SlideShare a Scribd company logo
1 of 42
Download to read offline
INVESTIGATION OF ERROR
PROPAGATION EFFECT OF AES AND TO
FIND TECHNIQUES AND SOLUTION OF
THE PROBLEM
Dr. Bikramjit sarkar
Associate Professor
Dept. of Computer Science and Engineering
Techno India – Salt Lake
Kolkata, India
Email: sarkar.Bikramjit@gmail.com
Abstract
The Rijndael algorithm (Advanced Encryption Standard)
suffers from a major limitation owing to error
propagation in the encryption process. The salient
objective of our research is
To study different schemes for modified redundancy
based AES scheme,
To analyze the proposed schemes in terms of speed of
encryption /decryption and the level of security offered,
To finally design an error-free AES.
Advance Encryption Standard
Due to the several reports of failure of security or key
of DES (Data Encryption Standard), The NIST decided
to launch in 1997 a new standardization process, known
as Advanced Encryption Standard (AES).
The algorithm “Rijndael” was finally selected to
become the AES through an open contest where anyone
including even non-American citizen and companies
was invited to submit a candidate algorithm and to send
public comments on the other proposals.
The algorithm was named after its designers Daemen
and Rijmen.
Advance Encryption Standard contd..
The replacement has aimed to augment the level of
security mainly with higher key size. Besides the higher
level of security, AES has aimed to provide higher
efficiency and better flexibility by means of encryption
at different levels and with different block sizes.
The AES encryption is done at several rounds of
iteration. Each round of iteration has different input
data and different keys. The input data and the keys of
different rounds are all generated from the original
source data and the source key respectively. On the
basis of this theory the input data and the keys at
rounds follow a data path and key path respectively.
Advance Encryption Standard contd..
Error Propagation Effect of AES
AES is designed for flexibility and for implementing in
both Hardware and software. Due to its low memory
requirement, AES suitable for implementation in
devices like smart card etc. But the AES algorithm
suffers from a major limitation Error Propagation
Effect.
An error, if inflicted in any round, propagates through
the rest of the rounds and finally results in a large
number of errors in the cipher.
Error Propagation Effect of AES contd..
The effect of the Error Propagation has been observed
through an experiment carried out with the following
plain text (128 bits) and the cipher key (128 bits):
Plain text Cipher key
Error Propagation Effect of AES contd..
Now a one-bit error is forcibly injected in 8th bit
position of the intermediate cipher generated after 1st
round and then encrypt it through 9 more rounds of
AES encryption. As a consequence, an erroneous
cipher is generated at the output as follows:
Erroneous Cipher text Error free Cipher text
Error Propagation Effect of AES contd..
In this way, the experiment has been conducted on
deliberately injecting one-bit errors after different
rounds and consequently a number of erroneous ciphers
were generated. Upon decryption of these erroneous
ciphers, corresponding plaintexts have been received,
which, to a great extent, differ from the actual plain
text.
The result found in the above experiment is shown in
Table 1.
Error Propagation Effect of AES contd..
One-bit Errors injected in
8th bit positions of the
intermediate states
Number of errors in the cipher text
after injection of one bit error at
intermediate states
Number of corresponding errors in
the plain text after decrypting the
erroneous cipher
After 1st round 71 67
After 2nd round 67 55
After 3rd round 56 69
After 4th round 51 61
After 5th round 69 73
After 6th round 62 65
After 7th round 70 57
After 8th round 53 70
After 9th round 62 61
Table 1
Available Techniques to fight the Error
Propagation Effect of AES
the limitation of error propagation in AES leads to low
speed of encryption, more processing overhead and
higher complexity, because until and unless error free
encryption be achieved the transmission of the cipher
will be meaningless.
In order to tackle the error propagation of AES, two
techniques have been studied in literatures:
Redundancy Based Technique
Byte Based Parity Technique
Redundancy Based Technique
The Redundancy Based Technique needs both the
encryption module and decryption module for
producing error-free cipher at the transmitter. The
output cipher of the encryption module is decrypted by
the decryption module. The decrypted output is
compared with the plain text to check whether there is
any error at all. If they match, the cipher is considered
to be error-free. The dual process of encryption and
decryption by the technique make the encryption
process slow and costly as well. Although this
technique is suitable for both software and hardware
implementation.
Byte Based Parity Technique
The byte based parity technique makes use of parity
checking at each byte of plain text to combat error. Yet
it has the limitation of not being able to correct all type
of error vectors. Also the byte based parity technique is
not suitable for software implementation.
Selective AES
The performance of any cryptosystem is measured by
two parameters: level of security it provides and speed
of encryption and decryption. Selective AES is the
process of AES encryption of a fraction of message
keeping the remaining portions unencrypted, which
increases in the speed of encryption / decryption.
Error Propagation in Selective AES
The Error Propagation Effect has been observed in
Selective AES through an experiment conducted with
the following 640-bit message and 128-bit key:
Message = “ERROR PROPAGATION EFFECT OF
AES HAS THROWN A GREAT RESEARCH
CHALLENGE BEFORE US.”
Key = “BIKRAMJIT SARKAR”
Error Propagation in Selective AES contd..
First the entire message is divided into 5 blocks B1, B2,
B3, B4 and B5, each of which is of 128 bits, where,
B1 = “ERROR PROPAGATIO”,
B2 = “N EFFECT OF AES ”,
B3 = “HAS THROWN A GRE”,
B4 = “AT RESEARCH CHAL” and
B5 = “LENGE BEFORE US.”
Error Propagation in Selective AES contd..
The AES Encryption process has 10 rounds. And after
each round (1st to 9th) one intermediate state is found
and after the 10 rounds the final cipher text is
generated. So, each block, while being encrypted to
generate the final cipher, generates 9 intermediate
states.
Now after first round, the eighth bit of each
intermediate state generated from each block get
forcibly infected. The same thing is repeated for each of
the blocks after fifth and ninth rounds.
Error Propagation in Selective AES contd..
Now the erroneous ciphers are compared with the error
free cipher and accordingly the total number of errors at
the output ciphers for each of the cases is found. The
average number of errors occurred in the output cipher
when a single bit error has occurred in any of the
intermediate states generated during the encryption
process is also obtained and Table 2 is formed:
Block(s) Encrypted Percentage of Selection
Average number of errors occurred at the output after the
execution of the Encryption Module
B1 20 45.67
B1 & B2 40 88.67
B1, B2 & B3 60 133.67
B1, B2, B3 & B4 80 176.0
B1, B2, B3, B4 & B5 100 221.33
Table 2
Error Propagation in Selective AES contd..
From Table 2 it is clear that lesser is the percentage of
selection, lesser is the number of errors occurred at the
output. But the processing speed is inversely
proportional to the percentage of selection. So there
must be a tread off between the security level and the
processing speed since, the level of security is
proportional to the percentage of selection.
Redundancy Based Technique redefined
It is clear that the Redundancy Based Technique
requires the comparison of 128 bits since the plain text
taken is of 128 bits. Here the proposed algorithm SBM
1.1 proposes to modify the Redundancy Based
Technique that will reduce the overhead of comparison
