Content uploaded by Tito Waluyo Purboyo
Author content
All content in this area was uploaded by Tito Waluyo Purboyo on Mar 04, 2016
Content may be subject to copyright.
Sosemanuk Algorithm forEncryption and Decryption
Video on Demand (VoD)
1Siska Selany, 2Surya Michrandi Nasution &3Tito Waluyo Purboyo
123 Security Laboratory Electrical Engineering Faculty
Telkom University
Bandung, Indonesia
1siskaselany@gmail.com 2michrandi@telkomuniversity.ac.id 3titowaluyo@telkomuniversity.ac.id
Abstract. Video on Demand (VoD) service have a concept
that the user will have complete freedom to choose what he
wants to see. VoD gives satisfaction is more personal than
the interests of the people. The development of information
technology through multimedia services such as Video on
Demand cannot be separated from the issue of data
security. Security is very important to maintain the data of
those who do not have the right to see it. Therefore, it needs
a method of concealment of information / data that
cryptography to be able resolve this security issue. This
research regarding Sosemanuk Algorithm for encrypt and
decrypt multimedia data such as Video on Demand and to
demonstrate the performance of this algorithm.
Performances that calculated in this paper is the encryption
time, decryption and avalanche effect.
Keywords : Video on Demand (VoD), Cryptography, Sosemanuk
Algorithm.
I. INTRODUCTION
Nowadays, people interested in Video on Demand because
it has many advantages. Everything is run interactively using
buttons and a simple command we can pause, rewind, fast
forward, or whatever we want [1]. Systems on Video on
Demand is the first video file has been stored on the server so
that the client can request the desired video to watch. [2]. But it
could not be separated from the issue of data security and it
requires cryptography to secure the data. Cryptography is the art
and science of keeping messages secure [4].
Cryptography has two kinds of algorithm, there are
symmetric algorithm and asymmetric algorithm. Symmetric
algorithm is divided into two types, there are stream cipher and
block cipher. Sosemanuk is a symmetric algorithm type of
stream cipher. Its symmetric algorithm because has the same key
for encryption and decryption process. Sosemanuk type of
stream cipher because the operation of plain text/cipher text on
the form of a single bit. In this case a series of bits encrypted /
described bit by bit.. Sosemanuk algorithm is based on two
cryptography algorithms, there is a stream cipher Snow and
block cipher Serpent in hopes the weakness of each algorithm is
covered by other algorithm strength [5].
This research presented encryption and decryption data of
video files using Sosemanuk Algorithm. This paper will explain
about the level of security from Sosemanuk Algorithm through
avalanche effect and calculate the time of encryption and
decryption.
II. VIDEO ON DEMAND
Video on Demand (VoD) is an interactive multimedia system,
the difference being that the user can select a movie from a large
video database. Individual user in an area is able to watch
different programs when they wish to, making the system a
realization of the video rental shop brought into the home. The
user can call on a range of services. While watching movies,
performing operations such as video selection, pause,
rewinding, etc. can be selected as if it were a video player.
Video on Demand is an important aspect of interactive media
technology that is currently conquering the information systems
worldwide. The main components of a video on demand service
are shown in figure: The video server to store and provide access
to programs, the data delivery network to interconnect the
subscriber [3].
As we know that the large companies such as PT.
Telekomunikasi Indonesia Tbk has a new flagship product is
Usee TV that uses the concept of Video on Demand. VoD in
Usee TV presenting video content in the form of a catalog, so
the customer can choose the desired video. Concept of video on
demand will be growing as technology develops today.
III. SOSEMANUK ALGORITHM
Sosemanuk algorithm was designed by Come Berbain,
Olivier Billet, Anne Canteaut, Nicolas Courtois, Henri
Gilbert, Louis Goubin, Aline Gouget, Louis Granboulan,
Cédric Lauradoux, Marine Minier, Thomas Pornin and
Hervé Sibert and in November 2004 followed the e-STREAM
competition. The competition aims to find the stream cipher
algorithm that robust and can be used for the different purposes.
After going through three stages on April 15th 2008,
Sosemanuk selected with three another algorithm as a finalist
in Software-Oriented field [6].
