Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

Modular Square One-Way Function: Analyzing

the Algorithm for Randomness, Regularity
Schematic (Codec System) and Vector
Normalization
Ahmed Mohammed Al-Fahdi1
1
Sultanate of Oman, MSc Communication Engineering, University of Birmingham

Publication Date: 2025/02/22

Abstract: Actually, The proposed paper was built as one part working through analyzing one-way modular square which is
a special case of modular exponentiation with exponent of 2, counter form of quadratic residue, for encryption and PKI
implementation. Later on, Due to the amount of data regarding this analysis, it seems be better to separate it to three parts.
This paper (Part-1) aims to analyze a modular square one-way function using integer factorization (IF) or approximation
methods. It is supposed that such function has random characteristics. This analysis considers a new innovative idea, original
as claimed, focusing in notable regularities that could be used as a trapdoor for practical applications. Such as random
generator, new codec system and 3D vector normalization. At the end, different types of quadratic residue algorithms and
square root will be considered.

Keywords: Square Root Algorithm, Modular Square, Modular Arithmetic, Quadratic Nonresidue Square Root, One-Way Function,
Newton-Raphson Method, Exponentiation By Squaring, Jacobi Prime, Legendre Symbol, Fast Inverse Square Root (Invsqrt), Vector
Normalization.

How to Cite: Ahmed Mohammed Al-Fahdi. (2025). Modular Square One-Way Function: Analyzing the Algorithm for Randomness,
Regularity Schematic (Codec System) and Vector Normalization. International Journal of Innovative Science
and Research Technology, 10(1), 79-87. https://doi.org/10.5281/zenodo.14908867.

I. INTRODUCTION Such arithmetic is similar to 12-hour clock, in which the day

is divided into two 12-hour periods. In general, according to
Since the far past square roots was fascinating task and studies and modular graphs of arithmetic function, the
this was utilitarian by the area computations. Through over modular for multiplication look more diverse than addition.
centuries studies in such roots add a lot to the arithmetic As a special case the modular of the modular exponent the
science. These days, it is realized that not every natural modular square look even more random and have no apparent
number is a perfect square. Such square roots are nonintegral symmetry. It is surprised that the modular graphs for higher
and even irrational. As squaring an integer is number result powers look more regular again. [3]
from multiplying the integer by itself. In the other hand, the
square root of an integer could result in irrational number. II. METHODOLOGY AND PAPER
Approximating such number to an integer result in a ORGANIZATION
remainder. Such reminder is called modular square or
quadratic residue or nonresidue. In addition to all that, A. Overview
calculating a square root of large numbers it will be tedious The overall paper take analyzing the one-way square
using integer factorization (IF). In such case there is a need modular algorithm for encryption and PKI implementation.
of square root algorithm. This algorithm exploits the Due to the amount of data and type of the information we
underlying algebra of the decimal place of the floating value. decide to divide the whole content to three parts. First part
[1] [2] (PART-1), analyze the square modular function in term of
arithmetic modular using integer factorization and Newton
𝑦 = 𝑓(𝑥) = 𝑥 2 (𝑚𝑜𝑑 𝑛) (1) Raphson approximation. Such mathematics allow us to
analyze the function in term of randomness. Also, it will
Usually the modular square function (1) characterize a allow us to figure out if there is any regularity schematic that
randomness. It is type of modular arithmetic where integers could be exploit as new codec system! In addition, exploring
wrap around when reaching certain value called modulus. the practical implementation of the square modular and fast

