View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

provided by Directory of Open Access Journals

Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

RESEARCH ARTICLE OPEN ACCESS

Modifying Authentication Techniques in Mobile Communication

Systems
Zakaria Zakaria Hassan, Talaat A. Elgarf, Abdelhalim Zekry
Communication Engineering Department Higher Technological Institute Cairo, Egypt
Communication Engineering Department Higher Technological Institute Cairo, Egypt
Communication Engineering Department Faculty of Engineering Ain shams University Cairo, Egypt

Abstract
Milenage algorithm applies the block cipher Rijnadael (AES) with 128 bit key and 128 bit block size. This
algorithm is used in the 3GPP authentication and key generation functions (f1, f1*, f2, f3, f4, f5 and f5*) for
mobile communication systems (GSM/UMTS/LTE). In this paper a modification of Milenage algorithm is
proposed through a dynamic change of S-box in AES depending on secret key. To get a new secret key for
every authentication process we add the random number (RAND) transmitted from the authentication center
(AUC) to the contents of the fixed stored secret key (Ki) and thus the initialization of the AES will be different
each new authentication process . For every change in secret key a new S-box is derived from the standard one
by permuting its rows and columns with the help of a new designed PN sequence generator. A complete
simulation of modified Milenage and PN sequence generator is done using Microcontroller (PIC18F452).
security analysis is applied using Avalanche test to compare between the original and modified Milenage . Tests
proved that the modified algorithm is more secure than the original one due to the dynamic behavior of S-box
with every change of the secret key and immunity against linear and differential cryptanalysis using Avalanche
tests. This makes the modified Milenage more suitable for the applications of authentication techniques
specially for mobile communication systems.
Keywords—Authentication vector (AV), Modified MILENAGE Algorithm for AKA Functions
(F1,F1*,F2,F3,F4,F5,F5*), AES ,Dynamic S-BOX and PN Sequence Generator(LFSR).

I. INTRODUCTION both encryption and decryption, the cipher begins

Authentication includes the authenticity of with an Add Round Key stage, followed by nine
the subscriber as well as the network. Authentication rounds that each includes all four stages, followed by
of mobile subscribers and network operators is a a tenth round of three stages.[4].
challenge of future researchers due to increasing This paper is organized as follows: In
security threats and attacks with the enhanced volume Section II, authentication schemes in mobile
of wireless traffic. Authentication schemes in mobile communications are described. In Section III, a
communication systems are initiated during IMSI proposed authentication scheme is presented
attach, location registration, location update with depending on the dynamic change of S-box in AES ,
serving network change, call setup (MOC, MTC), the new Secret key for every authentication process
activation of connectionless supplementary services and the new PN sequence generator. In Section IV, a
and short message services (SMS). complete simulation of the modified Milenage
Milenage algorithm is used for generating algorithm and the Avalanche test results are
authentication and key agreement of cryptographic introduced. Discussions and Conclusions are
generating functions (MAC, XRES, CK and IK). The presented in Section V.
main core of Milenage algorithm is the Advanced
Encryption Standard (AES) [1] which launched as a II. AUTHENTICATION SCHEMES IN
symmetrical cryptographic standard algorithm by the MOBILE COMMUNICATIONS.
National Institute of Standard and Technology (NIST) (i) GSM/GPRS Authentication and Key
in October 2000, after a four year effort to replace the Agreement vectors .
aging DES. The Rijndael proposal for AES defined a There exists a permanent, shared secret key
cipher in which the key length can be independently Ki for each subscriber. This permanent key is stored
specified to be 128 , 192 or 256 bits but the input and in two locations: in the subscriber’s SIM card and in
output block length is 128 bits.[2],[3].Four different the Authentication Centre (AuC). The key Ki is never
stages are used in AES :Sub Byte transformation, moved from either of these two locations.
Shift Rows, Mix Columns and Add Round Key. For Authentication of the subscriber is done by checking

