Honey Encryption

nd
Line 1: 2 Given Name Surname
line 2: dept. name of the organisation
(of Affiliation)
line 3: Name of the organisation
(of Affiliation)
line 4: City, Country
line 5: email address or ORCID

Abstract— A novel encryption technique,HoneyiEncryption

(HE), addressesithe weaknesses of standard password-based
encryptioni(PBE). A low-min-entropy key may be used to
encrypt the data while the system is resistant to brute-force
attacks. A honey message is created by decrypting a ciphertext
using an incorrect key. With an emphasis on the HE scheme,
we examine the idea of honey objects in this study. Research in
this subject is examined and actual applications that have been
proposed or implemented to date. In the end, the present
limitations of Honey Encryption and some possible future
directions are discussed.
Keywords— iCryptography, HoneyiEncryption (HE),iHoney
Objects,iBrute-force Attack, PasswordiCracking,iPassword-based
Encryptioni(PBE), Distribution-TransformingiEncoderi (DTE)
I. INTRODUCTION
Because users are infamous for choosing weak passwords,
which they typically use across numerous sites, a single
breach can compromise several user accounts. Weak
passwords are the cause of several high-profile data
breaches. One of the most significant password breaches
in history occurred in 2008 at RockYou, and the
credentials were published online. One of the most
popular passwords was "123456", and many of the most
popular ones were very similar. Similar breaches at
Yahoo, Adobe, and LinkedIn attested to the problem's
magnitude.
Websites commonly employ cryptographic hash functions
to protect user names and passwords in the case of a
breach. Many freely accessible hash-cracking tools may be
used to cycle through the most regularly used passwords,
including lists of passwords that have been released
through breaches like RockYou. But this only works if the
users choose strong passwords. It is common practice to
salt user passwords to make them more secure. A random
string called a "salt" is appended or pre-applied to the
password before hashing.
Known for its usage in password managers, password-based
encryption (PBE) is a well-known encryption method. A
master password is used to encrypt and manage the
passwords. A 2011 security breach of LastPass [1] revealed
that encrypted data stored in the cloud is vulnerable to
assault since it is usually stored online. This service is a
prime target for hackers because if the data can be
decrypted, all users' website passwords can be revealed.
Assuming an attacker obtains access to this file, they can
launch a brute-force offline assault that has a decent
probability of succeeding. Even though users are somewhat
at fault, it is reasonable that they create simple to remember
passwords and reuse them across many sites. It is therefore
necessary to look at additional ways to address the
problem, even though many services enforce strict
password requirements.
It is called Honey Encryption (HE), and it's a type of
encryption that goes beyond brute force. Offlineibrute-force
attacksican beicarried out by an adversary who has access to a
hashed password database or a file that has been encrypted
with PBE. When the decryption is successful, it is easy to tell
since the plaintext result will be obvious. The plaintexts
produced by decryption with an incorrect password will not fit
the structure of natural language and will often contain invalid
characters. HE creates plausible but invalid plaintext when the
wrong password is provided, making it difficult for an attacker
to verify whether the decryption was successful.
II. BACKGROUND
A. Honey Objects
Honey is a word used in computer security to describe
decoys. Honey objects include honeypots, a fake server
designed to divert attackers from genuine targets and
perhaps identify them [4]. In order to secure passwords,
honeywords have been proposed as an option. "Sugarword"
is a set of incorrect passwords saved for each user along
with their correct password. User credentials are transmitted
to "honeychecker," a secure server that checks if the user
has typed a honeyword when entering the system [5].
Users' accounts containing honeywords are likely to have
been compromised, and this would set off an alarm that may
prevent them from logging in. Users may choose
honeywords for their accounts; however, this poses an issue
since the user may unintentionally use a honeyword in the
future, which would raise the alarm. A better solution would
be for the server to produce honeywords, but they would
have to be realistic to be helpful. Another example of a
Honey Object is a Honey File.
Upright and secrecy are the two most important properties
of honey items. If a list of honeywords is highly similar yet
distinct from the sugarword, an attacker will have a higher
chance of finding a valid password. For a honeychecker that
keeps an index of which password is the sugarword, the
security offered by the system relies on how well this index
is kept secret from any prospective attackers.
B. Password-based Encryption
This scheme comprises of two functions: enc() and dec() ().
After M is encrypted using P, the result is C (the ciphertext).
Decryption with an erroneous password will result in an invalid
plaintext message, and the user will be alerted that the
password was incorrect. The attacker can try different
passwords until they get a valid plaintext message. Due to
this fundamental weakness in security, PBE-generated
ciphertext is only as safe as an adversary's ability to guess
the password by brute force [2].
Different sorts of sensitive information are encrypted with