IJISRT25JAN1916 www.ijisrt.com 2770

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

inverse square root algorithm allow to implement it in vector integer factorization (IF), discrete logarithmic (DL) and
normalization. [4][5] quadratic nonresidue (QR) leads to focus in the integer
factorization (IF) as simple start point. As a result of quadratic
B. Modular Square Function nonresidue Integers, the resulting arithmetic modular
In mathematics, there are only few functions assumed supposed to be irrational number, for noncomplete squares.
(conjectured) to be one-way. None of these functions is For such calculation, in this methodology, a way of numerical
proven to be one-way. It is theoretically not even known the analysis will be used called Newton’s approximation method
proof of existence of such functions. Modular square function which is also known as Newton-Raphson method. This way
(1) is supposed to be a one-way function. It is also called will help to find the roots or zeros with a real-valued function
square function. [1] 𝑓 and its derivative 𝑓′.For illustration, in geometrically
respective this allow to approximate the zeros using tangent
𝑦 = 𝑓(𝑥) = 𝑥 2 (𝑚𝑜𝑑 𝑛) of the function graph (Figure-1). The execution of a linear
approximation is in repeated process with a recursive
Analyzing the nonresidue quadratic root, for non- partitioning procedure. [6]
perfect squares, of this function using techniques such as

Fig 1: Newton Raphson Approximation Illustration as Intersect of the tangent /geeksforgeeks.org/newton-raphson-method/

C. Mathematics Such calculation will be done manually till a sufficiently

The mathematic for Newton-Raphson approximation is precise value is reached. This will allow us to approximate
as follow the root as well with the modular or remainder. [10]

𝑓(𝑥𝑖 ) Later on, we could do such calculations using the

𝑋𝑖+1 = 𝑋𝑛 − (2) calculator and proceed with the following procedure for the
𝑓 ′ (𝑥𝑖 )
remainder of the squares to analyze such values. Other
So, for modular square function parameters could be calculated for deeper view of numerical
theory analysis with such function.
𝑓(𝑥𝑖 ) = 𝑥 2 − 𝑥𝑖 (3)

𝑓 ′(𝑥𝑖 ) = 2𝑥 (4)

IJISRT25JAN1916 www.ijisrt.com 2771

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

From above, the square root of non-perfect square is So,,,

irrational number. Hence, the integer part could be
approximated to the nearest integer using the usual way of 𝑥02 − 𝑥0
threshold (0.5). 𝑥1 = 𝑥0 −
2𝑥0
Furthermore, approximating the rational part, rational 𝑥0 − 1
decimal, needs to evaluate thresholds using (1-3) functions as 𝑥1 = 𝑥0 −
2
follow:
𝑥0 + 1
𝑓(𝑥𝑖 ) 𝑥1 = (5)
𝑥𝑖+1 = 𝑥𝑛 − ′ 2
𝑓 (𝑥𝑖 )
But for rational decimals −1 ≤ 𝑥0 ≤ 1
Where
So
𝑓(𝑥𝑖 ) = 𝑥 2 − 𝑥𝑖
0 𝑓𝑜𝑟 𝑥0 = −1
′(𝑥𝑖 ) 1
𝑓 = 2𝑥 𝑥1 = { 2 𝑓𝑜𝑟 𝑥0 = 0 (6)
1 𝑓𝑜𝑟 𝑥0 = 1
Approximating 𝑥1 from 𝑥0 leads to
This means that the resulting rational decimal calculated
𝑥 2 − 𝑥0 from the approximation process need to be approximated to
𝑥1 = 𝑥0 −
2𝑥 the nearest value as what doing with the integers. For
quadratic nonresidue irrational number, there are infinite
But in this Newton iteration rational numbers of the rational part between 0 and 1. These
irrational parts follow a gaussian distribution with mean value
𝑥 2 ≈ 𝑥02 & 𝑥 ≈ 𝑥0 0.5, because it is the only exact finite number in this scale, or
either sides 0 and 1 using threshold of 0.25 or 0.75. This could
be illustrated in figure (2).

