International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248

Volume: 3 Issue: 11 548 – 552

_______________________________________________________________________________________________

VHDL Implementation of 128 bit Pipelined Blowfish Algorithm

Dakey Rahul Khanna1 N.Srikanth2 Dr. B.Subrahmaneswara Rao3

Student, M.Tech , VLSISD,Dept. of Associate Professor, Dept. of ECE, Professor, Dept. of ECE
ECE, Swarnandhra College of Engg. &Tech. Swarnandhra College of Engg. &Tech.
Swarnandhra College of Engg. &Tech. Narsapur , A.P,India. Narsapur , A.P,India.
Narsapur , A.P,India. e-mail: srikanth648@yahoo.co.in e-mail: boyina_a@yahoo.com
e-mail: rahulkhanna556@gmail.com

Abstract–– Communication through public networks imposes threat to our sensitive data. Information security plays an important role in public
networking and wireless communication. In order to achieve the protection of information or private data in networking, Cryptography can be
used. It is the automated method in which security goals are accomplished. Cryptographic algorithm is a mathematical function used in
Encryption and Decryption process. Blowfish is a keyed symmetric Cryptographic algorithm. It is a very fast and useful scheme for Encryption
and Decryption. A key is the strongest point of any algorithm but it can become the weakest point if it is not secured. Our information can be
secured if it is encrypted by using multiple keys. Hence implementation of Blowfish algorithm is very much use. Usually blowfish in existing
method is 64 bit block cipher and Throughput depends on the size of the blocks applied. In this proposed paper, the blowfish algorithm is
designed for 128 bit block size and pipelining operation is carried to improve the speed and reduce the delay accordingly, and so that this
architecture improves the throughput of an encoder. The implementation results indicate that the proposed pipelined architecture shows 10% of
improvement in Throughput. VHDL Implementation of proposed architecture has done by using XILINX ISE 9.1.

Key words: Cryptography, blowfish, pipelined architectures.

