TrApps

Abstract

The cloud computing paradigm enables the flexible and scalable outsourcing of workloads. However, cloud customers are often reluctant to entrust their sensitive data with cloud providers. This is due to the fact that the infrastructure is owned by another company and a resulting loss of control. With the recent advent of powerful ARM hardware targeted for data centres, there is the opportunity of using trusted execution technology provided by ARM TrustZone to enhance the protection of cloud customer's data. In this paper we propose TrApps, a secure platform for generalpurpose trusted execution in an untrusted cloud with multiple isolated tenants based on the ARM TrustZone technology. Our system targets the parallel execution of partitioned applications of distinct tenants with lean security-sensitive components, and is based on a minimal trusted code base in the secure world of ARM TrustZone when compared to similar systems. In our evaluation we show the feasibility of our approach, and demonstrate its performance with trusted execution of memcached with an overhead of only 36.9% compared to the vanilla implementation and execution.

Managing confidentiality leaks through private algorithms on Software Guard eXtensions (SGX) enclaves

“…The Horizon2020 funded research projects under SERECA 3 focus on a number of goals to build secure enclaves. These goals include application partitioning [36], trusted architectures for web services [8,33], container architecture [3] and library support for unmodified applications (SGX-LKL 4 ), better integrity [5] and isolation [7], and enclave memory safety [35] in the cloud.…”