to only 8 bits. The algorithm makes use of the
Longitudinal Redundancy Check (LRC) code.
Redundancy Based Technique redefined contd..
Proposed Algorithm: SBM 1.1
1. Input the plain text P of 128 bits
2. Generate an LRC code (8 bits), say L, out of P
3. Encrypt P with AES Encryptor to find the cipher
text, say C
4. Decrypt C with AES Decryptor to find P’
5. Generate an LRC code (8 bits), say L’, out of P’
6. L and L’ are now compared. If L and L’ are found to
be same, C is considered to be error-free and is
transmitted through the channel.
Redundancy Based Technique redefined contd..
Block Diagram: SBM 1.1
Redundancy Based Technique redefined contd..
Experimental Results:
Input Message P (128 bits):
42 69 6b 72 61 6d 6a 69 74 20 53 61 72 6b 61 72
LRC code generated from P (Say, L): 0 1 0 1 0 0 0 1
P is encrypted with the following key:
2B 28 AB 09 7E AE F7 CF 15 D2 15 4F 16 A6 88 3C
Redundancy Based Technique redefined contd..
Experimental Results:
A one-bit error is injected in the 8th bit position of the
intermediate cipher generated after 7th round and the
encryption process continues through 3 more rounds of
AES encryption. As a consequence, an erroneous
cipher text is generated at the output as follows:
8A 88 3E 2D DC 16 77 90 4D B3 05 3E CA 04 4D 0C
P’ generated from the above erroneous cipher:
B2 C9 A7 34 CF 60 C6 24 75 F5 4B CD 9F 97 3C 62
LRC code generated from P’(Say, L’): 1 1 1 1 0 1 0 1
Redundancy Based Technique redefined contd..
Experimental Results:
L and L’ are finally compared and it is observed that
they are not same. This indicates that some error has
occurred and the cipher text generated and hence it is
useless to transmit.
The proposed approach has minimized the overhead of
comparison from 128 pairs of bits to 8 pairs of bits.
Although an additional module of LRC generator that
has been introduced in our approach causes an extra
overhead of the new approach, yet the proposed
technique is superior.
A probabilistic approach to tackle Error
Propagation Effect of AES
The proposed scheme suggests that the plain text will
be encrypted odd number of times, say n, with the same
key to produce n number of cipher texts. It is assumed
that the probability of occurrence of a single bit error
amidst the rounds must not reach 0.5 so that out of n
cipher texts at most (n – 1)/2 number of cipher texts
may be erroneous whereas the least number of error
free cipher texts is (n + 1)/2. Majority Rule is then
applied over the n number of ciphers and as a result the
error free cipher text is achieved.
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Proposed Algorithm: SBM 1.2
1. Input the plain text P of 128 bits
2. Encrypt P n number of times with AES Encryptor to
find ciphers {Ci, i = 1 to n}, n being odd
3. Majority rule is applied over Ci to find the error-
free cipher text C which is directly transmitted
through the channel.
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Block diagram: SBM 1.2
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Experimental results:
Input Message P (128 bits):
42 69 6b 72 61 6d 6a 69 74 20 53 61 72 6b 61 72
Key K (128 bits):
2B 28 AB 09 7E AE F7 CF 15 D2 15 4F 16 A6 88 3C
Error-free Cipher:
FC 41 16 48 BE C0 16 A7 FC 5C 3F 43 F4 13 F4 A0
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Experimental results:
The Plain text P has been encrypted by the key 5
numbers of time and as per the assumption at most 2
numbers of ciphers texts may be erroneous. The
erroneous cipher texts are as follows:
Erroneous cipher text 1:
D2 06 F1 26 BE 27 E9 EE FA CF AF 6B 55 25 D6 0D
Erroneous cipher text 2:
24 E2 C9 55 1C 30 E0 58 5D D7 3C 64 EA 2F CA F7
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Experimental results:
As per the assumption, the rest of the cipher texts (3
numbers) are error free.
The Majority Rule is then applied over the 5 cipher
texts and finally the error-free output cipher is
generated as follows:
FC 41 16 48 BE C0 16 A7 FC 5C 3F 43 F4 13 F4 A0
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Experimental results:
As per the assumption, the rest of the cipher texts (3
numbers) are error free.
The Majority Rule is then applied over the 5 cipher
texts and finally the error-free output cipher is
generated as follows:
FC 41 16 48 BE C0 16 A7 FC 5C 3F 43 F4 13 F4 A0
A probabilistic approach to tackle Error
Propagation Effect of AES contd..
Although in the experiment the plain text has been
encrypted 5 times, It can be realized that with the
increase in the number (n) of cipher texts, the scheme
guarantees reliable communication on its part, even
with the occurrence of error with higher probability
(0.5 * (1 – 1/n)); that means that if n increases the
tolerance also increases.
A Hybrid Architecture to overcome the Error
Propagation Effect of AES
The algorithm SBM 1.3 proposes a hybrid architecture
to ward off the Error Propagation Effect of AES. The
proposed scheme is the combination of the algorithms
SBM 1.1 and SBM 1.2, that makes use of both the
Longitudinal Redundancy Check code and the Majority
Rule in parallel. The algorithm checks for the effect of
error propagation in 2 different paths and finally the
error-free cipher text is obtained ensuring a reliable
communication.
A Hybrid Architecture to overcome the Error
Propagation Effect of AES contd..
Proposed algorithm: SBM 1.3
1. Input the plain text P of 128 bits
2. Generate an LRC code (8 bits), say L1, out of P
3. Encrypt P with AES Encryptor to find the cipher
text, say C’
4. Decrypt C with AES Decryptor to find P’
5. Generate an LRC code (8 bits), say L2, out of P’
6. L1 and L2 are now compared. If L1 and L2 are found
to be same, C’ is fed to the comparator.
A Hybrid Architecture to overcome the Error
Propagation Effect of AES contd..
Proposed algorithm: SBM 1.3
7. Encrypt P n number of times with AES Encryptor to
find ciphers {Ci, i = 1 to n}, n being odd
8. Majority rule is applied over Ci to find the cipher
text C” which is also fed to the comparator.
9. If C’ and C” are found to be same, say C, it is
considered to be error-free and is transmitted
through the channel.
A Hybrid Architecture to overcome the Error
Propagation Effect of AES contd..
Block diagram: SBM 1.3
A Hybrid Architecture to overcome the Error
Propagation Effect in Selective AES
Proposed algorithm: SBM 1.4
Input Key Words for the message of N words. Message
is divided into K parts each of k blocks. Each block is
of M words.
Find the occurrence of any keyword in the blocks,
starting with the first block in the first part. If it occurs,
encrypt that block using the algorithm SBM 1.3 and all
the blocks thereafter in the part.
Repeat (1-2) for all parts, j=1 to K. When j=K, the
proposed scheme of encryption is complete.
A Hybrid Architecture to overcome the Error
Propagation Effect in Selective AES
Block diagram: SBM 1.4
Overall conclusion
In this thesis we have studied the effect in detail. The
relative mathematical analysis has been cited in the
thesis. Error propagation effect in case of selective AES
and its comparison with normal AES has also been
studied. In this thesis four algorithms, viz. SBM 1.1,
SBM 1.2, SBM 1.3 and SBM 1.4, have been proposed
for preventing Error Propagation Effect of AES. The
corresponding experimental results have also been
provided so as to prove the algorithms to be efficient to
overcome the error propagation effect of the Rijndael
used as Advanced Encryption Standard.
Future scope
It is assumed that the probability of occurrence of a
single bit error amidst the rounds is maximum 50%.
Further research works may be extended to achieve
certainty in preventing the Error Propagation Effect of
AES algorithm under all circumstances.
When the number of rounds is increased in AES, the
complexity of AES encryption and decryption also
increases. So, the length of the key may increased to
512 bits in order to increase the number of rounds.
Future research may include the above mentioned
considerations.
Error Propagation Effect of AES and Techniques to ward off the problem