Fig 2: Rational Part Approximation

IJISRT25JAN1916 www.ijisrt.com 2772

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

 Notice-1: The Methodology trick used here is that from III. RESULTS
above functions it is easy to calculate the square modular
(r) approximated to the closest integer using rational part The following table summarize the calculations that
approximation with threshold of 0.5. The same thing is have been done in the above section resulting from equations
done with the rational part of the root modular (n) (1-6) where:
approximated to the ratio of 0.5 using threshold of ¼ and
¾.  𝑥 2 :is the square, non-perfect squares chosed randomly
 𝑥: is the square root
 𝑟: is the square remainder approximated to integer (𝑥 2 −
2
𝑥𝑎𝑝𝑝 )
 𝑛 :is the ratio, from rational par, of the root remainder
 (𝑥𝑒𝑥𝑎𝑐𝑡 − 𝑥𝑎𝑝𝑝 ) × 𝑥𝑎𝑝𝑝
𝑛
 : is the notified correlation between 𝑛 and 𝑟
𝑟

Table 1: Summarizing Results of the Methodology Section

Num 𝒙𝟐 𝒙𝒂𝒑𝒑 𝒓 Exact approx 𝒏
𝒏 𝒏 𝒓
1 612 25 -13 6.5341 6.5 -0.5
2 570 24 -6 3.0078 3 -0.5
3 811 28 27 13.3857 13.5 0.5
4 550 23 21 10.3978 10.5 0.5
5 1011 32 -13 6.5207 6.5 -0.5
6 1111 33 22 10.9449 11 0.5
7 2022 45 -3 1.5005 1.5 -0.5
8 41 6 5 2.4187 2.5 0.5
9 59 8 -5 2.5508 2.5 -0.5
10 61 8 -3 1.5180 1.5 -0.5
11 103 10 3 1.4889 1.5 0.5
12 2131 46 27 7.4867 13.5 0.5
13 2311 48 15 3.4973 7.5 0.5
14 2777 53 7 16.0458 3.5 0.5
15 3119 56 -17 8.5115 8.5 -0.5
16 2130 46 14 6.9884 7 0.5
17 2312 48 -8 3.9965 4 -0.5
18 2778 53 -31 15.5430 15.5 -0.5
19 3118 56 -18 9.0129 9 -0.5

 Notice-2: From number analysis, it is clear that the square After the suggested approximations, the ratio of the root
number residue (r) could be directly extracted as double modular to the square modular has fixed magnitude of (0.5)
of the root modular ratio (n) without the traditional with alternating sign as shown in the below figure (3).
subtraction method.

Fig 3: Root Modular ratio to the Square Modular with Fixed Magnitude at (0.5)

IJISRT25JAN1916 www.ijisrt.com 2773

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

IV. RESULTS ANALYZING similar to the modular square function (1) which is a special
case of the modular exponent function as follow:
Form above methodology and results a case study has
been considered above for modular square one-way function  𝑄𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐 𝑟𝑒𝑠𝑖𝑑𝑢𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
using equation (7):
𝑥 2 = 𝑞(𝑚𝑜𝑑 𝑛) (11)
2 2
𝑟=𝑥 − 𝑥𝑎𝑝𝑝 (7)
 𝑀𝑜𝑑𝑢𝑙𝑎𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
Analyzing the relation of such residue to the ratio of the
rational part of the root using equation (8): 𝑞 = 𝑥 2 (𝑚𝑜𝑑 𝑛)