www.ijera.com 67 | P a g e
Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

that the subscriber has access to Ki. This can be To generate authentication quintuple, the
achieved by challenging the subscriber by sending a HLR\AUC computes a message authentication code
random 128-bit string (RAND) to the terminal. The for authentication MAC-A= f1k (SQN || RAND ||
terminal has to respond by computing a one-way AMF), the length of MAC-A is 64bits. an expected
function with inputs of RAND and the key Ki, and response XRES = f2k (RAND), the length of XRES is
returning the 32-bit output Signed Response (SRES) 64bits.. a cipher key CK = f3k (RAND). the length of
to the network. Inside the terminal, the computation of CK is 128bits . an integrity key IK = f4k (RAND) the
this one-way function, denoted by A3, happens in the length of IK is 128bits and an anonymity key AK =
SIM card. During the authentication procedure, a f5k (RAND), the length of AK is 48bits that is used to
temporary session key Kc is generated as an output of conceale sequence number SQN, the length of SQN is
another one-way function A8. The input parameters 48bits, SQN= SQN ⊕ AK. the HLR/AuC aggregates
for A8 are the same as for A3: Ki and RAND. The the authentication token AUTN = SQN [⊕ AK] ||
session key Kc is subsequently used to encrypt AMF(16bits) || MAC-A, the lengths of AUTN is
communication on the radio interface. The serving 128bits that forms the quintet Q =AV= (RAND,
network does not have direct access to the permanent XRES, CK, IK, AUTN). [7], [8], [9].
key Ki, so it cannot perform the authentication alone.
Instead, all relevant parameters, so called the . Authentication and key derivation in the
authentication triplet (RAND, SRES and Kc) are sent USIM.
to the serving network element Mobile Switching Upon receipt of a (RAND, AUTN), the
Centre/Visitor Location Register (MSC/VLR) (or USIM computes the anonymity key AK = f5k
Serving GPRS Support Node (SGSN) in the case of (RAND) and retrieves the unconcealed sequence
GPRS) from the AuC. [5], [6]. number SQN = (SQN ⊕AK) ⊕ AK, XMAC-A = f1k
(SQN || RAND|| AMF), the response RES = f2k
(ii) UMTS/LTE/Advanced LTE Authentication and (RAND), the cipher key CK = f3k (RAND) and the
Key Agreement Vectors. integrity key IK = f4k (RAND) as shown in fig.2. [5],
. UMTS Generation of Authentication vectors [6].
(Quintets) in the AUC. RAND AUTN

Upon the receipt of the authentication data

f5 SQN  AK AMF MAC
request from the VLR/SGSN, The HLR/AuC sends an
authentication response back to the VLR/SGSN that AK 
contains an ordered array of n authentication vectors
AV(1..n). The HLR/AuC starts with generating a SQN
K
fresh sequence number SQN and an unpredictable
challenge RAND. The authentication vectors are
ordered based on sequence number. [5]. f1 f2 f3 f4
There are eight Cryptographic functions used
in UMTS/LTE/Advanced LTE Authentication and XMAC RES CK IK
Key Agreement to generate Authentication vector
(AV). f0 is the random challenge-generating function. Verify MAC = XMAC

It should be a pseudo random number-generating Verify that SQN is in the correct range
function and map the internal state of the generator to Figure1. Authentication and key derivation in the
the challenge value RAND , the length of RAND is USIM.[7].
128 bits. The f1 is the network authentication
function, f1* is the re-synchronization message
(iii) LTE\AdvancedLTE Generation of
authentication function, it is used to provide data Authentication Vectors in the HSS.
origin authentication for synchronization failure The LTE architecture is built on the existing
information sent by the USIM to the AuC, f2 is the architecture from UMTS. LTE standards reuse the
user authentication function, f3 is the cipher key authentication and key-agreement of UMTS. The
derivation function , f4 is the integrity key derivation LTE/Advanced LTE AKA protocol also known as the
function, f5 is the anonymity key derivation function Evolved Packet System (EPS) AKA protocol. The
for normal operation and f5* is the anonymity key EPS-AKA protocol is executed between UE and the
derivation function for re-synchronization, f5* is only MME instead of between the USIM and the
used to provide user identity confidentiality during VLR/SGSN. The AuC generates UMTS AVs for EPS
resynchronization. K is the subscriber authentication AKA in exactly the same format as for UMTS AKA.
key stored in the USIM and at the AuC, The length of The HSS part outside the AuC derives KASME from
K is 128 bits. [5],[7],[8]. the CK and IK. EPS AV consists of [RAND, XRES, a