More Related Content

What's hot

AUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEM
AUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEMAUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEM
AUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEMAnamika Vinod
 
An introduction to deep reinforcement learning
An introduction to deep reinforcement learningAn introduction to deep reinforcement learning
An introduction to deep reinforcement learningBig Data Colombia
 
Kohonen self organizing maps
Kohonen self organizing mapsKohonen self organizing maps
Kohonen self organizing mapsraphaelkiminya
 
Support vector machines (svm)
Support vector machines (svm)Support vector machines (svm)
Support vector machines (svm)Sharayu Patil
 
K MEANS CLUSTERING
K MEANS CLUSTERINGK MEANS CLUSTERING
K MEANS CLUSTERINGsingh7599
 
Artificial Neural Networks Lect5: Multi-Layer Perceptron & Backpropagation
Artificial Neural Networks Lect5: Multi-Layer Perceptron & BackpropagationArtificial Neural Networks Lect5: Multi-Layer Perceptron & Backpropagation
Artificial Neural Networks Lect5: Multi-Layer Perceptron & BackpropagationMohammed Bennamoun
 
A thesis presentation on pothole detection
A thesis presentation on pothole detectionA thesis presentation on pothole detection
A thesis presentation on pothole detectionPrimeAsia University
 
. An introduction to machine learning and probabilistic ...
. An introduction to machine learning and probabilistic .... An introduction to machine learning and probabilistic ...
. An introduction to machine learning and probabilistic ...butest
 
MACHINE LEARNING - GENETIC ALGORITHM
MACHINE LEARNING - GENETIC ALGORITHMMACHINE LEARNING - GENETIC ALGORITHM
MACHINE LEARNING - GENETIC ALGORITHMPuneet Kulyana
 
