A Trusted and Privacy-Enhanced In-Memory Data Store

• Main Author: Anciães, Miguel de Lemos Dias Rosa
• Publication Date: 2021
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10362/130409
• topic: Hardware Security; Privacy-Enhanced Data Store; Homomorphic Encryption; Isolated Environments; Trusted Computing; Cloud Computing; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática

The recent advent of hardware-based trusted execution environments provides isolated execution, protected from untrusted operating systems, allowing for the establishment of hardware-shielded trust computing base components. As the processor provides such a “shielded” trusted execution environment (TEE), their use will allow users to run appli cations securely, for example on the remote cloud servers, whose operating systems and hardware are exposed to potentially malicious remote attackers, non-controlled system administrators and staff from the cloud providers. On the other hand, Linux containers managed by Docker or Kubernetes are interesting solutions to provide lower resource footprints, faster and flexible startup times, and higher I/O performance, compared with virtual machines (VM) enabled by hypervisors. However, these solutions suffer from soft ware kernel mechanisms, easier to be compromised in confidentiality and integrity as sumptions of supported application data. In this dissertation we designed, implemented and evaluated a Trusted and Privacy-Enhanced In-Memory Data Store, making use of a hardware-shielded containerised OS-library to support its trust-ability assumptions. To support large datasets, requiring data to be mapped outside those hardware-enabled con tainers, our solution uses partial homomorphic encryption, allowing trusted operations executed in the protected execution environment to manage in-memory always-encrypted data, that can be or not mapped inside the TEE.