Publication

Abstract : With the advent of Cloud computing as a dominant computing platform, data owners are provoked to handle complex and sensitive data by moving data from local sites to public cloud which provides great flexibility and low cost. For providing data discretion and privacy, unlike traditional data usage based on plaintext keyword search, sensitive data have to be encrypted before outsourcing. Thus, enabling searchable encryption service is of preeminent importance. Having the large number of data users and documents in cloud, it is necessary for the search service to provide optimal search with single keyword query and effective document retrieval based on relevance score of keywords. Related works on privacy preserving searchable encryption focus on single keyword search, Fuzzy keyword search or Boolean keyword search with or without similarity ranking. Cloud handles huge set of documents. When a cloud user posts a query to the cloud, it is known that the user would be retrieved with both relevant and irrelevant documents with which the user has to post-process the documents and sort out to get the relevant documents. This whole process is cumbersome to the user. In this paper, an architecture which provides an optimal search over encrypted cloud data has been proposed which allows the user to retrieve only relevant documents with respect to the query keyword. Elliptic Curve Cryptography (ECC) algorithm has been used for encryption of documents thereby securing the documents and Diffie-Hellman algorithm to check for authentication of key exchange. We have used inverted index to index each document. After the relevant documents are retrieved, we then rank each document based on TF-IDF scoring and provide the top K retrievals to the users. © Research India Publications.