PBE. It is often used in conjunction with password
managers, which are intended to assist users in storing
their passwords for multiple online accounts safely.
Because passwords are kept in plaintext rather than hashed
form, they are protected by a master password rather than
a hashing algorithm. It is common for people to use
password managers so that they don't have to remember
passwords for every website, some of which may need
frequent changes or impose stringent password policy.
A single point of failure is introduced by password
managers, even while making life easier for users and
encouraging safe passwords usage for particular websites.
Anyone with access to the encrypted vault may launch a
brute-force assault using a weak master password. This is
likely to be successful given a reasonable time, assuming
that the typical user's master password is as weak as past
password breaches have shown it to be. HE can create a list
of viable passwords when decrypting with the incorrect
password, forcing the adversary to verify login information
from every possible decryption manually.
C. Distribution-Transforming Encoders
This encoding system consists of an encoding algorithm and
a decoding algorithm to encodeiandidecode aimessage. A
message M is sent to the encode function, which encodes it
as a K-bit seed S∈{0, 1}K. There's no need for this
encoding to beiunique; manyiseeds might match M, in
which case one of the seeds will be chosen evenly and at
random. For the decode function to obtain the original
message, encode must be reversible. A message S is
decoded to produce a message M, equivalent to the original
message S. DTE is an excellent paradigm for message
distribution since it is asymmetrical. By utilising the DTE-
then-encrypt technique, it's straightforward to build a solid
high-entropy (HE) scheme.
D. Honey Encryption
Honey Encryption (HE) can prevent an opponent from
learning anything from the ciphertext that will help them
decrypt the plaintext. For security beyond brute-force, it
assures that any decryption of ciphertext with an incorrect
key would give a reasonable appearing plaintext, which the
attacker cannot readily use to identify whether the
decryption has been successful or not. To crack a weak
password, you need to perform a brute-force offline assault
to get the plaintext. This might help reduce the vulnerability
to brute-force assaults. This makes it hard to verify whether
the decryption was successful even when a weak password
had been used.
Its primary security aim is to prevent an attacker from
recovering the original plaintext from a ciphertext with a
high degree of certainty. For this to be achieved, the
message and key must have high entropy.
If both the message and the key are low-entropy, it is still
possible to establish message recovery security using HE.
Semantic security may be achieved by using HE if the key is
sufficiently unexpected, and adversaries are
computationally constrained [2].
E. Hash Cracking Tools
Many hash cracking tools are now readily accessible on the
market. When the passwords are weak, the attackers can
simply crack the hashes with these tools. These
programmes take use of the fact that users' passwords are
generally composed in a certain way. As an example, the
2014 Yahoo data breach was a direct result of the use of
hash cracking tools.
 JohnitheiRipper: JohnitheiRipper is a free and
Open Source password breaking programme. It's a
command-line-only programme for Linux.
 Ophcrack: Ophcrack is a free passwordicracking
programme for Windows that uses the Rainbow
Table. It's also compatible with Linux and Mac.
 Brutus: Many people believe Brutus to be the
fastest online password breaker.
 RainbowCrack: The RainbowCrack programme
cracks hashes by using a rainbow table. It employs
ailarge-scaleitime-memory trade-off to break
passwords quickly and efficiently. It is compatible
with Linux and Windows.
III. CURRENT STATE OF THE ART
A. Credit Card Details and RSA Secret Keys
A few real-world instances of HE may be found in [2]. In
the first instance, credit card numbers, PIN codes, and
CVVs are compatible with HE since the message space is
highly organised. This is even more so when we consider
the size and scope of the CC fraud business, which is fueled
by more skilled cyber-criminals. This means that decryption
efforts during brute-force attacks on PBE-protected CC data
will display when the correct plaintext has been recovered.
How HE works with RSA secret keys is also demonstrated.
It is recommended that users use PBE to encrypt their
private keys in case of a data breach. If a decryption attempt
succeeds owing to the structure and legality of the plaintext,
an attacker might conduct aniofflineibrute-force assault on
the encryptediRSA key. The authorsiuse a solid proof of
theorem to support their HE examples, which also discusses
deployment issues. To ensure that the decrypted result is
valid, it is vital to ensure that the password is entered
correctly.
B. Generic Alphabet and Simple Messaging System
This credit card HE method has been implemented in
practice by researchers in [6]. However, they do not
presume that the outputs must be words in the English
language but instead provide a general HE scheme for an
alphabet such as a-z. "bcde" may be produced by decrypting
a message M = "aabb" using the incorrect key K. A huge
hash table was explored, however, the researchers decided
to make use of each letter's independence. As a result, they
had to create a simple communications scheme that dealt
with a simple format; questions and replies.
