Home // DBKDA 2024, The Sixteenth International Conference on Advances in Databases, Knowledge, and Data Applications // View article

MosaicDB: An Efficient Trusted / Untrusted Memory Management for Location Data in Database

Authors:
Tomoya Suzuki
Taisho Sasada
Yuzo Taenaka
Youki Kadobayashi

Keywords: Database; Trusted Execution Environment; Intel SGX; Location data; Endpoint Privacy Zones; Cloud Computing.

Abstract:
Location data has been used for various purposes in digitized society but includes waypoints directly related to personal privacy, such as home addresses. To hide such sensitive waypoints, some applications provide Endpoint Privacy Zones (EPZs) that keep a portion of the track secret. However, most service providers placing databases on a cloud face the potential risk of exposing sensitive waypoints to cloud service providers. Existing studies have proposed databases using Trusted Execution Environments (TEE) that protects sensitive data in a trusted and completely secure memory region. However, as TEE inevitably limits the size of the trusted memory, databases proposed in these studies cannot use the trusted memory efficiently due to the fundamental design that handles all data in the trusted memory. Moreover, the memory outside of the trusted memory, called untrusted memory, remains vacant even if the trusted memory is fully used, thereby leading to insufficient memory utilization on a whole system and decreased database performance. In this study, we propose MosaicDB, a memory-efficient and trusted database for location data, using both trusted and untrusted memory. To enhance memory utilization efficiency, MosaicDB handles only sensitive waypoints within the EPZ in the trusted memory while handling non-sensitive waypoints in the untrusted memory. Experimental results show that MosaicDB improves memory utilization efficiency, thereby achieving a 25% reduction in execution time for selection queries compared to the database that handles all data in the trusted memory.

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2024

Publication date: March 10, 2024

Published in: conference

ISSN: 2308-4332

ISBN: 978-1-68558-138-1

Location: Athens, Greece

Dates: from March 10, 2024 to March 14, 2024