Weizmann Logo
ECCC
Electronic Colloquium on Computational Complexity

Under the auspices of the Computational Complexity Foundation (CCF)

Login | Register | Classic Style



REPORTS > DETAIL:

Paper:

TR18-109 | 29th May 2018 13:10

Yes, There is an Oblivious RAM Lower Bound!

RSS-Feed




TR18-109
Authors: Kasper Green Larsen, Jesper Buus Nielsen
Publication: 4th June 2018 19:57
Downloads: 220
Keywords: 


Abstract:

An Oblivious RAM (ORAM) introduced by Goldreich and Ostrovsky
[JACM'96] is a (possibly randomized) RAM, for which the memory access
pattern reveals no information about the operations
performed. The main performance metric of an ORAM is the bandwidth
overhead, i.e., the multiplicative factor extra memory blocks that must be
accessed to hide the operation sequence. In their seminal paper
introducing the ORAM, Goldreich and Ostrovsky proved an amortized
$\Omega(\lg n)$ bandwidth overhead lower bound for ORAMs with memory
size $n$. Their lower bound is
very strong in the sense that it applies to the ``offline'' setting in
which the ORAM knows the entire sequence of operations ahead of time.

However, as pointed out by Boyle and
Naor [ITCS'16] in the paper ``Is there an oblivious RAM lower bound?'', there are two caveats with the lower bound of
Goldreich and Ostrovsky: (1) it only applies to ``balls in bins'' algorithms,
i.e., algorithms where the ORAM may only shuffle blocks around and
not apply any sophisticated encoding of the data, and (2), it only
applies to statistically secure constructions. Boyle and Naor showed
that removing the ``balls in bins'' assumption would result in super
linear lower bounds for sorting circuits, a long standing open problem
in circuit complexity. As a way to circumventing this barrier, they also
proposed a notion of an ``online'' ORAM, which is an ORAM that remains
secure even if the operations arrive in an online manner. They argued that most
known ORAM constructions work in the online setting as well.

Our contribution is an
$\Omega(\lg n)$ lower bound on the bandwidth overhead of any online
ORAM, even if we require only computational security and allow arbitrary representations of data, thus greatly
strengthening the lower bound of Goldreich and Ostrovsky in the online setting.
Our lower bound applies to ORAMs with memory size $n$ and any word size
$r \geq 1$. The bound therefore asymptotically matches the
known upper bounds when $r = \Omega(\lg^2 n)$.



ISSN 1433-8092 | Imprint