𝑛 = (𝑥 − 𝑥𝑎𝑝𝑝 ) × 𝑥𝑎𝑝𝑝 (8) 𝑎

𝐿𝑒𝑔𝑒𝑛𝑑𝑟𝑒 ( ) =
𝑏
−1 𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐 𝑟𝑒𝑠𝑖𝑑𝑢𝑒 𝑚𝑜𝑑𝑢𝑙𝑜 𝑝
Going through the methodology above with the
{0 𝑎 ≡ 0(𝑚𝑜𝑑 𝑝) (12)
explained trick by calculating the residue of the rational root
1 𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐 𝑛𝑜𝑛𝑟𝑒𝑠𝑖𝑑𝑢𝑒 𝑚𝑜𝑑𝑢𝑙𝑜 𝑝
itself approximated to ½ ratio using thresholds of ¾ and ¼.
Such characteristic could be examined in another paper. It is
clear that:
That is, in the methodology above, the modular square
function (1) has been studied for the approximated integer
𝑟 =2×𝑛 (9)
root so that:
Or
𝑟 = 𝑥 2 (𝑚𝑜𝑑 𝑥𝑎𝑝𝑝 )
𝑟
𝑛= (10) Inspired by the Legendre symbol (12) and Jacobi symbol with
2 𝑛
the notified iteration of ( ) in table (1), it is helpful to
𝑟
Where: represent the ratio of the rational root modular to the square
 r: is the modular square relative to the square of the 𝑛
modular ( ) by a normalized symbol, lets represent it as
approximated root 𝑟
modular factor (M), with values of -1,0 and 1 as follow:
 n: is the rational ratio of the modular root relative to the
approximated root 𝑛
−1 𝑓𝑜𝑟 = −0.5
𝑟
 Notice: It is clear that 𝑛 work as a trapdoor for easy 𝑛 𝑛
calculation of the remainder 𝑟. 𝑀( ) = 0 𝑓𝑜𝑟 = 0 (13)
𝑟 𝑟
𝑛
For the results and functions analyzed above and {1 𝑓𝑜𝑟 = 0.5
𝑟
referring to the Legendre symbol that is used for the quadratic
residue as a quadratic character modulo of an odd prime As a result, this symbol could be examined for squares
number. In somehow, the quadratic residue function (11) is scale, chosen randomly [17-48] in the table (2) below

Table 2: Normalized modular ratio (Symbol M)

Num 𝟐 r n M (normalized residue)
Integers (𝒙 )
1 17 1 0.5 1
2 18 2 1 1
3 19 3 1.5 1
4 20 4 2 1
5 21 -4 2 -1
6 22 -3 1.5 -1
7 23 -2 1 -1
8 24 -1 .5 -1
9 25 0 0 0
10 26 1 .5 1
1111 27 2 1 1
12 28 3 1.5 1
13 29 4 2 1
14 30 5 2.5 1
15 31 -5 2.5 -1
16 32 -4 2 -1
17 33 -3 1.5 -1
18 34 -2 1 -1

IJISRT25JAN1916 www.ijisrt.com 2774

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

19 35 -1 .5 -1
20 36 0 0 0
21 37 1 .5 1
22 38 2 1 1
23 39 3 1.5 1
24 40 4 2 1
25 41 5 2.5 1
26 42 6 3 1
27 43 6 3 -1
28 44 5 2.5 -1
29 45 4 2 -1
30 46 3 1.5 -1
31 47 2 1 -1
32 48 1 .5 -1

From the table (2) above, although the known factor(M).Such regularity could be illustrated by plotting
randomness property of the square modular function, there is graphs (figure 4-5) below:
a striking regularity represented by the normalized modular

Fig 4: Line Chart of the Normalized Modular Factor (M) - Equation (13)

Fig 5: Column Chart of the Normalized Modular Factor (M) - Equation (13)