__________________________________________________*****________________________________________________

I. INTRODUCTION
Embedded devices are becoming more popular and that only intended user can use it and other devices are
used in network communications. These devices are now used prevented from using it without proper authentication. In
commonly for storing data and exchange of data through certain applications data security from tampering is mandatory
public networks. Wireless networking imposes threat to the by law. The applications like electronic engine controllers are
sensitive data. To protect sensitive data against threat required to be protected from tampering emission and
embedded devices are designed with inbuilt security features. performance data. Devices used to store medical data and
The sensitive data is encrypted before transmission so that utility meter like taxi, gas, electricity must be tamperproof.
only authorized user can have access to such information. Devices can be secured with passwords, identification
Hardware implementation of encryption algorithm is helpful in numbers, or tokens. The devices can be designed with better
designing secured Embedded System. security using cryptography.
Modern embedded systems need data security more
II. RELATED WORK
than ever before. It is now common to see portable electronic
devices of all sorts, ranging from cellular phones to military It was concluded in [8] that AES is faster and more
radio systems and from inventory tracking systems to medical efficient than other encryption algorithms. When the trans-
record tablets. Devices like PDAs and cell phones store mission of data is considered there is insignificant difference
personal e-mail and contact lists [1]. GPS receivers keep logs in performance of different symmetric key schemes (most of
of our movements and our automobiles record our driving the resources are consumed for data transmission rather than
habits. User-friendly products are in demand which can be computation). Even under the scenario of data transfer it
reprogrammed during normal use and add new features as would be advisable to use AES scheme in case the encrypted
firmware upgrades become available. The user expects that the data is stored at the other end and decrypted multiple times.
system should be protected from unauthorized access. A study in [9] is conducted for different popular
Embedded device manufacturers expect protection of secret key algorithms such as DES, 3DES, AES, and Blowfish.
their systems from unauthorized duplication or reverse They were implemented, and their performance was compared
engineering. Cryptography can be used to secure private data by encrypting input files of varying contents and sizes. The
and keep it private. Encrypted data transmission will protect algorithms were tested on two different hardware platforms, to
contact lists and personal e-mail from unauthorized or compare their performance. They had conducted it on two
unintended persons. Firmware upgrades can be encrypted so different machines: P-II 266 MHz and P-4 2.4 GHz. The
548
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 11 548 – 552
_______________________________________________________________________________________________
results showed that Blowfish had a very good performance IV. PROPOSED DESIGN
compared to other algorithms. Also it showed that AES had a There are software and hardware approaches to
better performance than 3DES and DES. It also shows that implement cryptographic Blowfish algorithm. Hardware
3DES has almost 1/3 throughput of DES, or in other words it implementation provides greater physical security and higher
needs 3 times than DES to process the same amount of data speed as compared to software implementation. Because of the
[10]. increasing requirement to implement cryptographic algorithms
A study in [11] provides evaluation of six of the most in fast rising high-speed network applications combined with
common encryption algorithms namely: AES, DES, 3DES, physical security, hardware implementation becomes essential.
RC2, Blowfish, and RC6. A comparison has been conducted
for those encryption algorithms at different settings for each
algorithm such as different sizes of data blocks, different data
types, battery power consumption, different key size and
finally encryption/decryption speed. Several points can be
concluded from the Experimental results. Firstly; there is no
significant difference when the results are displayed either in
hexadecimal base encoding or in base 64 encoding. Secondly;
in the case of changing packet size, it was concluded that
Blowfish has better performance than other common
encryption algorithms used, followed by RC6. Thirdly; we
find that 3DES still has low performance compared to Figure 1: Blowfish Algorithm Design Blocks
algorithm DES. Fourthly; we find RC2, has disadvantage over The encrypter is required to be designed as integrated
all other algorithms in terms of time consumption. Fifthly; we part of the system in order to attain superior performance.
find AES has better performance than RC2, DES, and 3DES. Embedded system designs are constraint by space, memory
In the case of audio and video files we found the result as the and time. Lookup tables are considered for S box and P array
same as in text and document. Finally, in the case of changing implementations which will generate keys fast and
key size, it can be seen that higher key size leads to clear precomputed values can be retained for encryption and
change in the battery and time consumption. decryption. The symmetric nature of Blowfish helps in
reutilizing hardware for encryption and decryption which
III. ALGORITHM further reduces area requirement or space. Blowfish is feistel
There are two parts to this algorithm; a part that handles the network. This simplifies encryption and decryption realization
expansion of the key and a part that handles the encryption of in hardware.
the data.The first step in the algorithm is to break the original Input plain text is applied to the core Blowfish cipher
key into a set of subkeys. Specifically, a key of no more than block. Input will be processed as block of 64 bits. Each 64 bit
448 bits is separated into 4168 bytes. There is a P-array and data further bifurcated into two 32bit data, in order to apply to
four 32-bit S-boxes. The P-array contains 18 32-bit subkeys, the cipher block. While applying data to the feistel block
while each S-box contains 256 entries. Padding will be done for framing data as multiple of 8 bits. Π
The following steps are used to calculate the subkeys: ROM will give initial key values to S Box. 32bit are divided
into four groups of 8 bit each. It acts as address for selecting
 Initialize the P-array and S-boxes precomputed substitution values of S box. Permutation values
 XOR P-array with the key bits. For example, P1 XOR are computed using lookup table. Data obtained from four S-
(first 32 bits of key), P2 XOR (second 32 bits of key), boxes further underwent to Arithmetic operations to get 32bit
 Use the above method to encrypt the all-zero string random data. The sub key of length 32 bit obtained from P
 This new output is now P1 and P2 array elements which are XORed and given to F function.
 Encrypt the new P1 and P2 with the modified
subkeys.
 This new output is now P3 and P4
 Repeat 521 times in order to calculate new subkeys
for the P-array and the four S-boxes
Blowfish has a 64-bit block size and a key length of
anywhere from 32 bits to 448 bits. It is a 16-round Feistel
cipher and uses large key-dependent S-boxes. It is similar in
structure to CAST-128, which uses fixed S-boxes. Here is the
Figure 2: Proposed Memory-based Method
visual representation of this encryption algorithm.
549
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 11 548 – 552
_______________________________________________________________________________________________
The architecture of the proposed Blowfish consists of path delay will be of the order of 21.618ns. The building block
a 128-bit block size and 64bit key size, whereby it comprises comprised of logic for performing XOR operation between
two parallel blocks of 64-bit Blowfish algorithm that are key and data and feistel function using S-boxes.
simultaneously executed. This design technique enables the
throughput of the Blowfish algorithm to be maximized. The
parallel blocks share the same S-box that is used for the F
function. As the implementation of the Blowfish design is
targeted to reduce the core size and timing delay, The
implementation for the blow fish using S-boxes can be
replaced with the proposed method uses a read-only memory
(ROM) that contains 1024 × 32-bit input data of addr. The
addr represents the data of four 32-bit S-boxes. The 32-bit
output data are read from the ROM. The proposed method can
Figure 4: Black box of building block of the Design
also lessen the total of slices used by the Blowfish design. A
implementation.
slice contains a set number of look-up tables (LUTs), FFs, and
multiplexers. Thus, less logic resources are used to perform
logic, arithmetic, and ROM functions that can lead to a faster
encryption/decryption process.
This implementation of Hardware block can be
reused two times to perform 128bit operation by means of Figure 5: internal implementation of building block.
selector and de-selector blocks aided to the above design. That
is the proposed implementation of the Blowfish algorithm. The As the single block to perform 64bit operation requires
design can be further modified to a pipelined implementation 21.618ns, to perform 128bit operation by reusing the block
of blowfish algorithm for improved performance as shown in for two times requires 43.236ns. Hence pipelining can be
the figure. introduced by inserting registers between every two building
blocks and make sure that the time lag between two clock ticks
must be at least 43.236ns. Hence the proposed improved
throughput pipelined blowfish Algorithm can work at the
application of maximum clock frequency of 23.128MHz.

