computing High Performance Cloud Computing Infrastructure, Platform, Application, Business, Social and Mobile Clouds Innovative Cloud Applications and Experiences Security, Privacy and Compliance Management for Public, Private and Hybrid Clouds Self-service Cloud Portal, Dashboard and Analytics Storage, Data and Analytics Clouds Welcome to submit your original manuscripts to us. For more information, please visit our website: You can click the bellows to follow us: Facebook: Twitter: LinkedIn: Google+:

11 SOP welcomes authors to contribute their research outcomes under the following rules: Although glad to publish all original and new research achievements, SOP can t bear any misbehavior: plagiarism, forgery or manipulation of experimental data. As an international publisher, SOP highly values different cultures and adopts cautious attitude towards religion, politics, race, war and ethics. SOP helps to propagate scientific results but shares no responsibility of any legal risks or harmful effects caused by article along with the authors. SOP maintains the strictest peer review, but holds a neutral attitude for all the published articles. SOP is an open platform, waiting for senior experts serving on the editorial boards to advance the progress of research together.