IJISRT25JAN1916 www.ijisrt.com 2775

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

From figure (4) above, it is notified that the pulse width direct, it is applicable for many applications such as codec,
is equal to the root of the lowest complete square. In addition cryptography and 3D graphic games. These are high tech
to that, the symbol of normalizing the ratio of the square fields in term of signal processing.
modular to the rational root modular could be simplified by
the follow definition: According to the results analyzing (part4) and the
previous background in signal processing, the analyzed
𝑀 characteristics of the studied algorithm is useful for the
1 𝑓𝑜𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛 𝑙𝑜𝑤𝑒𝑟 𝑐𝑜𝑚𝑝𝑙𝑒𝑡𝑒 𝑠𝑞𝑢𝑎𝑟𝑒 following fields:
={ 0 𝑓𝑜𝑟 𝑎𝑛𝑦 𝑐𝑜𝑚𝑝𝑙𝑒𝑡𝑒 𝑠𝑞𝑢𝑎𝑟𝑒
−1 𝑓𝑜𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛 𝑎𝑏𝑜𝑣𝑒 𝑐𝑜𝑚𝑝𝑙𝑒𝑡𝑒 𝑠𝑞𝑢𝑎𝑟𝑒 A. Randomness System Applications:
A.1 Although the randomness of the square modular, it
Analyzing figure (5), warped around any complete is notable that there is a striking regularity for the ratio of the
square there are equal a count of positive and negative rational root modular to the square modular and this
symbols equal to the root of that complete square. regularity schematic increase for the higher perfect squares as
shown in figure (5). This in turn, represent that the perfect
Notice: Implementing the methodology trick to squares space away for higher values. Considering the prime-
exponent higher than square, 3 ,4,5 does not work, need counting function 𝜋(𝑛) shown in the figure (6), It is
𝑟
further examine. That means such ( ) trapdoor does not work supersizing to figure out that there is a direct relationship
𝑛
with the exponential modular one-way algorithm higher than related to the growth rate of (M). This is something related to
2 as further it has been trayed. This enhances that the modular the prime number theorem. Furthermore, the prime number
square function as a special case of the modular exponent density (equation-14), has normal or gaussian probability
function. which in turn reflected to the rational part of the irrational
roots and hence the modular factor (M) around any perfect
V. IMPLEMENTATION AND PRACTICAL root shown in figure (5). [8]*
APPLICATIONS
𝜋(𝑥)
lim =1 (14)
𝑥→∞ 𝑥/log(𝑥)
Although it seems that the analysis of the above
algorithm done in the results analyzing section (4) simple and

Fig 6: Values of 𝜋(𝑛) for the first 60 positive integers [8]

B. Codec System Applications:  According to the above characteristics and the

characteristic of alternating sign with the Modular factor
 As shown above, the prime number density function 𝜋(𝑛) M , such algorithm could be used as practical application
follow similar characteristic to the density of modular for Hilbert envelop transform with special case of the
factor (M). Furthermore, it is known that from prime Riemann-Hilbert problem. [8]
number theorem such function follows Gaussian  Considering the figure (6), there is window schematic
probability and it is applied to what is called Multi- algorithm with scalable interval and zeros at perfect
Dimensional Gaussian classifier and Multivariate normal squares. Such Windows with gaussian characteristic is
distribution. This is an advantage for audio and video applicable for filters to remove the airy disks caused by
CODEC diffraction. This window with zero skewness at the