www.ijera.com 68 | P a g e
Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

local master key KASME and an AUTN] as shown in RAND

fig.1. [10], [11], [12]. OPC OP EK OPC

SQN||AMF||SQN||AMF EK

OPC OPC OPC OPC OPC

rotate rotate rotate rotate rotate

by r1 by r2 by r3 by r4 by r5
c1 c2 c3 c4 c5
EK EK EK EK EK

OPC OPC OPC OPC OPC

f1 f1* f5 f2 f3 f4 f5*
Figure 3. Computation of the MILENAGE functions.
[13].

AUTN = [SQN ⊕ AK || AMF || MAC]. Upgrade of S-box (Dynamic S-box) depends

UMTS (AV) = [RAND || XRES || CK || IK || AUTN]. on the new secret key (Key⊕RAND) for every
EPS (AV) = [RAND || XRES || KASME|| AUTN]. authentication process and the new PN Random
Figure2. Generation of UMTS and EPS authentication sequence generator[14] . The suggested generator
vectors. [6]. consists of three Maximal length Linear Feedback
Shift Register (LFSR) with thirty two, seventeen and
III. Proposed Authentication Scheme in fifteen taps. The period of this PN sequence= (232-1)
Mobile Communication. (217-1) (215-1) the 1st 128 bits of the PN sequence
A modification of Milenage algorithm is generator is taken as the secret key to upgrade the s-
proposed through a dynamic change of S-box in AES box. The 1st 64 bits to rearrange the columns and the
depending on the new secret key. To get a new secret 2nd 64 bits to rearrange the rows of original S-box. The
key for every authentication process we add the feedback functions of the LFSRs are: [15].
random number (RAND) transmitted from the LFSR 1: F1=X15+X14 +1
authentication center (AUC) to the contents of the LFSR 2: F2=X32+X22+X2+X+1
fixed stored secret key (Ki) and so, the initialization of LFSR 3 : F3= X17+X14 +1
the AES will be different for each authentication To initialize the PN sequence generator as
process. For every change in secret key a new S-box shown fig.4, the new secret key is divided into two
is derived from the standard one by permuting its vectors of 64 bit length that are XORed to produce the
columns and rows with the help of a new designed initial state of the PN sequence generator (64bits). Let
PN sequence generator. Finally to get a strong the fixed stored authentication key Ki = [9E5944 AE
Milenage algorithm generating all functions f1, f1*, A94B8116 5C82FBF9 F32DB751] and the RAND
f2, f3, f4, f5, and f5* and the outputs of the various =[CE83DBC54AC0274A157C17F8D017BD6], the
functions used in User Authentication, Network new secret key = Ki⊕RAND =[50DA9
Authentication, Data Integrity Check and Ciphering F6BE38BA65C49FEEC01FE2CCC87].
data. The outputs of the various functions are then The initialization vector of the PN sequence
defined as shown in fig.3. generator (reshaped new secret key )
 Output of f1 = MAC-A, where MAC-A[0] .. =[50DA9F6BE38BA6 5C]⊕[49FEEC01FE2CCC87]
MAC-A[63] = OUT1[0] .. OUT1[63] = [1924736A1D A76ADB]. The 1st 64 bits of the PN
 Output of f1* = MAC-S, where MAC-S[0] .. sequence generator will be [09AE48DBC37F5612]
MAC-S[63] = OUT1[64] .. OUT1[127] used to rearrange columns of S-box and the 2nd 64 bits
 Output of f2 = RES, where RES[0] .. RES[63] = of the PN sequence generator will be[B9DE60C327
OUT2[64] .. OUT2[127] 458F1A] to rearrange the rows of S-box to have the
 Output of f3 = CK, where CK[0] .. CK[127] = final modified form.
OUT3[0] .. OUT3[127]
 Output of f4= IK, where IK[0] .. IK[127] =
OUT4[0] .. OUT4[127]
 Output of f5= AK, where AK[0] .. AK[47] =
OUT2[0] .. OUT2[47]
 Output of f5* = AK, where AK[0] .. AK[47] =
OUT5[0] .. OUT5[47]
Figure 4. PN random sequence generator.
www.ijera.com 69 | P a g e
Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