Brain Tumor Classification using Support Vector Machine
Brain Tumor Classification using Support Vector MachineBrain Tumor Classification using Support Vector Machine
Brain Tumor Classification using Support Vector MachineIRJET Journal
 
Reinforcement Learning Q-Learning
Reinforcement Learning   Q-Learning Reinforcement Learning   Q-Learning
Reinforcement Learning Q-Learning Melaku Eneayehu
 
Deep Learning’s Application in Radar Signal Data II
Deep Learning’s Application in Radar Signal Data IIDeep Learning’s Application in Radar Signal Data II
Deep Learning’s Application in Radar Signal Data IIYu Huang
 
Threat Detection in Surveillance Videos
Threat Detection in Surveillance VideosThreat Detection in Surveillance Videos
Threat Detection in Surveillance VideosDatabricks
 
Credit Card Fraud Detection Using Unsupervised Machine Learning Algorithms
Credit Card Fraud Detection Using Unsupervised Machine Learning AlgorithmsCredit Card Fraud Detection Using Unsupervised Machine Learning Algorithms
Credit Card Fraud Detection Using Unsupervised Machine Learning AlgorithmsHariteja Bodepudi
 

What's hot (20)

AUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEM
AUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEMAUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEM
AUTOMATIC ACCIDENT DETECTION AND ALERT SYSTEM
 
Implements the histogram equalization algorithm
Implements the histogram equalization algorithmImplements the histogram equalization algorithm
Implements the histogram equalization algorithm
 
Flight Delay Prediction
Flight Delay PredictionFlight Delay Prediction
Flight Delay Prediction
 
04 Multi-layer Feedforward Networks
04 Multi-layer Feedforward Networks04 Multi-layer Feedforward Networks
04 Multi-layer Feedforward Networks
 
An introduction to deep reinforcement learning
An introduction to deep reinforcement learningAn introduction to deep reinforcement learning
An introduction to deep reinforcement learning
 
PROJECT FINAL PPT
PROJECT FINAL PPTPROJECT FINAL PPT
PROJECT FINAL PPT
 
Kohonen self organizing maps
Kohonen self organizing mapsKohonen self organizing maps
Kohonen self organizing maps
 
Support vector machines (svm)
Support vector machines (svm)Support vector machines (svm)
Support vector machines (svm)
 
K MEANS CLUSTERING
K MEANS CLUSTERINGK MEANS CLUSTERING
K MEANS CLUSTERING
 
Artificial Neural Networks Lect5: Multi-Layer Perceptron & Backpropagation
Artificial Neural Networks Lect5: Multi-Layer Perceptron & BackpropagationArtificial Neural Networks Lect5: Multi-Layer Perceptron & Backpropagation
Artificial Neural Networks Lect5: Multi-Layer Perceptron & Backpropagation
 
Ensemble methods
Ensemble methods Ensemble methods
Ensemble methods
 
A thesis presentation on pothole detection
A thesis presentation on pothole detectionA thesis presentation on pothole detection
A thesis presentation on pothole detection
 
. An introduction to machine learning and probabilistic ...
. An introduction to machine learning and probabilistic .... An introduction to machine learning and probabilistic ...
. An introduction to machine learning and probabilistic ...
 
Image segmentation
Image segmentation Image segmentation
Image segmentation
 
MACHINE LEARNING - GENETIC ALGORITHM
MACHINE LEARNING - GENETIC ALGORITHMMACHINE LEARNING - GENETIC ALGORITHM
MACHINE LEARNING - GENETIC ALGORITHM
 
Brain Tumor Classification using Support Vector Machine
Brain Tumor Classification using Support Vector MachineBrain Tumor Classification using Support Vector Machine
Brain Tumor Classification using Support Vector Machine
 
Reinforcement Learning Q-Learning
Reinforcement Learning   Q-Learning Reinforcement Learning   Q-Learning
Reinforcement Learning Q-Learning
 
Deep Learning’s Application in Radar Signal Data II
Deep Learning’s Application in Radar Signal Data IIDeep Learning’s Application in Radar Signal Data II
Deep Learning’s Application in Radar Signal Data II
 
Threat Detection in Surveillance Videos
Threat Detection in Surveillance VideosThreat Detection in Surveillance Videos
Threat Detection in Surveillance Videos
 
Credit Card Fraud Detection Using Unsupervised Machine Learning Algorithms
Credit Card Fraud Detection Using Unsupervised Machine Learning AlgorithmsCredit Card Fraud Detection Using Unsupervised Machine Learning Algorithms
Credit Card Fraud Detection Using Unsupervised Machine Learning Algorithms
 

Viewers also liked

The wire
The wireThe wire
The wiresdpable
 
Ionosphere and its effect in Aerial Propagation
Ionosphere and its effect in Aerial PropagationIonosphere and its effect in Aerial Propagation
Ionosphere and its effect in Aerial PropagationArman Sharif
 
Propagation Effects for Radar&Comm Systems
Propagation Effects for Radar&Comm SystemsPropagation Effects for Radar&Comm Systems
Propagation Effects for Radar&Comm SystemsJim Jenkins
 
DEF CON 23 - NSM 101 for ICS
DEF CON 23 - NSM 101 for ICSDEF CON 23 - NSM 101 for ICS
DEF CON 23 - NSM 101 for ICSChris Sistrunk
 
Space wave propagation ppt
Space wave propagation pptSpace wave propagation ppt
Space wave propagation pptdhanesh1994
 
