Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Modified Blowfish Algorithm

Adwaith Samod T Devika TV Hareesh P
Department of Computer Science Department of Computer Science Department of Computer Science
and Engineering and Engineering and Engineering
Mar Athanasius College of Engineering Mar Athanasius College of Engineering Mar Athanasius College of Engineering
Ernakulam, India Ernakulam, India Ernakulam, India

Leya Elizabeth Sunny

Assistant Professor, Department of Computer Science and Engineering
Mar Athanasius College of Engineering
Ernakulam, India

Abstract:- Every time a sender and receiver communicate, metric encryption. Popularly used cryptographic algorithms for
se-curity has always been a major concern. Numerous symmetric encryption include DES, 3DES, AES, Rivest Shamir
cryptographic techniques, such as AES (Advanced Adleman (RSA), and Blowfish. Each has strengths and
Encryption Standard), Data Encryption Standard (DES), weaknesses. Experimental findings and comparisons showed
Triple DES (3DES), Blowfish, etc., are employed to prevent that the Blowfish algorithm was the best among these when
security breaches. We make an effort to change the current taking time into account.
Blowfish method based on various factors, including time,
the avalanche effect, and non linearity. The Blowfish algorithm was initially developed by Bruce
Schneier in 1994 to replace the antiquated DES. The 64-bit
By combining the XOR and addition operations symmetric key block cypher with changeable length is a
employed in the original approach, the ”f” function is distinguishing feature of blowfish. With the exception of when
adjusted. Sbox is also changed. Text, image, audio, and changing keys, Blowfish provides a free alternative to currently
video files are encrypted and decrypted using the modified used encryption algorithms that have varying levels of security.
blowfish method, which is then tested. We generate and To test the security feature and speed provided by Blowfish,
examine nine cases. The findings of all the tests performed numerous studies conducted performance comparisons based
on these cases point to the conclusion that the modified on various assessment factors. The findings revealed that it is
Blowfish method is more compact and safe than the earlier definitely quick and secure.
because of the security of the modified algorithm with
various cases. The current standard requires a minimum of 128-bit block
size, which makes Blowfish unsuitable because it can only
Keywords:- Modified Blowfish Algorithm, Encryption, F- accommodate 64-bit blocks. This property is seen as
function, Multimedia Encryption. undesirable because it may lead to duplicate blocks that will
eventually make other types of attacks possible, compromising
I. INTRODUCTION data security. Blowfish is regarded as a remarkably fast block
cypher. Although Twofish, a Blowfish-related technique,
Information security has emerged as a crucial concern in supports 128-bit block sizes and offers a high level of security,
business, industry, and administration as a result of the rapid it is slower than Blowfish at encrypting data. The block size of
growth in digital communication and the sharing of electronic Blowfish has been extended to 128 bits by a number of
data. Security is the main concern for every communication researchers; however, the results show a significant increase in
between sender and recipient in the modern era. Sender and time and a requirement for more memory, which makes the
recipient would suffer significant losses if there are any security performance less favourable for use in applications that
lapses during communication. Today’s cryptography provides prioritise speed and renders it inefficient for use in small
a variety of crucial methods for securing data and safeguarding devices with limited memory.
information.
II. RELATED WORKS
By applying specific mathematical computational proce-
dures to change text into unintelligible form, cryptography A. Derivation Cases
offers a way to protect sensitive information. With the right key, Two s-boxes instead of one prevents symmetry in the pro-
the text may then be transformed back into readable form. cess. The process retained the overall structure of the blowfish
Utilizing cryptography provides information communication algorithm but provided two derivation in the f function. The
that is secure, private, and confidential. Data privacy is guar- second derivation proved to have better performance in terms
anteed by cryptographic methods using symmetric and asym- of avalanche effect and time. The time for key generation is

IJISRT22JUN1497 www.ijisrt.com 1757

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
independent of the input size while the time for encryption and and XOR is used. The only new operations for the four indexed
decryption depends upon the file size. This algorithm with 128 arrays are data lookups each round.
bit block size and 128 bit key size can be used to encrypt text,
image and other types in addition to electronic medical data.

B. Text Cryptography based on Modified Blowfish Algorithm

Text in digital platforms can be encrypted using modified
blowfish algorithm and Lempel-Ziv-Welch (LZW). Blowfish
uses variable keylength of 32 to 448 bits. Modified blowfish is
the fastest out of the two in encryption and decryption. These
algorithms provide scope for combining both these algorithms
to provide better security. These algorithms provide fast
encryption and decryption without compromising security in
digital platforms such as e-commerce portals, online marketing,
electronic banking etc.