IJISRT25JAN1916 www.ijisrt.com 2776

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

perfect squares could be used also for smoothing the science in term of speed, it was easy practical way for
signal in similar way to Hann function. modular manual calculations analysis. From such analysis a
 Using such codec for compression it is assumed to be normalized residue symbol M, similar Legendre, was
lossless compression which allows the data for further proposed as a trapdoor for the square modular calculation. By
processing including encryption or hashing. trial and error, such trapdoor does not work with higher
 Such regularity characteristics applied for PCM from modular exponentiation. The characteristics shown by M
analogue to digital modulation such as that with a-law and propose different practical applications in cryptography,
𝜇-law with manipulated property of sampling rate and the codec systems and vectors normalization.
bit depth to represent each sample. In such coding the
complete square could be synchronized to Nyquist Newton approximation method is avoided in computer
frequency or (or folding frequency). Such characteristic science as it contains slow division operation. In contrast to
allows to have free of aliasing distortion samples. This that, for computation purposes, fast inverse square root was
could be clarified by transferring the time-based analyzed. Such method depends on a bit fiddling technique
algorithm to frequency based using Laplace by shifting rational part decimal followed by Newton
transformation. [9] iteration. There are many other algorithms could be used for
square modular such as Tonelli-Shanks and Cipolla’s
C. Vector Normalizing applications: algorithms. They were for quadratic residue (equation-11)
Vector normalizing is done by scaling the vector to the similar to square modular (equation-1)
magnitude such that:
The square root function analysis may play pivotal role
𝑉 = 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 (15) in future with quantum computations. This is applied for the
quantum circuit gates and the quantum algorithms. This paper
2
Magnitude =|𝑉| = √𝑎2 + 𝑏 2 + 𝑐 2 (part-1), presented as a foundation of other two papers (part-
2 & part-3) dealing with cryptography and mathematical
𝑉 𝑎𝑥+𝑏𝑦+𝑐𝑧 conjectures.
Then normalized vector 𝑣 = |𝑉| = 2
√𝑎2 +𝑏 2 +𝑐 2
REFERENCES
That means multiplying the vector by the inverse square
1 [1]. Prof . Dr. Rolf Oppliger,Contemporary to
root ( 2 2 2 2 )
√𝑎 +𝑏 +𝑐 Cryptography,Infogaurd,Bar Switzerland ,2013
[2]. W. Diffie and M. Hellman”New Direction in
As a result of above, the square root algorithm in inverse Cryptography”,IEEE Trans, 1976
form is used for vector normalization. There is a c-code [3]. Center for education in mathematics and computing,
algorithm (Appendix [1]) called Fast Inverse Square Root Grade 6 Math Circles -Modular Arithmetic,University
algorithm referred as Fast InvSqrt (). It depends in the bit- of Waterloo,2020
fiddling techniques followed by Newton iterations, as it has [4]. Catalano,Fiore,Rosario,Vamvourellis,Algebraic
been done above in the methodology section. It was used by (Trapdoor) One-Way Functions and their
3dfx Interactive. In 2005, it become famous algorithm in the Applications, University of Catania
world of games as it is used by Id Software. This is because [5]. Evan Huynh,Rabin’s Cryptosystem, Department of
it was mysterious algorithm for the programmers as it Mathematics, Linnaeus University, Sweden 2021
depends on bits hacking playing with the rational part [6]. Anca-Maria Nica ,Quadratic Residues and
decimal, something similar to what have been done above in Applications in Cryptography, Alexandru Ioan Cuza
the methodology. The magic of such popularity was with the Universithy of Iasi, 2020
Magic number (0𝑥5𝐹3759𝐷𝐹) that use to minimize the [7]. Pranav Gokhale ,Implementation of Square Root
approximation error and not known precisely how it was Function Using Quantum Circuits, Princton
determined. It was derived through trial and error. University Class, 2015.
[8]. F.B. Roodenburg,Riemann’s Explicit Formula and the
Because of its speed the algorithm was used in graphics Prime Number Theorem, Delft Institute of Applied
of video games to normalize the surface for lightening and Mathmatics, 2020
shading calculations. Such algorithm can work as a hint to use [9]. Dushyant,Patrick,Nikolay, Non Intrusive Codec
the above result analysis for vector normalization. [10] Identification Algorithm ,Marsh 2012
[10]. Andrei Seymour-Howell,Fast inverse square-root
VI. CONCLUSION program,2021
In conclusion, throughout this paper one-way modular
square function was analyzed. This function is a special case
of modular exponentiation, which built the infrastructure of
public-key cryptography. Square root function was first
milestone in such analyzation. Integer factorization method
was used throughout Newton Raphson approximation
method. Although such method is not practical for computer

IJISRT25JAN1916 www.ijisrt.com 2777

Volume 10, Issue 1, January – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14908867

APPENDIX

Fast Inverse Square Root [10]

 Quote…
“If you couldn’t can’t explain it simply, you don’t understand it well enough “Albert Einstein…

IJISRT25JAN1916 www.ijisrt.com 2778