Table 1. AES standard S-BOX. (ii) For Modified AES-128

[Dynamic S-box] using PN sequence random
generator. Reshaped Secret Key 64 bit= [8E8C57BC
4EBCB9A6], Column dynamic S-box after
arrangement = [6093B714F2AEDC58] and final
dynamic S-box ROWs after arrangement = [A14FC
8D65B09E372].

Table 4. Modified AES (Dynamic S-box) – 128.

Plain text = CF 57 47 10 27 73 65 1A 6E 23
Table 2. For COLUMN Dynamic S-box after 88 18 A2 7C B9 EF
arrangement = [09AE48DBC37F5612]. Secret Key = 88 5C 36 49 B8 40 D9 E0 06 D0
61 F5 F6 FC 60 46
Cipher Text = 38 B1 4D 2A 56 81 2F 13 FF EE 38
69 FA A4 77 40

(iii) Avalanche test

Table 5. Results of Avalanche test due to change one

bit in plain text in standard AES.
Changed one bit (1)
Plain text = 4F 57 47 10 27 73 65 1A 6E
23 88 18 A2 7C B9 EF
Cipher Text= 23 B0 6F 4D 38 93 19 64 B4
DD 8A DE 7A 16 0A BC
Table3. Final S-box ROWs after arrangement Difference value = 91 A8 CA C2 99 1D AD D3
=[B9DE60C327458F1A] that used in Modified D0 AE F7 8F F9 21 81 F2
Milenage Algorithm during the new secret key to Ratio=51.56%
generate a new s-box so called [Dynamic Key (S- Changed one bit (15)
box)]. Plain text = CF 55 47 10 27 73 65 1A 6E
23 88 18 A2 7C B9 EF
Cipher Text= 51 75 5C B4 E9 B5 FD AF
C1 CA F3 FE F5 B4 5E E2
Difference value = E3 6D F9 3B 48 3B 49 18 A5
B9 8E AF 76 83 D5 AC
Ratio=53.90%
Changed one bit (69)
Plain text = CF 57 47 10 27 73 65 1A 66
23 88 18 A2 7C B9 EF
Cipher Text= 9F BB 05 72 7E 08 28 23 77
DE 80 21 2B 68 A1 BA
Difference value = 2D A3 A0 FD DF 86 9C 94 13
AD FD 70 A8 5F 2A F4
IV. SIMULAION AND RESULTS Ratio=53.90%
A complete simulation of the modified Changed one bit (115)
Milenage algorithm is achieved using Microcontroller Plain text = CF 57 47 10 27 73 65 1A 6E
(PIC18F452). The Avalanche tests are introduced to 23 88 18 A2 7C 99 EF
compare between the original and modified milenage. Cipher Text= DC 7F 25 84 3F 5D 38 46 73
86 0E 3B F7 54 AB 09
(i) For AES standard – 128 Difference value = 6E 67 80 0B 9E D3 8C F1 17
Plain text = [CF574710 2773651A6E238818 F5 73 6A 74 63 20 47 Ratio=50.00%
A27CB9EF], Secret Key=[885C3649 B840D9E0 Changed one bit (128)
06D061F5F6FC6046] and Cipher Text = Plain text = CF 57 47 10 27 73 65 1A 6E
[B218A58FA18EB4B764737D51 83378B4E]. 23 88 18 A2 7C B9 EE
Cipher Text= 8B B7 62 9E 65 73 58 29 38