Error analysis presentation
Error analysis presentationError analysis presentation
Error analysis presentationGeraldine Lopez
 

Viewers also liked (7)

The wire
The wireThe wire
The wire
 
Propagation effects: Reverb
Propagation effects: ReverbPropagation effects: Reverb
Propagation effects: Reverb
 
Ionosphere and its effect in Aerial Propagation
Ionosphere and its effect in Aerial PropagationIonosphere and its effect in Aerial Propagation
Ionosphere and its effect in Aerial Propagation
 
Propagation Effects for Radar&Comm Systems
Propagation Effects for Radar&Comm SystemsPropagation Effects for Radar&Comm Systems
Propagation Effects for Radar&Comm Systems
 
DEF CON 23 - NSM 101 for ICS
DEF CON 23 - NSM 101 for ICSDEF CON 23 - NSM 101 for ICS
DEF CON 23 - NSM 101 for ICS
 
Space wave propagation ppt
Space wave propagation pptSpace wave propagation ppt
Space wave propagation ppt
 
Error analysis presentation
Error analysis presentationError analysis presentation
Error analysis presentation
 

Similar to Error Propagation Effect of AES and Techniques to ward off the problem

Aes 128 192_256_bits_project_report
Aes 128 192_256_bits_project_reportAes 128 192_256_bits_project_report
Aes 128 192_256_bits_project_reportsakhi rehman
 
A Probabilistic Approach towards the Prevention of Error Propagation Effect o...
A Probabilistic Approach towards the Prevention of Error Propagation Effect o...A Probabilistic Approach towards the Prevention of Error Propagation Effect o...
A Probabilistic Approach towards the Prevention of Error Propagation Effect o...IDES Editor
 
Renas Rajab Asaad
Renas Rajab Asaad Renas Rajab Asaad
Renas Rajab Asaad Renas Rekany
 
IMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSOR
IMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSORIMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSOR
IMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSORacijjournal
 
Enhanced Advanced Encryption Standard (E-AES): using ESET
Enhanced Advanced Encryption Standard (E-AES): using ESETEnhanced Advanced Encryption Standard (E-AES): using ESET
Enhanced Advanced Encryption Standard (E-AES): using ESETIRJET Journal
 
High throughput FPGA Implementation of Advanced Encryption Standard Algorithm
High throughput FPGA Implementation of Advanced Encryption Standard AlgorithmHigh throughput FPGA Implementation of Advanced Encryption Standard Algorithm
High throughput FPGA Implementation of Advanced Encryption Standard AlgorithmTELKOMNIKA JOURNAL
 
Iaetsd implementation of lsb image steganography system using edge detection
Iaetsd implementation of lsb image steganography system using edge detectionIaetsd implementation of lsb image steganography system using edge detection
Iaetsd implementation of lsb image steganography system using edge detectionIaetsd Iaetsd
 
Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...
Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...
Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...Bikash Chandra Prusty
 
Design and Implementation of Area Efficiency AES Algoritham with FPGA and ASIC
Design and Implementation of Area Efficiency AES Algoritham with FPGA and ASICDesign and Implementation of Area Efficiency AES Algoritham with FPGA and ASIC
Design and Implementation of Area Efficiency AES Algoritham with FPGA and ASICpaperpublications3
 
Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...
Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...
Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...IJMTST Journal
 
Advanced Encryption Standard (AES) Implementaion using Java
Advanced Encryption Standard (AES) Implementaion using JavaAdvanced Encryption Standard (AES) Implementaion using Java
Advanced Encryption Standard (AES) Implementaion using JavaSunil Kumar R
 
An Efficient VLSI Architecture for AES and It's FPGA Implementation
An Efficient VLSI Architecture for AES and It's FPGA ImplementationAn Efficient VLSI Architecture for AES and It's FPGA Implementation
An Efficient VLSI Architecture for AES and It's FPGA ImplementationIRJET Journal
 
Module 2 network and computer security
Module 2 network and computer securityModule 2 network and computer security
Module 2 network and computer securityDeepak John
 

Similar to Error Propagation Effect of AES and Techniques to ward off the problem (20)

Aes 128 192_256_bits_project_report
Aes 128 192_256_bits_project_reportAes 128 192_256_bits_project_report
Aes 128 192_256_bits_project_report
 
A Probabilistic Approach towards the Prevention of Error Propagation Effect o...
A Probabilistic Approach towards the Prevention of Error Propagation Effect o...A Probabilistic Approach towards the Prevention of Error Propagation Effect o...
A Probabilistic Approach towards the Prevention of Error Propagation Effect o...
 
Renas Rajab Asaad
Renas Rajab Asaad Renas Rajab Asaad
Renas Rajab Asaad
 
Fault Detection AES
Fault Detection AESFault Detection AES
Fault Detection AES
 
IMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSOR
IMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSORIMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSOR
IMPLEMENTATION OF AES AS A CUSTOM HARDWARE USING NIOS II PROCESSOR
 
Enhanced Advanced Encryption Standard (E-AES): using ESET
Enhanced Advanced Encryption Standard (E-AES): using ESETEnhanced Advanced Encryption Standard (E-AES): using ESET
Enhanced Advanced Encryption Standard (E-AES): using ESET
 
Implementation of Fast Pipelined AES Algorithm on Xilinx FPGA
Implementation of Fast Pipelined AES Algorithm on Xilinx FPGAImplementation of Fast Pipelined AES Algorithm on Xilinx FPGA
Implementation of Fast Pipelined AES Algorithm on Xilinx FPGA
 
High throughput FPGA Implementation of Advanced Encryption Standard Algorithm
High throughput FPGA Implementation of Advanced Encryption Standard AlgorithmHigh throughput FPGA Implementation of Advanced Encryption Standard Algorithm
High throughput FPGA Implementation of Advanced Encryption Standard Algorithm
 
J0445255
J0445255J0445255
J0445255
 
A03530107
A03530107A03530107
A03530107
 
Iaetsd implementation of lsb image steganography system using edge detection
Iaetsd implementation of lsb image steganography system using edge detectionIaetsd implementation of lsb image steganography system using edge detection
Iaetsd implementation of lsb image steganography system using edge detection
 
Aes
AesAes
Aes
 
Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...
Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...
Dual Layer Security Of Data Using LSB Image Steganography And AES Encryption ...
 