2015 IEEE Asia Pacific Conference on Wireless and Mobile
125
The key length between 128 – 256 bits and IV (Initial Value)
128 bit. Sosemanuk has a keystream generation process before
do the encryption and decryption process, here are the
Sosemanuk algorithm process [4] :
1. 128 bit key is the first step of the process
2. Padding bits, 128 bit key are expanded into 256 bits. The
way is added bit 1 followed by bit 0 and so on until it
reaches 256 bits, then divided into 8 blocks ( W0 – W7 )
. Each block consists of 32 bits.
3. Key schedule, the key scheduling process established 25
sub key ( K0-K24 ) using Serpent Algorithm’s principle.
W0 – W7 block expanded into W0 – W99 based equation
:
W[i] = (Wi-8Wi-5 Wi-3 Wi-1 Фi) <<< 11 …..(3.1)
And then to get the 25 sub keys ( K0-K24 ), blocks above
are grouped based on the following rules :
Ki = {K4i, K4i+1, K4i+2, K4i+3} …..(3.2)
Where
{K4i, K4i+1, K4i+2, K4i+3}={W4i, W4i+1, W4i+2, K4i+3};
0≤i ≤24
4. From the key schedule, we get 25 sub keys that become
the input to the next process Serpent24. Serpent24 is a
Serpent algorithm implementation with 24 rounds. At
the beginning of this process, B0 = Initial Value. The
process for each round are :
a. Key mixing :
Ki Bi …..(3.3)
b. Linear Transformation : The result of key
mixing, we do linear transformation to produce
Bi+1 . Except for the last round is not done linear
transformation but replaced with additional
mixing key operation. Linear transformation
can be defined as follow :
X1, X2, X3, X4=S0 (Bi Ki)
X3<<< 13 = X3
X1<<< 3 = X1
X2X3 X1 = X2
X0X1 (X3<< 3) = X0
X2<<< 1 = X2
X0 <<< 7 = X0
X3 X2X0 = X3
X1X0(X2<< 7) = X1
X3<<< 5 = X3
X1<<< 22 = X1 …..(3.4)
S0 is Serpent S-boxes. Serpent S-boxes, implemented in
bitslice mode. These circuits have been published by Dag
Arne Osvik ("Speeding up Serpent", published in the 3rd
AES Candidate Conference) and work on five 32-bit
registers: the four inputs, and a fifth scratch register. There
are meant to be quite fast on Pentium-class processors.
These are not the fastest published, but they are "fast
enough" and they are unencumbered as far as intellectual
property is concerned (note: these are rewritten from the
article itself, and hence are not covered by the GPL on Dag's
code, which was not used here) [7].
The output bits are permuted. Here is the
correspondence:
S0: X1, X4,X2, X0
S1: X2, X0, X3, X1
S2: X2, X3, X1, X4
S3: X1, X2, X3, X4
S4: X1, X4, X0, X3
S5: X1, X3, X0, X2
S6: X0, X1, X4, X2
S7: X4, X3, X1, X0 …..(3.5)
5. Results from the 12th round, 18th, and 24th selected
as the initial state of LFSR and FSM.
S[6], S[7], S[8], S[9] = (X3, X2, X1, X0)12
S[5], R[2], S[4], R[1] = (X3, X2, X1, X0)18
S[0], S[1], S[2], S[3] = (X3, X2, X1, X0)24…..(3.6)
Before keystream formed, we do update LFSR and FSM.
- LFSR (Linear Feedback Shift Register) should be
small, so that both the code and the storage size
will be better so we selected LFSR of length 10.
LFSR operates on Sosemanuk as follows :
St+9= St+9 α-1 St+3 αSt …..(3.7)
- FSM (Finite State Machine) is a component with
64 bits of memory, corresponding to two 32 bit
register R1 and R2. At each step, the FSM takes as
inputs some words from the LFSR state. It updates
the memory bits and produces a 32 bit output [5] .
The FSM operates on Sosemanuk as follows :
R1t = (R2t-1 + mux (lsb (R1t-1), St+1, St+1 St+8)) mod 232
R2t = Trans (R1t-1) …..(3.8)
Where :
- LSB(X) is Least Significant Bit from X
- mux(C,X,Y) is X if C = 0 or Y if C = 1
- Trans(Z) = M . Z mod 232<<< 7
M is constanta 0x54655307 (Hexadecimal from
ten digits beginning of π).