www.ijera.com 70 | P a g e
Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

A0 2E C9 42 CE DB F9 Difference value = B4 88 CF FB 38 5F 9C FB
Difference value = 39 AF C7 11 C4 FD EC 9E 10 B6 CA 85 38 BD B0 8D
5C D3 53 98 C1 F9 50 B7 Ratio=53.12%
Ratio=54.68% Changed one bit (2)
Secret key = C8 5C 36 49 B8 40 D9 E0 06
Table 6. Results of Avalanche test due to change one D0 61 F5 F6 FC 60 46
bit in plain text in modified AES. Cipher Text = B3 3B 73 D4 77 9C F5 CB
Changed one bit (1) B5 0A DD FB 4F 27 70 47
Plain text = 4F 57 47 10 27 73 65 1A 6E Difference value = 01 23 D6 5B D6 12 41 7C
23 88 18 A2 7C B9 EF D1 79 A0 AA CC 10 FB 09
Cipher Text = A6 3B 8A BD B0 84 37 EA Ratio=44.53%
BF 7E F5 A4 1D 8F F9 0C Changed one bit (16)
Difference Value = 9E 8A C7 97 E6 05 18 F9 40 Secret key = 88 5D 36 49 B8 40 D9 E0 06
90 CD CD E7 2B 8E 4C Ratio=49.21% D0 61 F5 F6 FC 60 46
Changed one bit (15) Cipher Text = C7 43 E5 6C 4C D6 65 F1
Plain text = CF 55 47 10 27 73 65 1A 6E 60 31 D9 43 E5 0D 0E 65
23 88 18 A2 7C B9 EF Difference value = 75 5B 40 E3 ED 58 D1 46 04
Cipher Text = 5F EC 44 96 1B C5 F4 36 42 A4 12 66 3A 85 2B
52 F2 E4 72 5E 25 22 9B Ratio=42.96%
Difference Value = 67 5D 09 BC 4D 44 DB 25 Changed one bit (99)
AD 1C DC 1B A4 81 55 DB Secret key = 88 5C 36 49 B8 40 D9 E0 06
Ratio=50.00% D0 61 F5 D6 FC 60 46
Changed one bit (69) Cipher Text = CB 67 F2 A1 C9 4D B7 21
Plain text = CF 57 47 10 27 73 65 1A 66 C0 06 A1 4C DC D6 9B 05
23 88 18 A2 7C B9 EF Difference value = 79 7F 57 2E 68 C3 03 96 A4
Cipher Text = 3B E8 CF 4F 38 2C 25 26 75 DC 1D 5F E1 10 4B
C1 7E B7 B1 4C 9F 81 C7 Ratio=51.56%
Difference Value = 03 59 82 65 6E AD 0A 35 3E Changed one bit (128)
90 8F D8 B6 3B F6 87 Secret key = 88 5C 36 49 B8 40 D9 E0 06
Ratio=50.00% D0 61 F5 F6 FC 60 47
Changed one bit (115) Cipher Text = 5D F7 E1 35 B8 9D 17 77
Plain text = CF 57 47 10 27 73 65 1A 6E 68 17 02 CB 12 EF 61 33
23 88 18 A2 7C 99 EF Difference value = EF EF 44 BA 19 13 A3 C0
Cipher Text = 70 F7 F4 7C 5D AD FE E1 0C 64 7F 9A 91 D8 EA 7D
08 02 F4 13 1C 2E DE 62 Ratio=52.34%
Difference Value = 48 46 B9 56 0B 2C D1 F2 F7
EC CC 7A E6 8A A9 22 70.00%
Avalanche effect Ratio

