Searchable Encryption in Apache Cassandra

In today’s cloud computing applications it is common practice for clients to outsource their data to cloud storage providers. That data may contain sensitive information, which the client wishes to protect against this untrustworthy environment. Confidentiality can be preserved by the use of encryption. Unfortunately that makes it difficult to perform efficient searches. There are a couple of different schemes proposed in order to overcome this issue, but only very few of them have been implemented and tested with database servers yet. While traditional databases usually rely on the SQL model, a lot of alternative approaches, commonly referred to as NoSQL (short for “Not only SQL”) databases, occurred in the last years to meet the new requirements of the so called “Web 2.0”, especially in terms of availability and partition tolerance. In this paper we implement three different approaches for searching over encrypted data in the popular NoSQL database Apache Cassandra (offered by many cloud storage providers) and run tests in a distributed environment. Furthermore we quantify their performances and explore options for optimization.