Searchable symmetric encryption: Sequential scan can be practical

Abstract

The proliferation of cloud computing highlights the importance of techniques that allow both securing sensitive data and flexible data management at the same time. One line of research with this double motivation is the study of Searchable Symmetric Encryption (SSE) that has provided several outstanding results in the recent years. These solutions allow sublinear keyword search in huge databases by using various data structures to store keywords and document identifiers. In this work, we focus on certain scenarios in which search over the whole database is not necessary and show that the otherwise inefficient sequential scan (in linear time) can be very practical. This is due to the fact that adding new entries to the database comes for free in this case while updating a complex data structure without information leakage is rather complicated. To demonstrate the practicality of our approach we build a simple SSE scheme based on bilinear pairings and prove its security against adaptive chosen-keyword attacks in the standard model under the widely used Symmetric eXternal Diffie-Hellman (SXDH) assumption.