Output from FSM :
Ft =(St+9+ R1t mod 232) R2t …..(3.9)
6. We form the keystream of Sosemanuk Algorithm with
this equation :
(Z[t+3], Z[t+2], Z[t+1], Z[t]) =
(f[t+3],f[t+2], f[t+1], f[t] )(S[t+3],S[t+2],S[t+1],S[t])…..(3.10)
7. Encryption on Sosemanuk Algorithm is or between
plain text with keystream and the decryption on
Sosemanuk Algorithm is or between cipher text with
keystream.
2015 IEEE Asia Pacific Conference on Wireless and Mobile
126
IV. SYSTEM DESIGN
System design includes general groove of the system,
flowcharts of system, the generate keystream process and
encryption decryption process of the system.
A. System Descryption
Specification of the system is to implement the Sosemanuk
Algorithm on video file. The algorithm will encrypt and decrypt
frame by frame of a video file. The application has two users,
there are member and nonmember. Non members will see the
encrypted video and the member will see the decrypted video.
The application that design is based on VoD concept. That user
can see the video what he wants that can play back, stop and
pause the video.
VoD
APPLICATION
Server
Decrypted
video Member
Encrypted
video
Non member
Key of Sosemanuk
Fig1. Design of the system
B. Generate Keystream on Sosemanuk Algorithm
A stream cipher is a symmetric key cipher where plaintext
digits are combined with a pseudorandom cipher digit stream
(keystream).There are the step of generate keystream on
Sosemanuk algorithm [8] :
1. Initialize Key and IV
2. Padding bit key into 256 bits
3. Generate 25 subkeys
4. Key Mixing
5. Linear Transformation
6. Update LFSR and FSM
After the above steps do then we can form the keystream of
Sosemanuk and then can do the encryption and decryption
process.
C. Encryption & Decryption Process
Flowchart diagrams of encryption and decryption process
using Sosemanuk algorithm can be seen below:
Start
Original video
Video into frames
Encryption of video
frames (keystream
⊕ data frame)
Finish
Frames into video
Encrypted video
Start
Encrypted video
Video into frames
Decryption of video
frames (keystream
⊕ data frame)
Finish
Frames into video
Decrypted video
(a) (b)
V. EXPERIMENTAL TESTING
Experimental testing is performed how long time from
encrypt and decrypt video files and calculates the avalanche
effect of the Sosemanuk algorithm that implemented on the
video.
A. Encryption and Decryption Time
Here is a list of tables testing how long it needed for
encryption and decryption process.
Table 5.1 Video data and time
Video
The
number of
frames
Encryption
time (minute)
Decryption
time
(minute)
Video1.
mp4
421
15,42
3,41
Video2.
mp4
602
20,22
4,44
Video3.
mp4
783
28,36
6,43
B. Encryption and Decryption Process
Table 4.2 shows the comparison of frame video before and
after encryption. That is the sample frame of video. The table
shows the plaintext (original frame), ciphertext (encrypted
frame) and decrypted frame.
Fig3. (a). Sosemanuk encryption (b).
Sosemanuk decryption process
2015 IEEE Asia Pacific Conference on Wireless and Mobile
127
Table 5.2 Encryption and Decryption Frame Video Result
Plaintext
Ciphertext
Decryption result
C. Avalanche Effect
Avalanche effect is a way to determine whether or not a
cryptographic algorithm. By counting avalanche effect, we will
know how much the changes that occur in the ciphertext bits due
to the encryption process. The higher value of the avalanche
effect will make better the level of security algorithms. Here is
an equation of avalanche effect :
….(5.1)
Here is an avalanche effect table testing of Sosemanuk
Algorithm. The first test is on Table 4.3 with the different video
that showed in table and the key and IV are same for each video.
Key : 4F46129B1ADFAD325B5CCB469D1AFB87
IV : 11 81 BC A5 AD 1E 17 B4 45 24 83 85 36 BA AC 21
Table 5.3 Avalanche Effect Testing 1
No
Name
Plaintext
Ciphertext
AE (%)
1.
Video1.mp4
d1 20 0 20 5f df
9b befb ef be fb ef
be fb efbe fb ef be
fb ef be fbef be fb
ef be fb ef be
6d 4a 88 ff d8 f8
a9 9fcb 23 45 11
fa 41 64 14ab 92
46 b7 55 5a f0 10
6c 9d e2 36 6a
49 8f 2d
52,34375
2.
