| Resumo: | In this dissertation, we researched techniques to support trustable and privacy enhanced solutions for on-line applications accessing to “always encrypted” data in remote DBaaS (data-base-as-a-service) or Cloud SQL-enabled backend solutions. Although solutions for SQL-querying of encrypted databases have been proposed in recent research, they fail in providing: (i) flexible multimodal query facilities includ ing online image searching and retrieval as extended queries to conventional SQL-based searches, (ii) searchable cryptographic constructions for image-indexing, searching and retrieving operations, (iii) reusable client-appliances for transparent integration of multi modal applications, and (iv) lack of performance and effectiveness validations for Cloud based DBaaS integrated deployments. At the same time, the study of partial homomorphic encryption and multimodal searchable encryption constructions is yet an ongoing research field. In this research direction, the need for a study and practical evaluations of such cryptographic is essential, to evaluate those cryptographic methods and techniques towards the materialization of effective solutions for practical applications. The objective of the dissertation is to design, implement and perform experimental evaluation of a security middleware solution, implementing a client/client-proxy/server appliance software architecture, to support the execution of applications requiring on line multimodal queries on “always encrypted” data maintained in outsourced cloud DBaaS backends. In this objective we include the support for SQL-based text-queries enhanced with searchable encrypted image-retrieval capabilities. We implemented a prototype of the proposed solution and we conducted an experimental benchmarking evaluation, to observe the effectiveness, latency and performance conditions in support ing those queries. The dissertation addressed the envisaged security middleware solution, as an experimental and usable solution that can be extended for future experimental testbench evaluations using different real cloud DBaaS deployments, as offered by well known cloud-providers.
|
|---|