There are about 2^154 distinct relevant meanings in
English, even when sentences are confined to 140
characters. Even more challenging is the fact that English
has many dependencies and grammatical restrictions,
mainly when the response is limited to a yes-or-no, as in
this implementation, questions and responses may be
described as having a simple structure. It was possible to
produce over 100,000,000 distinct phrases by combining the
Natural Language Toolkit (nltk) with a bespoke Probability
Distribution Function. In [6], you can find the code for all
three implementations.
C. Kamouflage and NoCrack
When N – 1 decoy vaults are encrypted with decoy master
passwords, Kamouflage provides cracking resistance for
vaults. As much as Kamouflage's theory looks to satisfy
security objectives, researchers in [3] have discovered minor
flaws that degrade security below standard PBE. Decoy
master passwords follow the same template as their real
counterparts. A brute-force offline attack will disclose the
template whenever an attacker obtains any password
(genuine or fake).
Decoy master passwords are chosen evenly in line with the
master password template, which creates a second
vulnerability. Because the distribution of decoy passwords
is different from actual users, an attacker is more likely to
guess the correct password fast if they guess passwords in
order of popularity. Researchers can crack the password
with less effort using these two flaws; therefore,
Kamouflage has decreased security. To break the
Kamouflage vault, they only need 44 per cent as much work
as PBE demands for 10^3, and 24 per cent as much effort as
PBE requires for 10^4.
The decoys should not give away the password. In addition
to breaking the Kamouflage method, the researchers
developed a new system called NoCrack, a honey-vault
service comparable to commercial password vaults such as
LastPass. So as not to introduce vulnerabilities, they
carefully designed and executed a HE scheme and
successfully reaped the security benefits of HE while
incurring minimum performance overheads. Open-source
software, NoCrack's code may be found at [14].
D. Honey Chatting
[7] Introduces a new secure instant messaging system based
on HE. Honey Chatting is a Java programme that encrypts
communications using a statistical coding technique (DTE)
(AES). In the statistical coding system, the transmitter and
receiver utilise the n-gram language model of the text
corpus. Using a text corpus for training data is advantageous
since its literary style largely influences the output of honey
messages. Making the honey messages look like real chat
conversations by using text databases such as movie scripts
with a lot of dialogue instead of descriptions.
It is demonstrated in a demonstration of the software in action
how a recipient with the right password may effectively
decrypt and decode chat messages. An attacker who intercepts
communications and attempts to decode them with the
incorrect password is presented with a collection of invalid but
convincing messages. These tests show the difference between
encrypted text when using the right key and an erroneous key.
According to their experiment study, Honey Chatting's efficacy
is impacted by some elements, including the message's length.
E. GenoGuard
Scientists have developed GenoGuard, a programme that
employs HE methods to safeguard genetic data. This has led
to an increase in the processing of genetic data that may be
utilised in healthcare, research and consumer services, and
legal proceedings or for forensic purposes. Cloud platforms
are being used to store, analyse, and share genetic data by
large corporations like Google. The potential for scientific
and medical development is enormous, but the security and
privacy of the data is also a worry.
Health insurance firms and businesses might use personal
genetic data to discriminate against workers or to refuse
treatment to patients if it falls into the wrong hands. It could
even be used for blackmail if it falls into the wrong hands.
Long-term secure storage of genomic data is complex since
genomic data may be traced back to family members.
Security passwords protect data, but computer power and
password cracking technologies may develop to the point
where decrypting data we presently deem safe is a simple
operation in the future. The majority of earlier research and
implementations of HE relied on basic plaintext structures
such as credit card data and encryption keys. However, this
is changing now. For increasingly complex data structures,
such as genetic data, the authors acknowledged the
challenge of developing a HE scheme. Using a new tree-
based approach, the GenoGuard system encodes genetic
data, then encrypts it with a password given by the patient.
An adversary might utilise genotype-phenotype connections
to discover a victim's genome, an example of a side-channel
assault. As a result, GenoGuard's protection is superior to
traditional HE.
When an attacker gets hold of a patient's genomic data and
attempts to decrypt it with the wrong password, GenoGuard
produces an erroneous genomic sequence that appears
convincing. For service providers that offer DTC services to
store genomes securely, the GenoGuard system was proven
to be very efficient. Developing an efficient and precise
DTE was a challenge for the GenoGuard authors, who
expressed the hope that their work may encourage other
academics to build HE schemes for actual, particular use
cases based on their work.
F. Visual Honey Encryption
As stated in [2], the HE standard cannot be easily applied to
various application domains, such as natural language-