A04660105
A04660105A04660105
A04660105
 
Design and Implementation of Area Efficiency AES Algoritham with FPGA and ASIC
Design and Implementation of Area Efficiency AES Algoritham with FPGA and ASICDesign and Implementation of Area Efficiency AES Algoritham with FPGA and ASIC
Design and Implementation of Area Efficiency AES Algoritham with FPGA and ASIC
 
Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...
Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...
Novel Adaptive Hold Logic Circuit for the Multiplier using Add Round Key and ...
 
Advanced Encryption Standard (AES) Implementaion using Java
Advanced Encryption Standard (AES) Implementaion using JavaAdvanced Encryption Standard (AES) Implementaion using Java
Advanced Encryption Standard (AES) Implementaion using Java
 
An Efficient VLSI Architecture for AES and It's FPGA Implementation
An Efficient VLSI Architecture for AES and It's FPGA ImplementationAn Efficient VLSI Architecture for AES and It's FPGA Implementation
An Efficient VLSI Architecture for AES and It's FPGA Implementation
 
D0412032038
D0412032038D0412032038
D0412032038
 
Module 2 network and computer security
Module 2 network and computer securityModule 2 network and computer security
Module 2 network and computer security
 

Error Propagation Effect of AES and Techniques to ward off the problem

  • 1. INVESTIGATION OF ERROR PROPAGATION EFFECT OF AES AND TO FIND TECHNIQUES AND SOLUTION OF THE PROBLEM Dr. Bikramjit sarkar Associate Professor Dept. of Computer Science and Engineering Techno India – Salt Lake Kolkata, India Email: sarkar.Bikramjit@gmail.com
  • 2. Abstract The Rijndael algorithm (Advanced Encryption Standard) suffers from a major limitation owing to error propagation in the encryption process. The salient objective of our research is To study different schemes for modified redundancy based AES scheme, To analyze the proposed schemes in terms of speed of encryption /decryption and the level of security offered, To finally design an error-free AES.
  • 3. Advance Encryption Standard Due to the several reports of failure of security or key of DES (Data Encryption Standard), The NIST decided to launch in 1997 a new standardization process, known as Advanced Encryption Standard (AES). The algorithm “Rijndael” was finally selected to become the AES through an open contest where anyone including even non-American citizen and companies was invited to submit a candidate algorithm and to send public comments on the other proposals. The algorithm was named after its designers Daemen and Rijmen.
  • 4. Advance Encryption Standard contd.. The replacement has aimed to augment the level of security mainly with higher key size. Besides the higher level of security, AES has aimed to provide higher efficiency and better flexibility by means of encryption at different levels and with different block sizes. The AES encryption is done at several rounds of iteration. Each round of iteration has different input data and different keys. The input data and the keys of different rounds are all generated from the original source data and the source key respectively. On the basis of this theory the input data and the keys at rounds follow a data path and key path respectively.
  • 6. Error Propagation Effect of AES AES is designed for flexibility and for implementing in both Hardware and software. Due to its low memory requirement, AES suitable for implementation in devices like smart card etc. But the AES algorithm suffers from a major limitation Error Propagation Effect. An error, if inflicted in any round, propagates through the rest of the rounds and finally results in a large number of errors in the cipher.
  • 7. Error Propagation Effect of AES contd.. The effect of the Error Propagation has been observed through an experiment carried out with the following plain text (128 bits) and the cipher key (128 bits): Plain text Cipher key
  • 8. Error Propagation Effect of AES contd.. Now a one-bit error is forcibly injected in 8th bit position of the intermediate cipher generated after 1st round and then encrypt it through 9 more rounds of AES encryption. As a consequence, an erroneous cipher is generated at the output as follows: Erroneous Cipher text Error free Cipher text
  • 9. Error Propagation Effect of AES contd.. In this way, the experiment has been conducted on deliberately injecting one-bit errors after different rounds and consequently a number of erroneous ciphers were generated. Upon decryption of these erroneous ciphers, corresponding plaintexts have been received, which, to a great extent, differ from the actual plain text. The result found in the above experiment is shown in Table 1.
  • 10. Error Propagation Effect of AES contd.. One-bit Errors injected in 8th bit positions of the intermediate states Number of errors in the cipher text after injection of one bit error at intermediate states Number of corresponding errors in the plain text after decrypting the erroneous cipher After 1st round 71 67 After 2nd round 67 55 After 3rd round 56 69 After 4th round 51 61 After 5th round 69 73 After 6th round 62 65 After 7th round 70 57 After 8th round 53 70 After 9th round 62 61 Table 1
  • 11. Available Techniques to fight the Error Propagation Effect of AES the limitation of error propagation in AES leads to low speed of encryption, more processing overhead and higher complexity, because until and unless error free encryption be achieved the transmission of the cipher will be meaningless. In order to tackle the error propagation of AES, two techniques have been studied in literatures: Redundancy Based Technique Byte Based Parity Technique
