Abstract:
Information security has become one of the major focus areas for any organization.
More often, organizations see the need of outsourcing their data storages in meeting
the operational and security objectives. This gives rise to a new problem of privacy
protection of the data stored with a third party. As a solution the data is encrypted
before storing with a third party data service provider. Thus when the users need to
process the data, the safer option is to download the data into a secure user machine
and perform the operations on the decrypted data. This creates an additional overhead
of having to download a large amount of data and decrypt them even to perform a
simple calculation on the data stored in the encrypted form. Therefore the possibility
of secure data processing at the remote third party storage has become an interesting
problem to solve. In order to preserve the privacy the data cannot be allowed to be
decrypted at the third party storage. One form of the solution is to facilitate computations
on the data stored in encrypted form. The users can make requests from the data
service provider and if the service provider can perform operations on the encrypted
data itself and provide the answer the above mentioned overhead can be avoided. This
brings the focus of this research on to the studying of computing on encrypted data
with specific focus on searchable encryption. As pert of the research, the current literature
of computing on encrypted data is studied to identify a suitable searchable
encryption scheme for practical use. Followed by the literature study, an existing symmetric
searchable encryption scheme is selected for a detailed study. Here a complete
implementation of the scheme is proposed and the test results are analyzed. Based on
the results, a keyword extraction mechanism is proposed to improve the performance
of the scheme. Finally significant performance improvements, 89.83% reduction in
extra space usage due to searchable encryption and 92.11% improvement in single key
word search time has been achieved. In addition to that, use cases in capital markets
are studied to understand the possibilities of practical use and challenges.
