Transactions on Dependable and Secure Computing

Fast Generation of Pseudo-Random Cryptographic Keys

using Variable Speed String Rotations
Fo
Journal: Transactions on Dependable and Secure Computing

Manuscript ID: Draft

rP
Manuscript Type: Short Paper

Pseudo-Random Numbers, Crypto Keys, String Rotations,

Keywords:
Marsaglia’s MWC, Alpha-numeric string
ee
rR
ev
ie
w
On
ly

http://mc.manuscriptcentral.com/tdsc-cs
Page 1 of 5 Transactions on Dependable and Secure Computing

1
2
3
4 Fast Generation of Pseudo-Random Cryptographic Keys
5
6 using Variable Speed String Rotations
7
8
9
Suhas J Manangi, Parul Chaurasia, Mahendra Pratap Singh
10 [email protected], [email protected], [email protected]
11 Department of Computer Engineering, NITK-Surathkal, India
12
13
14 Abstract: own specific range of characters and numbers and
Fo
15 A new method for the fast generation of pseudo special characters, by producing a key using these
16 random cryptographic keys (variable size -any length ranges gives ease for us to remember the lengthy
17 keys without compromising the security features in
depending on the seed) with high degree of
18 producing random keys. This paper proposes a
randomness and high resistance to replication. The
rP
19 general technique to produce human friendly random
20 method uses a small random number that can be
key making use of existing fast random number
21 easily obtained by basic bit shifting. a initial key generators called “User Defined Pseudo Random Key
22 string is produced using all the components of seed Generator” (Ud-PRKG).
ee
23 considered like system time, processes information,
24 cursor position, CPU register information etc. using 2. Algorithm Specification: “Ud-PRKG”
25 the small random number produced earlier each bit
26 Following are considered for the algorithm to be
rR

in this initial key string is randomized using variable

27 presented in this paper. In this paper along with
28 speed string rotations.
having randomness, we can even specify the set of
29 characters for the random key generated to have to
30 Key words: Pseudo-Random Numbers, Crypto Keys,
make it easier for user to understand and memorise,
String Rotations, Marsaglia’s MWC, Alpha-numeric
ev

31 thus producing user friendly cryptographic keys.

32 string.
33 Computer clocks, or similar operating system or
34 1. Introduction hardware values, provide significantly fewer real bits
ie

35 of unpredictability than might appear from their

Cryptography provides security to data and specification, here in this paper we are considering
36
communication. However there is a potential flaw. time string obtained from system time just to
37
w

All cryptographic algorithms need generation of demonstrate how rotating string and a random
38
secret, un-guessable random numbers. The main number generator together can be clubbed to obtain a
39
reason for the need of good key generator is because pseudo random cryptographic keys.
40
“Humans are unable to generate and remember strong
On

41 secrets, and thus have difficulty managing

42 The proposed Ud-PRKG algorithm has 4 parts:
cryptographic keys” [1]. And another reason to
43 consider pseudo random keys than random keys is i. Seed
44 because of energy [2] and hardware requirements [3] ii. Random Number Generator
45
ly

for producing real random numbers thus keys. iii. Initial key String
46
iv. User Preference String
47 Lot of work and research is done in the area of
48 Random Number Generators, Random Key Following are considered for the algorithm to be
49 Generators, Pseudo- Random Keys and Numbers, presented in this paper.
50 Bio-metric Key generators, Hybrid RNG etc. Most of
51 the algorithms concentrate on cost/performance,  Seed: Seed can consist of system process
52 energy efficiency [2], randomness [3], seed length [4] information, clocks, memory and other system
53 and security [5]. But the main reason for the work in information, or even a user input etc. For
54 this paper is the need for human friendly random demonstration of algorithm we are considering
55 keys generator. time string containing month-day-time-year,
56 system processes information, CPU register
57 Human has tendency to remember more in their
58
59
60
http://mc.manuscriptcentral.com/tdsc-cs
 Transactions on Dependable and Secure Computing Page 2 of 5