  • 12. Redundancy Based Technique The Redundancy Based Technique needs both the encryption module and decryption module for producing error-free cipher at the transmitter. The output cipher of the encryption module is decrypted by the decryption module. The decrypted output is compared with the plain text to check whether there is any error at all. If they match, the cipher is considered to be error-free. The dual process of encryption and decryption by the technique make the encryption process slow and costly as well. Although this technique is suitable for both software and hardware implementation.
  • 13. Byte Based Parity Technique The byte based parity technique makes use of parity checking at each byte of plain text to combat error. Yet it has the limitation of not being able to correct all type of error vectors. Also the byte based parity technique is not suitable for software implementation.
  • 14. Selective AES The performance of any cryptosystem is measured by two parameters: level of security it provides and speed of encryption and decryption. Selective AES is the process of AES encryption of a fraction of message keeping the remaining portions unencrypted, which increases in the speed of encryption / decryption.
  • 15. Error Propagation in Selective AES The Error Propagation Effect has been observed in Selective AES through an experiment conducted with the following 640-bit message and 128-bit key: Message = “ERROR PROPAGATION EFFECT OF AES HAS THROWN A GREAT RESEARCH CHALLENGE BEFORE US.” Key = “BIKRAMJIT SARKAR”
  • 16. Error Propagation in Selective AES contd.. First the entire message is divided into 5 blocks B1, B2, B3, B4 and B5, each of which is of 128 bits, where, B1 = “ERROR PROPAGATIO”, B2 = “N EFFECT OF AES ”, B3 = “HAS THROWN A GRE”, B4 = “AT RESEARCH CHAL” and B5 = “LENGE BEFORE US.”
  • 17. Error Propagation in Selective AES contd.. The AES Encryption process has 10 rounds. And after each round (1st to 9th) one intermediate state is found and after the 10 rounds the final cipher text is generated. So, each block, while being encrypted to generate the final cipher, generates 9 intermediate states. Now after first round, the eighth bit of each intermediate state generated from each block get forcibly infected. The same thing is repeated for each of the blocks after fifth and ninth rounds.
  • 18. Error Propagation in Selective AES contd.. Now the erroneous ciphers are compared with the error free cipher and accordingly the total number of errors at the output ciphers for each of the cases is found. The average number of errors occurred in the output cipher when a single bit error has occurred in any of the intermediate states generated during the encryption process is also obtained and Table 2 is formed: Block(s) Encrypted Percentage of Selection Average number of errors occurred at the output after the execution of the Encryption Module B1 20 45.67 B1 & B2 40 88.67 B1, B2 & B3 60 133.67 B1, B2, B3 & B4 80 176.0 B1, B2, B3, B4 & B5 100 221.33 Table 2
  • 19. Error Propagation in Selective AES contd.. From Table 2 it is clear that lesser is the percentage of selection, lesser is the number of errors occurred at the output. But the processing speed is inversely proportional to the percentage of selection. So there must be a tread off between the security level and the processing speed since, the level of security is proportional to the percentage of selection.
  • 20. Redundancy Based Technique redefined It is clear that the Redundancy Based Technique requires the comparison of 128 bits since the plain text taken is of 128 bits. Here the proposed algorithm SBM 1.1 proposes to modify the Redundancy Based Technique that will reduce the overhead of comparison to only 8 bits. The algorithm makes use of the Longitudinal Redundancy Check (LRC) code.
  • 21. Redundancy Based Technique redefined contd.. Proposed Algorithm: SBM 1.1 1. Input the plain text P of 128 bits 2. Generate an LRC code (8 bits), say L, out of P 3. Encrypt P with AES Encryptor to find the cipher text, say C 4. Decrypt C with AES Decryptor to find P’ 5. Generate an LRC code (8 bits), say L’, out of P’ 6. L and L’ are now compared. If L and L’ are found to be same, C is considered to be error-free and is transmitted through the channel.
  • 22. Redundancy Based Technique redefined contd.. Block Diagram: SBM 1.1
  • 23. Redundancy Based Technique redefined contd.. Experimental Results: Input Message P (128 bits): 42 69 6b 72 61 6d 6a 69 74 20 53 61 72 6b 61 72 LRC code generated from P (Say, L): 0 1 0 1 0 0 0 1 P is encrypted with the following key: 2B 28 AB 09 7E AE F7 CF 15 D2 15 4F 16 A6 88 3C
  • 24. Redundancy Based Technique redefined contd.. Experimental Results: A one-bit error is injected in the 8th bit position of the intermediate cipher generated after 7th round and the encryption process continues through 3 more rounds of AES encryption. As a consequence, an erroneous cipher text is generated at the output as follows: 8A 88 3E 2D DC 16 77 90 4D B3 05 3E CA 04 4D 0C P’ generated from the above erroneous cipher: B2 C9 A7 34 CF 60 C6 24 75 F5 4B CD 9F 97 3C 62 LRC code generated from P’(Say, L’): 1 1 1 1 0 1 0 1
  • 25. Redundancy Based Technique redefined contd.. Experimental Results: L and L’ are finally compared and it is observed that they are not same. This indicates that some error has occurred and the cipher text generated and hence it is useless to transmit. The proposed approach has minimized the overhead of comparison from 128 pairs of bits to 8 pairs of bits. Although an additional module of LRC generator that has been introduced in our approach causes an extra overhead of the new approach, yet the proposed technique is superior.
