TY - GEN
T1 - ConcurORAM
T2 - 26th Annual Network and Distributed System Security Symposium, NDSS 2019
AU - Chakraborti, Anrin
AU - Sion, Radu
N1 - Publisher Copyright:
© NDSS 2019.All rights reserved.
PY - 2019
Y1 - 2019
N2 - ConcurORAM is a parallel, multi-client oblivious RAM (ORAM) that eliminates waiting for concurrent stateless clients and allows overall throughput to scale gracefully, without requiring trusted third party components (proxies) or direct inter-client coordination. A key insight behind ConcurORAM is the fact that, during multi-client data access, only a subset of the concurrently-accessed server-hosted data structures require access privacy guarantees. Everything else can be safely implemented as oblivious data structures that are later synced securely and efficiently during an ORAM “eviction”. Further, since a major contributor to latency is the eviction – in which client-resident data is reshuffled and reinserted back encrypted into the main server database – ConcurORAM also enables multiple concurrent clients to evict asynchronously, in parallel (without compromising consistency), and in the background without having to block ongoing queries. As a result, throughput scales well with increasing number of concurrent clients and is not significantly impacted by evictions. For example, about 65 queries per second can be executed in parallel by 30 concurrent clients, a 2x speedup over the state-of-the-art [13]. The query access time for individual clients increases by only 2x when compared to a single-client deployment.
AB - ConcurORAM is a parallel, multi-client oblivious RAM (ORAM) that eliminates waiting for concurrent stateless clients and allows overall throughput to scale gracefully, without requiring trusted third party components (proxies) or direct inter-client coordination. A key insight behind ConcurORAM is the fact that, during multi-client data access, only a subset of the concurrently-accessed server-hosted data structures require access privacy guarantees. Everything else can be safely implemented as oblivious data structures that are later synced securely and efficiently during an ORAM “eviction”. Further, since a major contributor to latency is the eviction – in which client-resident data is reshuffled and reinserted back encrypted into the main server database – ConcurORAM also enables multiple concurrent clients to evict asynchronously, in parallel (without compromising consistency), and in the background without having to block ongoing queries. As a result, throughput scales well with increasing number of concurrent clients and is not significantly impacted by evictions. For example, about 65 queries per second can be executed in parallel by 30 concurrent clients, a 2x speedup over the state-of-the-art [13]. The query access time for individual clients increases by only 2x when compared to a single-client deployment.
UR - https://www.scopus.com/pages/publications/85091902015
U2 - 10.14722/ndss.2019.23411
DO - 10.14722/ndss.2019.23411
M3 - Conference contribution
AN - SCOPUS:85091902015
T3 - 26th Annual Network and Distributed System Security Symposium, NDSS 2019
BT - 26th Annual Network and Distributed System Security Symposium, NDSS 2019
PB - The Internet Society
Y2 - 24 February 2019 through 27 February 2019
ER -