1
2
3 values etc depending upon the length of the key For the above algorithm, the 2 variables needed m_z
4 needed. and m_w are obtained from system time
5 m_z = time difference between Sun Apr 7 02:25:53
6  Random number generator: Is required which 1974 and present time in seconds.
7 produces a small length random number. For m_w = present time in hh: min: sec format.
8 demonstration we are using Marsaglia's MWC [6] Thus a small random_number (r_n) is generated.
9 (multiply with carry) algorithm.
10
11  Initial Key String: Seeds are gathered together
12 to use it as initial key string which will be later
13 randomised. The length of this depends on the
14 length of key needed. In this paper for
Fo
15 demonstration we are using time string as it is
16 which accounts for 25 characters (200 bits).
17
18  User Preference String: User can decide the set
rP
19 of characters that makes up the cryptographic
20 key depending upon his/her ease of remembering
21 the keys with those set of characters. Here for
22 demonstration of algorithm we are considering
alpha-numeric string containing al characters and
ee
23
24 digits.
25
Seed System time string
26
rR

27 RNG Marsaglia's MWC

28
29 Initial Key String <day><month><date><hour: min: sec> <year>
30
“abcdefghijklmnopqurstuwxyz123456789
ev

31 User Preference
32 String 0ABCDEFGHIJKLMNOPQRSTUVWX
YZ”
33
34 Table 1: Summarising the parameters considered for demonstration
ie

35 of the proposed Algorithm

36
37 3. Marsaglia’s MWC
w

38
39 Marsaglia's MWC (multiply with carry) algorithm:
40
On

m_z 36969*(m_z & 65535) + (m_z >> 16)

41
42
m_w 18000*(m_w & 65535)+ (m_w >> 16) 4. String Rotations
random_number (m_z << 16) + m_w
43 Table 2: Marsaglia’s Parameters Alpha-Numeric String (User Preference String)
44
and Initial key string rotates at speeds m_z and m_w
45
ly

respectively as shown in Figures 4 and 5. At any

46
instance for given system time we get different set of
47 speeds.
48
49
50
51
52
53
54
55
56
57
58
59
60
http://mc.manuscriptcentral.com/tdsc-cs
Page 3 of 5 Transactions on Dependable and Secure Computing

1
2
3 E.g. of Initial key string: Mon Feb 09 02:25:53 2009 Alpha Numeric and
4 65*25 1550
Initial Key String
5 System Time 24*60*60 86400
6 Total Permutation 1550*86400 133920000
7 Table 3: Randomness in this demonstration
8
9 This is only to demonstrate the algorithm, to
10 increase the security and possibility of this algorithm
11 we can choose better seeds and complex ways of
12 fixing speeds of string rotations. To calculate the
13 randomness of the keys produced we have considered
14
a statistical approach of considering 100 keys
Fo
15
16 produced continuously by the implementation of this
5. Randomization of the Key
17 algorithm with the parameters chosen in this paper
18 r_n (random number) produced above using for demonstration. 100 keys are shown in the Fig. 6.
rP
19 Marsaglia's algorithm is added to ASCII values of
20 each character of Initial Key String thus injecting
21 randomness to each bit of the Key String.
22
Since the final key should have characters only
ee
23
24 from the Alpha-Numeric String (User Preference
25 String) given, each randomized character is
26 substituted by characters from Alpha-Numeric String.
rR

27 This is carried out as shown in Figure 6. 62 present in

28
this figure is because of the length of the User
29
30 Preference String considered here.
ev

31
For the alpha-numeric string considered the average
32
deviation of all its elements comes up to 26 and with
33
random key generator proposed in this paper, average
34
difference between corresponding key elements for
ie

35
first 100 key samples comes up to 18-20 which
36
shows the randomness of the key.
37
w

38
7. Application of the Ud-PRKG
39
40
This can be used in scenarios where a user needs
On

41
42 to set a very strong password or random key and
43 he/she is not able to frame one. Such as following
44 application:
45
ly