C. Message encryption technique based on enhanced blowfish

algorithm
By lowering the number of rounds and raising the block
length with a constant length during encryption and decryption,
as well as adding a transformation method on some rounds, an
improved version of the Blowfish algorithm is created. The
Blowfish Algorithm has been improved, resulting in a
significant reduction in execution time without sacrificing the
complexity of the encrypted file content. The algorithm runs
Fig. 1. Fiestel Structure of Blowfish
more quickly because of the reduction in iterations.
Additionally, the complexity of the encrypted file’s content was
C. Multimedia encryption
increased by block size improvements and other manipulations
In order to create an interactive presentation, multimedia
such the use of byte splitting, block size transposition, shift
mixes several content types as text, audio, photos, animations,
rows, and mix rows.
and video. Security of multimedia applications is a crucial and
difficult topic since they are especially easy to intercept over
III. BACKGROUND
wireless networks. The primary enabling technology for
maintaining secrecy and preventing unauthorised access to the
A. Cryptography
content is multimedia encryption.
The study of secure communication methods, such as
encryption, that only the message’s sender and intended
D. Fisher-Yates Shuffle
recipient can access, is known as cryptography. The word is
Ronald Fisher and Frank Yates proposed the Fisher-Yates
derived from kryptos, a concealed word in Greek. It is closely
shuffle (FYS), a technique for producing a random permutation
related to encryption, which is the process of converting plain
of a finite linear array. Every variation of FYS in an array
text into ciphertext before sending it and then back again after
produces results that are identical. The Fisher-Yates shuffle in
receiving it. The obscuring of information in photographs using
its contemporary form is introduced by Richard Durstenfeld.
methods like microdots or merging is also covered by
Donald E. Knuth, in his groundbreaking book The Art of Com-
cryptography. The most typical use of cryptography is to
puter Programming, popularised the work of Durstenfeld. The
encrypt and decrypt email and other plain-text messages while
more efficient version of today is an in-place shuffle, which
sending electronic data.
uses no additional storage space and only time proportionate to
the number of elements shuffled.
B. Blowfish Algorithm
Blowfish algorithm is a symmetric block cipher. A secret
—To shuffle an array a of n elements (indices 0...n-1):
key with a variable length that can be any length up to 448 bits
is used with a 64-bit data block cypher. The algorithm consists
for i from n-1 down to 1 do
of pieces for both data encryption and key expansion. The
expansion key converts a key with a maximum length of 448
j ← random integer such that 0 j i exchange a[j] and a[i]
bits into several subkey arrays with a combined length of 4168
bytes. Data is encrypted using a 16-round Feistel Network. It
has a key-dependent permutation and a key-and-data-dependent
switch for each round. For all operations, a 32-bit word addition

IJISRT22JUN1497 www.ijisrt.com 1758

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
E. Avalanche effect in Cryptography B. S-box permutation
The term ”avalanche effect” refers to a particular manner The existing S-box of blowfish algorithm is shuffled using
in which mathematical operations employed in encryption Fisher Yates Shuffle algorithm for generating a random per-
behave. One of the desirable properties of any encryption muation of the S-box. This randomness can enhance the non
scheme is the avalanche effect. A small change in the plain text linearity parameter of the S-box.
or the key should cause a large change in the encrypted text.
Avalanche effect is the name for this characteristic. It measures C. Case 1
the impact of a modest alteration to the plain text or the key on
the cipher-text. The following relation should always be met by
a decent encryption algorithm:

Avalanche effect > 50%

The effect makes sure that a plain-text cannot be readily

predicted by an attacker using statistical analysis. That is, the
encrypted text can be easily decrypted if a change in a single bit
of the input causes a change in only a single bit of the desired
output.

F. Non Linearity
Boolean function non-linearity standards are divided into
groups based on how well they work with cryptographic
architecture. A criterion invariant is left by the largest
transformation group while classifying an object. These
functions concurrently have minimum correlation to affine Fig. 2. Case 1
functions, maximum distance to linear structures, and both
maximum distance to affine functions and maximum distance In this case F(xL) can be calculated as:
to structures. Frequently, the Walsh-Hadamard Transform
(WHT), which makes use of the sign function, is used to assess F(xL)=((S1 + S2)mod 232)XOR((S3+S4)mod 232)
the nonlinearity. The highest level of nonlinearity is 120.
The 32 bit text is split into 4 parts with 8 bit each and used
IV. IMPLEMENTATION to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with second and XOR ed
A. Blowfish Algorithm with the result of ADD function between the third and fourth
64 bit input plaintext is divided into 32 bit left(LE0) and elements. This gives a 32 bit text which is the output of the
right (RE0). LE0 is XORed to P1 in the P-array. All entries in function.
the P-array are 32-bit. The result is given as input to the F-
function. The output of the F-function is XOR ed with
right(RE0). The current left and right are interchanged. This
process continues until the number of rounds reaches 16 or its
corresponding value. Finally the left and right are XOR ed with
the last two values of the P-array and XORed with each other
to produce the cipher text. The reverse process is used to
decrypt the cipher text into the plaintext. Round: A round
consists of mainly four functions