V. RESULTS AND COMPARISONS

64bit fixed block cipher implemented in standard
blowfish hardware architecture by make use of 4 S-boxes
whose entries are calculated from the key and are independent
of both data and key while executing. The RTL (Register
Transfer Level) diagram and Design summary of implemented
Blowfish Algorithm are as shown below. The hardware
description done using VHDL and the synthesis of the design
done for Vertex E FPGA.

Figure 6: Simulated Waveform Diagram of 64bit blowfish

Algorithm
Figure 3: Block Diagram of improved throughput pipelined
The black box of improved throughput
blowfish Algorithm
blowfish Algorithm is as shown below. It generates 128bit
If we observe the feistel block implementation, which
is the building block, in hardware using VHDL the critical
550
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 11 548 – 552
_______________________________________________________________________________________________
cipher output from 64bit design by time sharing using select Here by comparing result obtained from 64bit fixed
input. block size blowfish algorithms implemented using S-box
method and ROM for feistel network, we can conclude that the
ROM based design is optimised to produce less delay and
lesser equivalent gate count. The logic utilization for the both
implementations are almost same.

S-box and ROM Designs

350000
300000
250000
Figure 7: Black Box of area efficient blowfish Algorithm. 200000
The black box of implemented improved throughput 150000
100000
pipelined blowfish Algorithm is as shown below. Here the 50000
select input makes the 64bit design to operate twice in order 0
produce 128bi output. The clock signal drives the registers Logic Equvalent Delay(ps)
separating the combinational building block of blowfish utilization Gate Count
hardware architecture and triggers with approximate minimum
lapse of 50ns. The select signal trigger once between two S-box Design ROM Design
clock triggers. The clear signal clears all the registers
Figure 9: Comparison between 64bit blowfish algorithms
whenever asserted.
using S-box and ROM
Throughput is defined as the average rate of
successful message delivery over a communication channel.
Throughput is directed toward evaluating each architecture’s
characteristic and performance. Throughput is calculated as
Eq.1.

Latency is the encryption/decryption time that is calculated in

clock cycles. Latency should be as small as possible to achieve
Figure 8: Black Box of Improved throughput pipelined a power-saving system. Furthermore, a long battery life is
blowfish Algorithm. necessary, particularly for mobile devices.
The cipher result obtained from the plain data applied to the
implemented design for two inputs is as given below, showing Comparisions among 128bit
minimum devotion of 56 of128. implementations
Plaintext:
700000
00000000000000000000000000000000000000000000000000
600000
00000000000000000000000000000000000000000000000000 500000
0000000000000000000000000000 400000
300000
Ciphertext: 200000
00010011110000010111011100110110001001000001011000 100000
00101101010011000100111100000101110111001101100010 0
0100000101100000101101010011 Logic Equivalent Critical Area Delay
Plaintext: utilization Gate Count Delay(ps) Product(ADP)
in um.ps
11111111111111111111111111111111111111111111111111
11111111111111111111111111111111111111111111111111
basic parallel implementation
1111111111111111111111111111
Ciphertext: proposed area efficient implementation
01101111011000001101101001111110001110111011000001 Improved throughput pipelined
00101111111011011011110110000011011010011111100011
1011101100000100101111111011 Figure 10: Comparison among implemented 128bit blowfish
algorithm implementations.
551
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 11 548 – 552
_______________________________________________________________________________________________