Video2.mp4
40 0 40 c7 fd ea fb
efbe fb ef be fb ef
be fbef be fb ef be
fb ef befb ef be fb
ef be fb ef
fc 6a c8 18 7a cd
c9 ce8e 37 14 54
ee 10 21 0fa d7
52 e6 10 4e a1
5578 cc a7 22 3b
c 9b 7c
52,34375
3.
Video3.mp4
7a 17 0 1 2 f5 bb
77df 7d f7 df 7d
f7 df 7df7 df 7d f7
df 7d f7 df7d f7 df
7d f7 df 7d f7
c6 7d 88 de 85
d2 89 56ef b1 c
35 68 8 40 86e2
b6 d4 fe 71 c8 b9
34fe d4 c6 a4 23
6d 1d 64
52,34375
The second test is on Table 4.4 with the different key that
showed on table, the same video that is Video1.mp4 and the
same IV. IV that used is : 11 81 BC A5 AD 1E 17 B4 45 24 83
85 36 BA AC 21
Table 5.4 Avalanche Effect Testing 2
No
Name
Key
Plaintext
Ciphertext
AE (%)
1.
Video1.
mp4
4f 46 12
9b1adfad
32 5b5ccb
46
9d1afb87
d1 20 0 20
5f df 9b befb
ef be fb ef be
fb efbe fb ef
be fb ef be
fbef be fb ef
be fb ef be
6d 4a 88 ff d8
f8 a9 9fcb 23
45 11 fa 41 64
14ab 92 46 b7
55 5a f0 10
6c 9d e2 36
6a 49 8f 2d
52,34375
2.
Video1.
mp4
4e 46 12
9b1adfad
32 5b5ccb
46 9d1afb
87
d1 20 0 20
5f df 9b befb
ef be fb ef be
fb efbe fb ef
be fb ef be
fbef be fb ef
be fb ef be
ac 2d 0 92 2d
b3 7c d92d 24
d9 31 e2 ec c4
8029 cb 6b 76
e5 fc 13 6e
b4 76 4a 1e 3
fd 7c 19
51,171875
3.
Video1.
mp4
4e 56 12
9b1adfad
32 5b5ccb
46 9d1afb
87
d1 20 0 20
5f df 9b befb
ef be fb ef be
fb efbe fb ef
be fb ef be
fbef be fb ef
be fb ef be
a c0 7d 86 64
b4 f4 c171 e5
ed c8 97 f4 18
7916 a6 d0 1
5e 55 f9 ce
ae 96 5a e1 4d
35 ab 67
53,515625
The third test is on Table 4.5 with the different IV that
showed on table, the same video and the same key. Key that used
is : 4F46129B1ADFAD325B5CCB469D1AFB87
Table 5.5 Avalanche Effect Testing 3
No
Name
IV
Plaintext
Ciphertext
AE (%)
1.
Video1.
mp4
11 81 bc
a5 ad 1E
17 b4 45
24 83 85
36 ba ac
21
d1 20 0 20
5f df 9b befb
ef be fb ef be
fb efbe fb ef
be fb ef be
fbef be fb ef
be fb ef be
6d 4a 88 ff
d8 f8 a9 9fcb
23 45 11 fa
41 64 14ab
92 46 b7 55
5a f0 106c 9d
e2 36 6a 49
8f 2d
52,34375
2.
Video1.
mp4
12 81 bc
a5 ad 1E
17 b4 45
24 83 85
36 ba ac
21
d1 20 0 20
5f df 9b befb
ef be fb ef be
fb efbe fb ef
be fb ef be
fbef be fb ef
be fb ef be
3f b1 cd 16
f7 55 47 c2fe
cc 74 b9 2a
61 af 8e88 d1
51 14 6c 77
91 bd85 3e
89 b9 93 15
15 76
49,609375
3.
Video1.
mp4
12 91 bc
a5 ad 1E
17 b4 45
24 83 85
36 ba ac
21
d1 20 0 20
5f df 9b befb
ef be fb ef be
fb efbe fb ef
be fb ef be
fbef be fb ef
be fb ef be
6b 12 7a d3
74 b 18 b21f
e0 47 77 fa
64 49 9e
62 78 b 9b 8c
e5 62 cb1b
84 99 5d 47
da b6 c8
50,0
D. Analysis
From the Table 5.1 above, we can see the comparison of the
time required by three videos with different frame. The first
video with 421 frames has duration of time from encryption and
decryption is more efficient than the second video with 602
frames. The first video needs 15,42 minutes for encryption and
3,41minute for decryption while the second video is needed
2015 IEEE Asia Pacific Conference on Wireless and Mobile
128
20,22minute for encryption and 4,44 minutes for decryption.