• LE0 is XORed with the number of P-array (corresponding

to the number of round)
• The output is given as input to the F-function
• The output from the F-function is XORed with RE0 gives
next RE
• LE and RE s are exchanged

(In the funcion the S-box consists of 256 entries of 32 bit

each)
Fig. 3. Case 1 Avalanche Effects

IJISRT22JUN1497 www.ijisrt.com 1759

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
D. Case 2 E. Case 3
In this case F(xL) can be calculated as:

F(xL)=(((S1 + S3)mod 232))XOR((S2+S4)mod 232)

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with third and XOR ed with
the result of ADD function between the second and fourth
elements. This gives a 32 bit text which is the output of the
function.

Fig. 4. Case 2

In this case F(xL) can be calculated as:

F(xL)=((S1 XOR S2)mod 2322)+((S3 XOR S4)mod 232)

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is XOR ed with second and ADD ed
with the result of XOR function between the third and fourth
elements. This gives a 32 bit text which is the output of the Fig. 6. Case 3
function.

Fig. 5. Case 2 Avalanche Effects Fig. 7. Case 3 Avalanche Effects

IJISRT22JUN1497 www.ijisrt.com 1760

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
F. Case 4
In this case F(xL) can be calculated as:

F(xL)=(((S1 XOR S3)mod 232))+((S2 XOR S4)mod 232).

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is XOR ed with third and ADD ed with
the result of XOR function between the second and fourth
elements. This gives a 32 bit text which is the output of the
function.

Fig. 10. Case 5

Fig. 8. Case 4

Fig. 11. Case 5 Avalanche Effects

G. Case 5
In this case F(xL) can be calculated as:

F(xL)=(((S1 + S2)mod 232))XOR((S3+S4)mod 232)) ˆ (((S1

XOR S3)mod 232))+((S2 XOR S4)mod 232))

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with second and XOR ed
with the result of ADD function between the third and fourth
elements.The result is then XORed with the result of the first
element from S-box is XOR ed with third and ADD ed with the
result of XOR function between the second and fourth
Fig. 9. Case 4 Avalanche Effects elements. This gives a 32 bit text which is the output of the
function.

IJISRT22JUN1497 www.ijisrt.com 1761

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
H. Case 6
In this case F(xL) can be calculated as:

F(xL)=(((S1 + S2)mod 232))XOR((S3+S4)mod 232)) ˆ (((S1

XOR S2)mod 232))+((S3 XOR S4)mod 232))

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with second and XOR ed
with the result of ADD function between the third and fourth
elements. The result is then XORed with the result of the first
element from S-box is XOR ed with second and ADD ed with
the result of XOR function between the third and fourth
elements. This gives a 32 bit text which is the output of the
function.

Fig. 14. Case 7

Fig. 12. Case 6

Fig. 15. Case 7 Avalanche Effects

I. Case 7
In this case F(xL) can be calculated as:

F(xL)=(((S1 + S2)mod 232))XOR((S3+S4)mod 232)) ˆ (((S1 +

S2)mod 232))XOR((S3+S4)mod 232))

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with second and XOR ed
with the result of ADD function between the third and fourth
elements. The result is then XORed with the result of the first
element from S-box is ADD ed with second and XOR ed with
the result of ADD function between the third and fourth
elements. This gives a 32 bit text which is the output of the
function.

Fig. 13. Case 6 Avalanche Effects

IJISRT22JUN1497 www.ijisrt.com 1762

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
J. Case 8
In this case F(xL) can be calculated as:

F(xL)=(((S1 + S2)mod 232))XOR((S3+S4)mod 232)) ˆ (((S1 +

S3)mod 232))XOR((S2+S4)mod 232))

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with second and XOR ed
with the result of ADD function between the third and fourth
elements. the result is then XORed with the result of the first
element from S-box is ADDed with third and XOR ed with the
result of ADD function between the second and fourth
elements. This gives a 32 bit text which is the output of the
function.

