Recent databases are implemented as in-memory columnstores. Adjustable encryption offers a solution to encrypted database processing in the cloud.We show that the two technologies play well together by providing an analysis and prototype results that demonstrate the impact of mechanisms at the database side (dictionaries and their compression) and cryptographic mechanisms at the adjustable encryption side (order-preserving, homomorphic, deterministic and probabilistic encryption).

Abstract. In order to perform a join in a deterministically, adjustably encrypted database one has to re-encrypt at least one column. The problem is to select that column that will result in the minimum number of re-encryptions even under an unknown schedule of joins. Naive strategies may perform too many or even infinitely many re-encryptions. We provide two strategies that allow for a much better performance. In particular the asymptotic behavior is O(n 3/2 ) resp. O(n log n) re-encryptions for n columns. We show that there can be no algorithm better than O(n log n). We further extend our result to element-wise re-encryptions and show experimentally that our algorithm results in the optimal cost in 41% of the cases.

The software-as-a-service (SaaS) market is growing very fast, but still many clients are concerned about the confidentiality of their data in the cloud. Motivated hackers or malicious insiders could try to steal the clients' data. Encryption is a potential solution, but supporting the necessary functionality also in existing applications is difficult. In this paper, we examine encrypting analytical web applications that perform extensive number processing operations in the database. Existing solutions for encrypting data in web applications poorly support such encryption. We employ a proxy that adjusts the encryption to the level necessary for the client's usage and also supports additively homomorphic encryption. This proxy is deployed at the client and all encryption keys are stored and managed there, while the application is running in the cloud. Our proxy is stateless and we only need to modify the database driver of the application. We evaluate an instantiation of our architecture on an exemplary application. We only slightly increase page load time on average from 3.1 seconds to 4.7. However, roughly 40% of all data columns remain probabilistic encrypted. The client can set the desired security level for each column using our policy mechanism. Hence our proxy architecture offers a solution to increase the confidentiality of the data at the cloud provider at a moderate performance penalty.

No abstract

Processing encrypted queries in the cloud has been extended by CryptDB's approach of adjustable onion encryption. This adjustment of the encryption entails a translation of an SQL query to an equivalent query on encrypted data. We investigate in more detail this translation and in particular the problem of selecting the right onion layer. Our algorithm extends CryptDB's approach by three new functions: configurable onions, local execution and searchable encryption. We have evaluated our new algorithm in a prototypical implementation in an in-memory column store database system.