60.00%
Ratio=50.00% 50.00%
Changed one bit (128) 40.00%
30.00%
Plain text = CF 57 47 10 27 73 65 1A 6E 20.00%
10.00%
23 88 18 A2 7C B9 EE 0.00%

Cipher Text = E6 58 E8 C3 91 53 46 B6
1
7

103
109
115
121
127
13
19
25
31
37
43
49
55
61
67
73
79
85
91
97

CA F4 A9 BC 6C A2 D8 56
Difference Value = DE E9 A5 E9 C7 D2 69 A5 128-bits
35 1A 91 D5 96 06 AF 16 Figure 5. Avalanche effects of AES standard due to
Ratio=52.34% change one bit in Secret Key.

Table 7. Samples of Avalanche test due to change one

bit in Secret Key of AES-128 standard.
Changed one bit (1)
Secret key = 08 5C 36 49 B8 40 D9 E0 06
D0 61 F5 F6 FC 60 46
Cipher Text = 06 90 6A 74 99 D1 28 4C 74
C5 B7 D4 BB 8A 3B C3

www.ijera.com 71 | P a g e
Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

Table 8. Samples of Avalanche test due to change one Table 9. Result Outputs of Modified MILENAGE
bit in Secret Key of Modified AES. Algorthim to derive a stronger Authentication Vector
Changed one bit (1) (AV) than Output of standard Milenage Algorthim
Secret key = 08 5C 36 49 B8 40 D9 E0 06 (Authentication Vector) in 3GPP. [16], [17].
D0 61 F5 F6 FC 60 46 Key 9E5944AE A94B8116 5C82FBF9
Cipher Text = 10 8D DA 81 C6 9C 0A F1 F32DB751
70 58 10 9A 79 04 08 07 RAND CE83DBC54AC0274A 157C17F8
Difference value = 28 3C 97 AB 90 1D 25 E2 8F D017BD6
B6 28 F3 83 A0 7F 47 Ratio = Dynamic 50DA9F6BE38BA65C49FEEC01FE
49.21% Key 2CCC87
Changed one bit (2) SQN 0B604A81 ECA8
Secret key = C8 5C 36 49 B8 40 D9 E0 06 AMF 9E09
D0 61 F5 F6 FC 60 46 OP 223014C5806694C0 07CA1EEE
F57F004F
Cipher Text = 7A 0F EA 22 23 2A 1B 89 94
8B 23 16 02 A8 6A 74 OPC 1B3D2E3E625213D9BC49DBC552
BCDE 4C
Difference value = 42 BE A7 08 75 AB 34 9A 6B
TEMP
65 1B 7F F8 0C 1D 34 Ratio =
C8B662E237F3E58D24E7B5A96D2
50.78%
C2C9F
Changed one bit (16)
OUT1 BA328BB5831B71029111AC8D2332
Secret key = 88 5D 36 49 B8 40 D9 E0 06 E862
D0 61 F5 F6 FC 60 46
OUT2 13DD09518BE63818E23EDE87AC
Cipher Text = 80 D1 05 BD 8B A4 40 40 6C 73F109
05 28 36 A2 8B 0C F1 OUT3 EF3523C87886D0637FD2B501D02
Difference value = B8 60 48 97 DD 25 6F 53 93 E3BA2
EB 10 5F 58 2F 7B B1 Ratio = OUT4 0FE930F07931B15882B88BEB4F4D
52.34% E654
Changed one bit (99) OUT5 93A92189A493A876B8F1801A9F8A
Secret key = 88 5C 36 49 B8 40 D9 E0 06 B8C3
D0 61 F5 D6 FC 60 46 F1(MAC- BA328BB5831B7102
Cipher Text = 4A 6C 51 DC 6D B9 99 13 69 A)
B3 72 43 21 7E C3 E6 F1*(MAC 9111AC8D2332E862
Difference value = 72 DD 1C F6 3B 38 B6 00 96 -S)
5D 4A 2A DB DA B4 A6 Ratio = F2(RES) E23EDE87AC73F109
51.56% F3(CK) EF3523C87886D0637FD2B501D02
Changed one bit (128) E3BA2
Secret key = 88 5C 36 49 B8 40 D9 E0 06 F4(IK) 0FE930F07931B15882B88BEB4F4D
D0 61 F5 F6 FC 60 47 E654
Cipher Text = 00 A6 7B 81 86 50 04 40 94 F5(AK) 13DD09518BE6
5E 64 9A 0F 57 18 03 F5*(AK) 93A92189A493
Difference value = 38 17 36 AB D0 D1 2B 53 6B AUTN 18BD43D0674E9E09BA328BB5831
B0 5C F3 F5 F3 6F 43 Ratio = B7102
54.68% CE83DBC54AC0274A157C17F80D
017BD6E23EDE87AC73F109EF352
70.00% AV 3C87886D063EF3523C87886D0637
Avalanche effect Ratio