46 i. Online Web based File Sharing: - User can get

47 a random key for the file shared thus able to
48 Thus after randomizing each character in the pass the key to intended readers with ease of
49 remembering the key.
Initial Key String we get the Random Key String
50
51 with required length.
ii. Email Passwords: - Mail service can produce
52
53 6. Analysis of the Ud-PRKG Algorithm passwords from set of characters which are
54 sent when the new account is opened or for
55 Considering the date to be same, for the resetting the passwords.
56 parameters considered in this demonstration the
57 possibilities is summarised in the Table 3.
58
59
60
http://mc.manuscriptcentral.com/tdsc-cs
 Transactions on Dependable and Secure Computing Page 4 of 5

1
2
3
4
5
6
7
8
9
10
11
12
13
14
Fo
15
16
17
18
rP
19
20
21
22
ee
23
24
25
26
rR

27
28
29
30
ev

31
32
33
34
ie

35
36
37
w

38
39
40
On

41
42
43
44
45
ly

46
47
48
49
50
51
52
53
54
55
56
57
58 Figure 8: 100 Keys produced
59
60
http://mc.manuscriptcentral.com/tdsc-cs
Page 5 of 5 Transactions on Dependable and Secure Computing

1
2
3 iii. Folder Locks: - For keeping files in an 10. Conclusion
4
encrypted form with not just passwords but
5 The proposed Ud-PRKG algorithm in this paper is
6 random keys with ease to remember them.
of advantage when a set of characters are defined
7 such that Cryptographic Key can have elements from
8 8. Advantages only this set. This algorithm generates human
9 friendly keys with ease to remember them. This
10  This algorithm gives flexibility to choose the
algorithm is explained here for limited capabilities,
11 character set of the elements that can be
with a better choice in Initial RNG and seed,
12 present in the Cryptographic Keys. randomness of higher degree can be achievable.
13 Providing Human friendly keys thus ease of
14 remembering the keys. References:
Fo
15  This algorithm doesn’t involve complex
16 [1]. “Robust techniques for evaluating biometric
mathematical functions. So computation
17 cryptographic key generators” ISBN: 978-0-549-
18 power needed is very less and produces
58505-3 Author: Lucas Kevin Ballard
rP
19 random keys at very fast rates
[2]. “Designing cryptographic key generators with
20  Since seed can include a user input, and the low power consumption” ISBN: 0-7695-2500-8
21 character set is defined by the user which Author: M PUCZKO and V. N. YARMOLIK
22 can be any combination of available [3]. “An Enhanced Hardware-based Hybrid Random
ee
23 characters, the recovering the key by Number Generator for Cryptosystem”
24 ISBN:978-0-7695-3595-1 Author: N. M.
guessing the seed values becomes more
25 Thamrin, G. Witjaksono, A. Nuruddin, M. S.
26 complex.
Abdullah
rR

27  The above demonstration included very [4]. “Reducing The Seed Length In The Nisan-
28 small seed for which the permutation needed Wigderson Generator” ISSN:0209-9683 Author:
29 to recover the key is 133920000. This Russell Impagliazzo, Ronen Shaltiel, Avi
30 tremendously increases with bigger seed and Wigderson
ev

31 complex functions of calculating speeds of [5]. "Bayesian estimates in cryptography" ISBN:960-

32 8457-29-7 Author: George Stephanides
string rotations.
33 [6]. Marsaglia's MWC (multiply with carry)
34 algorithm by George Marsaglia, Professor
ie

35 9. Disadvantages and Future Works: Emeritus Department of Statistics Florida State

36 University
37  This paper tries to explain the randomness
w

38 of this algorithm using statistical data, and

39 random numbers can’t be completely
40 explained by statistical data so better ways
of explaining the randomness has to be
On

41
proposed.
42
43  This algorithm needs a small random
44 number generator without which the security
45 of the algorithm decreases so choosing a
ly

strong random number generator becomes

46
essential.
47
48  Since the security and randomness provided
49 by this algorithm is driven by the input
string given by user it can’t be widely used
50
in network applications which run in the
51
backend and need no user interference like
52
SSL etc.
53
54
55
56
57
58
59
60
http://mc.manuscriptcentral.com/tdsc-cs