Abstract: With the advent of communication systems and high speed data networks, the amount of data and files an individual or an organization has to maintain and access have also increased multi fold. But with the limited storage space available in a mobile device or a laptop it is not feasible to store all the files in the personal systems. Hence we require a common file sharing platform. These systems provide functionality to data owners to upload their private data and users to search and retrieve the required files. But performing search and retrieval while preserving data privacy is a challenge as data is stored globally by a third party. To improve data security almost all file sharing platforms store the data in an encrypted manner. For enabling search and retrieval functionality, it is required to create a searchable index which contains the keywords present in the files and its respective weight or relevance in the file. The initial systems only supported single keyword search and the keywords were not encrypted as computations on encrypted data is complex. To enhance security we propose a system that implements two-round search with multi-keyword search support. The keywords are also encrypted to enhance the security. To improve efficiency and system performance a user cache is implemented which helps is removing the duplicate files.

Keywords: Homomorphic encryption, Multi-Keyword based file retrieval, relevance search