And for the third video, it is taking a more time to encryption
and decryption, it is about 28,36 minute for encryption and 6,43
minute for decryption. From that comparison, it can be
concluded that the number of frames affects the time of
encryption and decryption. The more number of frames so the
longer the time of the encryption and decryption process
The design of the encryption process video to MP4 video can
be implemented and successfully executed. Encryption and
decryption process can produce a video with the same format as
the original video, which Mp4 and can be run on the
VoDapplication.We can see in Table 5.2 that show the original
frame, the encrypted and decrypted frame of the video.
From test resultsTable 5.3 above, we can see the avalanche
effect value from the different video, the same key and IV is
static on 52,34375%. So if we use the same key and IV for
different video it will have the same avalanche effect value.
Test result from Table.5.4that testing use the different key for
the same video and the same IV produce the different value of
avalanche effect for each test. The first is 52,34375%, the
second is 51,171875 % and the last is 53,515625 %. So if we
use the different key for the same video, it will have the
different avalanche effect value.
The last result from Table 4.5 that using the different IV for
the same video and the same key produce the different value of
avalanche effect for each test. The first is 52,34375 %.The
second is 49,609375 % and the last test is 50,0 %. It means if
we use the different IV for the same video and the same key it
will produce the different avalanche effect value.
All result shows that the avalanche effect test produce
changes bits is about half the number of bits in its output.
Cryptographic algorithmswill bedifficult to resolvewhen
thekeyused is notknownandthe resultsofthe outputwill
beveryunique. It means Sosemanuk algorithm has a good
performance on security level.
VI. CONCLUSIONS
This paper presents an implementation of Sosemanuk
Algorithm for Video on Demand. The algorithm encrypted the
video and then decrypt the video frame by frame.
From the implementation and test results, we draw some
concluding remarks, as follows :
1. The difference of video frames can affect to time of
encryption and decryption. The larger the number of
frames, the longer they need time to encrypt and
decrypt the video file.
2. Video encryption with Sosemanuk Algorithms can
generate frames are very random, even different from
the original frame.
3. The changes bits after testing avalanche effect value is
about half of number of bits in its output. It means
Sosemanuk Algorithm has a good performance on
security level.
4. Sosemanuk Algorithm that based on stream cipher
Snow Algorithm and block cipher Serpent Algorithm in
a reduced version on duration time encrypts, decrypt
and the avalanche effect test becomes more efficient
and safer.
REFERENCES
[1]
Peltoniemi, Jarry, "Video on Demand
Overview," 1995.
[2]
Shujun Li, Guanrong Chen, Albert Cheung,
Bharat Bhargava, Kwok-Tung Lo;, "On the Design
of Perceptual MPEG-Video Encryption
Algorithms," IEEE Transactions on Circuit and
Systems for Video Technology, vol. 17 no 2, pp.
214-223, Februari 2007.
[3]
Kong Feng, Yang Xudong, "A Study on
Grid Based VOD System in the E-Learning,"
International Forum on Information Technology
and Applications, pp. 407-410, 2009.
[4]
R. Munir, Kriptografi, Bandung:
Informatika, 2006.
[5]
Jansen, "Studi dan Analisis Algoritma
Sream Cipher Sosemanuk," Institut Teknologi
Bandung.
[6]
C.Berbain, O.Billet, A.Canteaut,
N.Courtois, H.Gilbert, L.Goubin, A.Gouget,
L.Granboulan, C. Lauradoux, M.Minier, T. Pornin
and H.Sibert, "Sosemanuk, a Fast Software-
Oriented Stream Cipher," 2007.
[7]
"e-Crypt," [Online]. Available:
http://www.ecrypt.eu.org/stream/sosemanuk.html.
[Accessed 1 September 2014].
[8]
Endro Ariyanto, Trisya Indah Pravitasari,
Setyorini, "Analisa Implementasi Algoritma
Stream Cipher Sosemanuk dan Dicing dalam
Proses Enkripsi Data," ISSN 1979-2328, pp. 351-
357, 2008.
2015 IEEE Asia Pacific Conference on Wireless and Mobile
129