  • 26. A probabilistic approach to tackle Error Propagation Effect of AES The proposed scheme suggests that the plain text will be encrypted odd number of times, say n, with the same key to produce n number of cipher texts. It is assumed that the probability of occurrence of a single bit error amidst the rounds must not reach 0.5 so that out of n cipher texts at most (n – 1)/2 number of cipher texts may be erroneous whereas the least number of error free cipher texts is (n + 1)/2. Majority Rule is then applied over the n number of ciphers and as a result the error free cipher text is achieved.
  • 27. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Proposed Algorithm: SBM 1.2 1. Input the plain text P of 128 bits 2. Encrypt P n number of times with AES Encryptor to find ciphers {Ci, i = 1 to n}, n being odd 3. Majority rule is applied over Ci to find the error- free cipher text C which is directly transmitted through the channel.
  • 28. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Block diagram: SBM 1.2
  • 29. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Experimental results: Input Message P (128 bits): 42 69 6b 72 61 6d 6a 69 74 20 53 61 72 6b 61 72 Key K (128 bits): 2B 28 AB 09 7E AE F7 CF 15 D2 15 4F 16 A6 88 3C Error-free Cipher: FC 41 16 48 BE C0 16 A7 FC 5C 3F 43 F4 13 F4 A0
  • 30. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Experimental results: The Plain text P has been encrypted by the key 5 numbers of time and as per the assumption at most 2 numbers of ciphers texts may be erroneous. The erroneous cipher texts are as follows: Erroneous cipher text 1: D2 06 F1 26 BE 27 E9 EE FA CF AF 6B 55 25 D6 0D Erroneous cipher text 2: 24 E2 C9 55 1C 30 E0 58 5D D7 3C 64 EA 2F CA F7
  • 31. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Experimental results: As per the assumption, the rest of the cipher texts (3 numbers) are error free. The Majority Rule is then applied over the 5 cipher texts and finally the error-free output cipher is generated as follows: FC 41 16 48 BE C0 16 A7 FC 5C 3F 43 F4 13 F4 A0
  • 32. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Experimental results: As per the assumption, the rest of the cipher texts (3 numbers) are error free. The Majority Rule is then applied over the 5 cipher texts and finally the error-free output cipher is generated as follows: FC 41 16 48 BE C0 16 A7 FC 5C 3F 43 F4 13 F4 A0
  • 33. A probabilistic approach to tackle Error Propagation Effect of AES contd.. Although in the experiment the plain text has been encrypted 5 times, It can be realized that with the increase in the number (n) of cipher texts, the scheme guarantees reliable communication on its part, even with the occurrence of error with higher probability (0.5 * (1 – 1/n)); that means that if n increases the tolerance also increases.
  • 34. A Hybrid Architecture to overcome the Error Propagation Effect of AES The algorithm SBM 1.3 proposes a hybrid architecture to ward off the Error Propagation Effect of AES. The proposed scheme is the combination of the algorithms SBM 1.1 and SBM 1.2, that makes use of both the Longitudinal Redundancy Check code and the Majority Rule in parallel. The algorithm checks for the effect of error propagation in 2 different paths and finally the error-free cipher text is obtained ensuring a reliable communication.
  • 35. A Hybrid Architecture to overcome the Error Propagation Effect of AES contd.. Proposed algorithm: SBM 1.3 1. Input the plain text P of 128 bits 2. Generate an LRC code (8 bits), say L1, out of P 3. Encrypt P with AES Encryptor to find the cipher text, say C’ 4. Decrypt C with AES Decryptor to find P’ 5. Generate an LRC code (8 bits), say L2, out of P’ 6. L1 and L2 are now compared. If L1 and L2 are found to be same, C’ is fed to the comparator.
  • 36. A Hybrid Architecture to overcome the Error Propagation Effect of AES contd.. Proposed algorithm: SBM 1.3 7. Encrypt P n number of times with AES Encryptor to find ciphers {Ci, i = 1 to n}, n being odd 8. Majority rule is applied over Ci to find the cipher text C” which is also fed to the comparator. 9. If C’ and C” are found to be same, say C, it is considered to be error-free and is transmitted through the channel.
  • 37. A Hybrid Architecture to overcome the Error Propagation Effect of AES contd.. Block diagram: SBM 1.3
  • 38. A Hybrid Architecture to overcome the Error Propagation Effect in Selective AES Proposed algorithm: SBM 1.4 Input Key Words for the message of N words. Message is divided into K parts each of k blocks. Each block is of M words. Find the occurrence of any keyword in the blocks, starting with the first block in the first part. If it occurs, encrypt that block using the algorithm SBM 1.3 and all the blocks thereafter in the part. Repeat (1-2) for all parts, j=1 to K. When j=K, the proposed scheme of encryption is complete.
  • 39. A Hybrid Architecture to overcome the Error Propagation Effect in Selective AES Block diagram: SBM 1.4
  • 40. Overall conclusion In this thesis we have studied the effect in detail. The relative mathematical analysis has been cited in the thesis. Error propagation effect in case of selective AES and its comparison with normal AES has also been studied. In this thesis four algorithms, viz. SBM 1.1, SBM 1.2, SBM 1.3 and SBM 1.4, have been proposed for preventing Error Propagation Effect of AES. The corresponding experimental results have also been provided so as to prove the algorithms to be efficient to overcome the error propagation effect of the Rijndael used as Advanced Encryption Standard.
  • 41. Future scope It is assumed that the probability of occurrence of a single bit error amidst the rounds is maximum 50%. Further research works may be extended to achieve certainty in preventing the Error Propagation Effect of AES algorithm under all circumstances. When the number of rounds is increased in AES, the complexity of AES encryption and decryption also increases. So, the length of the key may increased to 512 bits in order to increase the number of rounds. Future research may include the above mentioned considerations.