60.00%
50.00% FD2B501D02E3BA20FE930F07931
40.00%
30.00%
B15882B88BEB4F4DE65418BD43D
20.00% 0674E9E09BA328BB5831B7102
10.00%
0.00%

V. DISCUSSION AND CONCLUSIONS

13
19
25
31
37
43
49
55
61
67
73
79
85
91
97
1
7

103
109
115
121
127

128 - bits
(i) The main weakness in Milenage, as stated by the
Cryptanalysts, is the use bit rotations and constant
Figure 6. Avalanche effects of Modified AES due to XORs in the middle part of the milenage specially if
change one bit in Secret Key. the kernel block cipher in milenage algorithm is
susceptible to differential cryptanalysis, then an
attacker is capable to proceed a variety of attacks on

www.ijera.com 72 | P a g e
 Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

milenage algorithm. An attacker cannot predict any Sussex PO19 8SQ, ISBN 0-470-84794-8,
useful information if the kernel block cipher in 2003.
milenage algorithm is a strong secure. [6] Dan Forsberg, Gunther Horn, Wolf-Dietrich
This paper modifies the standard Milenage Moeller and Valtteri Niemi.‖LTE security‖.
Authentication algorithm through the dynamic change United Kingdom: John Wiley & Sons Ltd,
of the kernel block cipher AES. For every the Atrium, Southern Gate, Chichester, West
Authentication process a new S-box will be generated Sussex, PO19 8SQ, 2013.
using a combination of received random sequence [7] 3GPP TS 33.102 V11.5.1 (2013-06)
number (RAND), stored Authentication key (Ki) and Technical Specification; Third Generation
PN sequence generator to rearrange the columns and Partnership Project; Technical Specification
rows of standard S-box in AES. Tests proved that the Group Services and System Aspects; 3G
modified AES is more secure than the standard one Security; Security architecture (Release 11).
due to its dynamic structure in addition to increasing [8] 3GPP TS 33.105 V11.0.0 (2012-09)
its immunity to linear and differential cryptanalysis as Technical Specification; Third Generation
shown by avalanche test results in table 12. Partnership Project; Technical Specification
Group Services and System Aspects; 3G
Table 10. Average value of avalanche tests for (plain Security; Cryptographic algorithm
text – Secret key) in AES and Modified AES. requirements (Release 11).
Input type of Type of Avalanche average [9] Stefan Pütz, Roland Schmitz, Tobias Martin
data algorithm value ―Security Mechanisms in UMTS‖ DBLP:
Plaintext Modified AES 50.15% journals/dud/PutzSM01, Vol.25, No.6, June
Plaintext AES 49.71% 2001.
Secret key Modified AES 49.86% [10] 3GPP TS 33.401 V12.9.0 (2013-09)
Secret key AES 49.84% Technical Specification; Third Generation
Partnership Project; Technical Specification
(ii) Execution time can be reduced as follows: Group Services and System Aspects; 3GPP
The implementation of the modified System Architecture Evolution
authentication algorithm required more operation than (SAE);Security architecture (Release 12).
the standard one due to the dynamic nature of its S- [11] Sebastian Banescu and Simona Posea