based texts, sounds and pictures. In contrast to the original
system, these data types have a complex synthetic or
semantic structure. HE may be used to complicate
multidimensional data using a 2D Markovian process in a
Bayesian framework, according to [10]. Multidimensional
high-end imaging (HE) is linked to covert data encryption
(steganography) for specialised applications (commonly
multimedia such as images). It's a form of steganography
based on visual honey encryption that they term honey
steganography. To circumvent the necessity of typical HE to
predetermine the DTE, the researchers develop an adaptive
system of DTE.
Additionally, they offer a novel variant of HE that
overcomes the original's practical limitations and
weaknesses, allowing HE to be used to complicate data,
such as pictures. As a final step, an improved steganography
approach was proposed based on visual honey encryption,
which improves the security given by previous
steganography systems, all of which have been cracked by
steganalysis. A picture decoded with a wrong key is
indistinguishable from one decrypted with the correct key
when analysed by a computer, which is the primary goal of
honey steganography.
IV. FUTURE DIRECTIONS
A. Typo-safety
A significant concern with HE implementations is typo
safety. Should an unintentional mistake occur during the
decryption process, the user would be presented with
incorrect but plausible-looking data. As an example, if a
system administrator wants to decrypt a list of usernames
and passwords for a website they control, this might be
problematic if they will be presented with convincing-
looking usernames and passwords if they enter the
decryption password incorrectly, but they may not realise
that the plaintext is wrong until they have done additional
actions, such as migrating the erroneous data to a different
server. A number of typo-safety solutions exist, such as
checksums [2] or dynamic security layers [3], but this is a
topic that might be explored more in the future.
B. Honey Encryption HE Applications
Due to the complexity of DTE creation for each unique use
case, HE is still a relatively new idea with little practical
applicability at this time. There are a limited number of HE
implementations that focus on sophisticated natural
English language. The development of DTEs for highly
structured data may be given more attention. Due to
Kamouflage [3], what appears to be a secure way of using
high-encryption (HE) can deteriorate security beyond
conventional PBE if not adequately verified through attack
simulation.
V. CONCLUSION
Weak passwords in PBE is a persistent concern in computer
security and will continue for many years. When paired with
computer speed improvements, password cracking
techniques are projected to improve, allowing more
encrypted data to be hacked by brute force. A remedy to this
dilemma may be found in HIM's teachings. To demonstrate
how traditional PBE may be made more secure, we
proposed the idea of HE in this work. Some of the most
cutting-edge research in this field was discussed, and some
of the actual implementations created to date.
To minimise data loss due to leakage and abuse, private
information should be adequately safeguarded. Due to the
output of the decryption process, brute-force assaults can
detect if a password-basediencryptioni(PBE) mechanism
used to secure private data is valid or not. This issue may be
countered by using the honey encryption method. It is
possible to design and build applicationsiwithiuniformly or
nonuniformlyidistributedimessageispaces, as well as
symmetricior asymmetricikeyiencryption methods. It was
determined that our honey encryption technique was not
performing as well as it should, and a solution was
recommended to reduce the overhead.
For honey encryption to work well, the message space must
be carefully defined for each application. Honey
encryption's capacity to secure sensitive private information
differs depending on the application. Due to the fact that the
message space varies, honey encryption must be configured
for each application.
REFERENCES
[1] Ieeexplore.ieee.org. 2021. Honey Encryption:
Encryption beyond the Brute-Force Barrier. [online]
Availableiat:<http://ieeexplore.ieee.org/abstract/document/6
876246/> [Accessed 13 August 2021].
[2] Juels, A. and Ristenpart, T., 2021. Honey Encryption:
Security Beyond the Brute-Force Bound.
[3] Ieeexplore.ieee.org. 2021. Cracking-Resistant Password
Vaults Using Natural Language Encoders. [online]
Availableiat:<http://ieeexplore.ieee.org/abstract/document/7
163043/> [Accessed 13 August 2021].
[4] Dl.acm.org. 2021. A bodyguard of lies | Proceedings of
the 19th ACM symposium on Access control models and
technologies. [online] Available at:
<http://dl.acm.org/citation.cfm?id=2613088> [Accessed 13
August 2021].
[5] Dl.acm.org. 2021. Honeywords | Proceedings of the
2013 ACM SIGSAC conference on Computer &
communications security. [online] Available at:
<http://dl.acm.org/citation.cfm?id=2516671> [Accessed 13
August 2021].
[6] Courses.csail.mit.edu. 2021. [online] Available at:
<https://courses.csail.mit.edu/6.857/2016/files/tyagi-wang-
wen-zuo.pdf> [Accessed 14 August 2021].
[7] Ieeexplore.ieee.org. 2021. Honey chatting: A novel
instant messaging system robust to eavesdropping over
communication. [online] Available at:
<http://ieeexplore.ieee.org/abstract/document/7472064>
[Accessed 14 August 2021].
[8] Ieeexplore.ieee.org. 2021. GenoGuard: Protecting
Genomic Data against Brute-Force Attacks. [online]
Available at:
<http://ieeexplore.ieee.org/abstract/document/7163041/>
[Accessed 14 August 2021].

