IEEE Consumer Electronics Magazine - January/February 2022 - 81

password encryption and brute-force attack are
both described. We also present the general
architecture of password recovery systems and
introduce some typical password encryption
algorithms.
Password Recovery and Brute-Force Attack
Even in this era, when biometrics, such as finFigure
1. Illustration of a brute-force attack for
password recovery.
in their organization, whichhelps themfind the vulnerability
of their IT infrastructure. Moreover, it is
needed by the law enforcement to recover the digital
evidence that is encrypted by criminals. As the
strength of password encryption algorithm
increases with time, specialized, fast, and efficient
password recovery accelerators are becoming
more of a necessity.
This article examines various password
recovery systems, including the popular hardware
platforms ranging from traditional application-specific
integrated circuits (ASICs) and fieldprogrammable
gate arrays (FPGAs) to modern
graphic process unit (GPUs) and hybrid CPUFPGA
chips, the password guessing methods,
and the acceleration techniques adopted to crack
the password encryption algorithms. The content
of this article can be summarized as follows.
1) An overview about how password recovery
works, the basic components of the password
recovery systems, and the password
encryption algorithms.
2) Various hardware platforms used to build
the password recovery systems are introduced
and compared.
3) Frequently used password guessing methods
are presented.
4) Various password recovery accelerator designs
are presented and the underneath key techniques
are analyzed.
BACKGROUND ON PASSWORD
RECOVERY
In this section, we provide some basic concepts
of password recovery. The scheme of
January/February 2022
gerprints and facial information, are widely used
for authentication, passwords are still holding
their unshakable positions in protecting personal
data, because biometrics are unchangeable
and can be easily observed and captured
while passwords do not have such flaws.1 The
idea of password encryption comes from the
fact that there is no absolutely secure system.
Any system could be invaded and the passwords
could be stolen. If the invaders could get clear
text passwords directly, they will produce
severe damages by accessing sensitive data,
such as user information. As a result, many operating
systems and applications adopt a password
encryption algorithm [also known as
password hash function (PHF)] to transform the
clear text password to a password hash. When
user authentication is initiated, the system transforms
the user's password with the PHF and
compares the generated password hash with the
correct password hash to check the legitimacy
of the user trying to log in. In this way, even if
the attacker holds the password hash, as the
PHF is a one-way function, the attacker will still
experience difficulties in retrieving the actual
clear text password.
Since recovering the clear text password
directly from the password hash is theoretically
impossible, brute-force attack, illustrated in
Figure 1, becomes the only way to accomplish
the goal. The password hash H is generated
from the correct password P. If an attacker
wants to crack the password hash H and find P,
he should first specify a search space that
includes the possible candidates of the correct
password. Then, the attacker applies a PHF on
each password candidate and compares the generated
password hashes (H1;H2,...,HN) with the
password hash H. When HX matches with H,
the correct password is found to be PX. In general,
the size of search space is in the range of
1010 for a recovery probability of about 80%,
81
IEEE Consumer Electronics Magazine - January/February 2022

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - January/February 2022