box. Using the PIC18F452 microcontroller, the ―Security of 3G and LTE‖. Faculty of
execution time of the modified algorithm can be Computer Science , Eindhoven University of
greatly decreased to about 50.333 ms ( instead of Technology.
500ms taken by the standard algorithm using IC card. [12] Mun, H., Han, K., & Kim, K. 3G-WLAN
[19] ). interworking: security analysis and new
authentication and key agreement based on
EAP-AKA. Wireless Telecommunications
REFERENCES Symposium. WTS2009 (pp. 18). IEEE.
[1] P. Kitsos*, N. Sklavos, O. Koufopavlou (2009).
―UMTS security: system architecture and [13] 3GPP TS 35.206 V11.0.0 (2012-09)
hardware implementation‖ in Wireless Technical Specification; 3G Security;
Communications and Mobile Computing.- Specification of the MILENAGE Algorithm
May 2007.-Issue (4):Vol. (7).-pp. 483-494. Set: An example algorithm set for the 3GPP
[2] Federal Information Processing Standards authentication and key generation functions
Publications (FIPS 197), "Advanced f1, f1*, f2, f3, f4, f5 and f5*; Document 2:
Encryption Standard (AES) ", 26 Nov. 2001. Algorithm Specification (Release 11).
[3] J. Daemen and V. Rijmen, The blocks cipher [14] khaled suwais and Azman samsudin ―New
Rijndael, Smart Card research and Classification of Existing Stream Ciphers‖
Applications, LNCS 1820, Springer-Verlag, INTECH Journal ,1 Feb.2010. [15] Shinsaku
pp. 288-296. Kiyomoto , Toshiaki Tanaka and Kouichi
Sakurai ―K2: A Stream Cipher Algorithm
[4] Reshma Nadaf and Veena Desai ―Hardware using Dynamic Feedback Control‖ Springer,
Implementation of Modified AES with Key Communications in Computer and
Dependent Dynamic S-Box‖ IEEE ICARET Information Science,, Vol.23, 2009, pp 214-
2012. 226.
[5] Valterri Niemi and Kaisa Nyberg ―UMTS [16] 3GPP TS 35.207 V11.0.0 (2012-09)
security‖. England: John Wiley & Sons Ltd, Technical Specification; 3G Security;
the Atrium, Southern Gate, Chichester, West Specification of the MILENAGE Algorithm

www.ijera.com 73 | P a g e
Zakaria Zakaria Hassan et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 6), March 2014, pp.67-74

Set: An example algorithm set for the 3GPP

authentication and key generation functions
f1, f1*, f2, f3, f4, f5 and f5*; Document 3:
Implementors' test Data (Release 11).
[17] 3GPP TS 35.208 V11.0.0 (2012-09)
Technical Specification; 3G Security;
Specification of the MILENAGE Algorithm
Set: An example algorithm set for the 3GPP
authentication and key generation functions
f1, f1*, f2, f3, f4, f5 and f5*; Document 4:
Design Conformance Test Data (Release
11).
[18] S3-010014 3GPP TSG SA WG3 Security
―Analysis of the Milenage Algorithm
Set.‖QUALCOMM International,
Gothenburg, Sweden, 27 February - 02
March, 2001.
[19] 3GPP TS 35.909 V10.0.0 (2011-03)
Technical Report; 3G Security;
Specification of the MILENAGE Algorithm
Set: Document 5: Summary and results of
design and evaluation (Release 10).

www.ijera.com 74 | P a g e