Fig. 18. Case 9

Fig. 16. Case 8

Fig. 19. Case 9 Avalanche Effects

K. Case 9
In this case F(xL) can be calculated as:

F(xL)=(((S1 + S2)mod 232))XOR((S3+S4)mod 232)) ˆ (((S11 +

S13)mod 232))XOR((S12+S14)mod 232))

The 32 bit text is split into 4 parts with 8 bit each and used
to find corresponding S-box 32 bit values for each. Then the
first element from S-box is ADD ed with second and XOR ed
with the result of ADD function between the third and fourth
elements. the result is then XORed with the result of the first
element from S-box2 is XOR ed with third and ADD ed with
the result of XOR function between the second and fourth
elements. This gives a 32 bit text which is the output of the
function.
Fig. 17. Case 8 Avalanche Effects

IJISRT22JUN1497 www.ijisrt.com 1763

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
V. RESULT VII. CONCLUSION

The proposed algorithm uses a different f function and s- Modification have been done in the f function of the blow-
box to encrypt and decrypt different types of files like text, fish algorithm to test out its implication in the parameters of the
images, audio and video. The avalanche effect and non-linearity algorithm for comparison. The use of modified f function and
is found to be slightly increased than original blowfish s-box to encrypt and decrypt different types of files like text,
algorithm since the proposed algorithm includes complex images, audio and video has shown promising results in the
calculations which increases the security. The modified performance of Blowfish Algorithm by increasing avalanche
blowfish algorithm obtained avalanche effect of 53.246 effect and non-linearity with little to no difference in time.
percentage and non lin-earity of 117.343. Different cases of f-function have been implemented to
illustrate and compare the different modification of the
VI. FUTURE RESEARCH blowfish algorithm and the results suggest that case 5 with an
average avalanche effect of 53.24 tend to perform better than
Promising results of modified blowfish algorithm shows the others.
the possibility of testing its application in cloud computing.
Testing out the implication of the modified algorithm by REFERENCES
exploiting a real world application would show the significance
of the performance. [1]. V. Poonia and N. S. Yadav, ”Analysis of modified
Blowfish algorithm in different cases with various
parameters,” 2015 International Conference on Advanced
Computing and Communication Systems, 2015, pp. 1-5
[2]. Corpuz, Reynaldo Gerardo, Bobby Medina, Ruji. (2018).
A Modified Approach of Blowfish Algorithm Based On
S-Box Permutation using Shuffle Algorithm. 140-145.
10.1145/3301326.3301331.
[3]. Quilala, Theda Flare Quilala, Rogel. (2021). Modified
Blowfish algorithm analysis using derivation cases.
Bulletin of Electrical Engineering and Informatics. 10.
2192-2200. 10.11591/eei.v10i4.2292.
[4]. S. Akhter and M. B. Chowdhury, ”Bangla and English
Text Cryptography Based on Modified Blowfish and
Lempel-Ziv-Welch Algorithm to Minimize Execution
Time,” 2019 International Conference on
Robotics,Electrical and Signal Processing Techniques
(ICREST), 2019, pp. 96-101
Fig. 20. Avalanche effects comparison [5]. M. Ali, Nada Abead, Suaad. (2016). Modified Blowfish
Algorithm for Image Encryption using Multi Keys based
on five Sboxes. 57. 2968-2978.
[6]. Godfrey L Dulla (2019). A unique message encryption
technique based on enhanced blowfish algorithm 2019
IOP Conf. Ser.: Mater. Sci. Eng. 482 012001
[7]. B. K. Maram and J. M. Gnanasekar, —Evaluation of Key
Dependent S-Box Based Data Security Algorithm using
Hamming Distance and Balanced Output,Journal, vol.5,
no. 1, pp. 67–75, 2016
[8]. U. C¸avus¸oglu,˘ A. Zengin, I. Pehlivan, and S. Kac¸ar, —
A novel approach for strong S-Box generation algorithm
design based on chaotic scaled Zhongtang system,||
Nonlinear Dyn., vol. 87, no. 2, pp. 1081–1094, 2017
[9]. M. Eberl, —The Fisher – Yates shuffle,|| pp. 1–9, 2018.
[10]. M. Ahmad, P. M. Khan, and M. Z. Ansari, —A Simple
and Efficient Key-Dependent S-Box Design Using Fisher-
Yates Shuffle Technique,|| Commun. Comput. Inf. Sci.,
Fig. 21. Analysis of Modified Algorithm vol. 420 CCIS, pp. 540–550, 2014

IJISRT22JUN1497 www.ijisrt.com 1764