REFERENCES
Throughput(Normalized) [1] Ravi, Srivaths, et al. "Security in embedded systems: Design
challenges." ACM Transactions on Embedded Computing
140 Systems (TECS) 3.3 (2004): 461-491.
120
[2] Arya, S. "An Implementation of Blowfish Algorithm Using
100
80 FPGA." International Journal of Engineering Research and
60 Technology. Vol. 2. No. 8 (August-2013). IJERT, 2013.
40 [3] Patel C.R., Gohil N.B., Shah V.“FPGA - Hardware Based
20 DES & Blowfish Symmetric Cipher Algorithms For
0 Encryption & Decryption Of Secured Wireless Data
Communication”, Journal of Information, Knowledge and
basic parallel proposed area Improved Research in Electronics and Communication Engineering,
2(2), 739–744. (2013).
implementatio efficient throughput
[4] Dakate, Deepak Kumar, and Pawan Dubey. "Blowfish
implementation pipelined encryption: A comparative analysis using VHDL."
International of Engineering and Advanced Technology
Figure 11: Throughput Comparison among implemented (IJEAT) 1.5 (2012): 177-179.
128bit blowfish algorithms [5] W. Stalling, “Cryptography and Network Security Principles
and Practices”, Prentice Hall, 4th ed., 2005.
From the comparison results shown above, the [6] A. Kahate, “Cryptography and Network Security”, Tata
proposed improved throughput method of implementation McGraw Hill, 2nd ed., 2007.
[7] National Institute of Standards and Technology, Federal
shows off less requirement of LUTs (Look up Tables),
Information Processing Standards Publication 46-3: Data
equivalent gates when compared to parallel implementation
Encryption Standard, 1999.
method. The evaluation delay may be little higher due to [8] S. Hirani, Energy Consumption of Encryption Schemes in
inclusion of selector and de-selector, but overall performance Wireless Devices Thesis, University ofPittsburgh, Apr.
metric ADP is lower for proposed design which is the index 9,2003, Retrieved Oct. 1, 2008.
for improved performance. The proposed design is further [9] A. Nadeem, \A performance comparison of data encryption
modified to get improved throughput pipelined blowfish algorithms," IEEE Information and Communication
architecture, which shows off extremely optimised Technologies, pp. 84-89, 2006.
performance. Results in less logic utilisation, less delay and [10] Results of Comparing Tens of Encryption Algorithms Using
Di®erent Settings- Crypto++ Benchmark,RetrievedOct. 1,
lesser Area Delay Product for superior performance. Still it
2008. (http://www.eskimo.com/ weidai/benchmarks.html)
seems to occupy more equivalent gate count than the proposed
[11] DiaaSalamaAbdElminaam, et,al. “Evaluating The
design due to inclusion of Registers. Performance of SymmetricEncryption Algorithms”
The performance metric throughput is seems to be International Journal of Network Security, Vol.10, No.3,
doubled for proposed implementation, due to parallel PP.213{219, May 2010.
implementation of hardware, whereas for pipelined
architecture throughput is n times, where n is the number of
pipeline stages. Overall performance of the proposed
improved throughput pipelined architecture is better compared
to others.

VI. CONCLUSION
Implemented 64bit fixed block size blowfish algorithm using
VHDL. Simulated for satisfactory result in MODELSIM and
synthesized in XILINX ISE simulator, also implemented
feistel network of blowfish algorithm with ROM and observed
the performance improvement. The implementation of 128bit
blowfish algorithm using proposed improved throughput
method, proposed area efficient method and proposed
improved throughput pipelined method is done and obtained
results shows that the proposed improved throughput pipelined
method poses superior performance. The implementation
results indicate that the proposed pipelined architecture shows
10% of improvement in Throughput.
552
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________