[9] Onlinelibrary.wiley.com. 2021. Screening the covert key

using honey encryption to rule out the brute force attack of
AES—a survey. [online] Available at:
<http://onlinelibrary.wiley.com/doi/10.1002/sec.1753/full>
[Accessed 14 August 2021].

[10] Dl.acm.org. 2021. Visual Honey Encryption

Proceedings of the 3rd ACM Workshop on Information
Hiding and Multimedia Security. [online] Available at:
<http://dl.acm.org/citation.cfm?id=2756606> [Accessed 14
August 2021].

[11] Eprint.iacr.org. 2021. [online] Available at:

<http://eprint.iacr.org/2013/696.pdf> [Accessed 14 August
2021].

[12] International Journal of Biology, Pharmacy and Allied

Sciences, 2021. https://ijbpas.com/archive/archive-detail-
pdf/VOLUME-10-ISSUE-1. 10(1).

[13] G, S., 2020.

https://medwinpublishers.com/NNOA/NNOA16000183.pdf.
Nanomedicine & Nanotechnology Open Access, 5(2).

[14] BU, M., 2021. An Intro to Honey Encryption:

Cryptographic Parlor Tricks for Passwords | McAfee Blogs.
[online] McAfee Blogs. Available at:
<https://www.skyhighnetworks.com/cloud-security-
blog/cryptographic-parlor-tricks-for-passwords-an-
introduction-to-honey-encryption/> [Accessed 15 August
2021].

[15] Threatpost.com. 2021. Honey Encryption Tricks

Hackers with Decryption Deception. [online] Available at:
<https://threatpost.com/honey-encryption-tricks-hackers-
with-decryption-deception/103950/> [Accessed 15 August
2021].