Added actual files to repo.

README

@@ -0,0 +1,17 @@
+# why-RSA-works/README
+# Copyright 2012 Peter Franusic.
+# All rights reserved.
+#
+
+This directory contains LaTeX source files for an article titled "Why RSA Works".
+The article describes the math behind the RSA algorithm.
+The makefile script assumes that the pdflatex program is installed.
+
+To generate a PDF of the article:
+
+  $ make why-RSA-works.pdf
+
+To generate a tarfile of the sources:
+
+  $ make why-RSA-works.tar
+

block-diagram.tex

@@ -0,0 +1,85 @@
+%%%% why-RSA-works/block-diagram.tex
+%%%% Copyright 2012 Peter Franusic.
+%%%% All rights reserved.
+%%%%
+%%
+%%     2         3         4         5         6         7
+%% 6789012345678901234567890123456789012345678901234567890123456
+%%             Alice's                Alice's
+%%            public key            private key
+%%           {---------}            {---------}
+%%            n       e              n       d
+%%            |       |              |       |
+%%            |       |              |       |
+%%            |       |              |       |
+%%         +-------------+        +-------------+
+%%    m    |             |   c    |             |   y
+%% --------|    m # x    |--------|    m # x'   |--------
+%%         |             |        |             |
+%%         +-------------+        +-------------+
+%%
+%%                    An RSA cryptosystem
+
+% graphic macro definitions
+
+\setlength{\unitlength}{0.05in} % for pictures
+\newsavebox{\bigblock}
+\savebox{\bigblock}(16,12)[bl]{
+  \put( 0,  0){\line(1,0){16}}
+  \put( 0, 12){\line(1,0){16}}
+  \put( 0,  0){\line(0,1){12}}
+  \put(16,  0){\line(0,1){12}}}
+\newsavebox{\smallblock}
+\savebox{\smallblock}(9,12)[bl]{
+  \put( 0,  0){\line(1,0){ 9}}
+  \put( 0, 12){\line(1,0){ 9}}
+  \put( 0,  0){\line(0,1){12}}
+  \put( 9,  0){\line(0,1){12}}}
+
+% The block diagram
+\begin{picture}(90,45)(0,0)
+% Box around picture
+%\put(  0.0,  0.0){\line(1,0){90}}
+%\put(  0.0, 42.0){\line(1,0){90}}
+%\put(  0.0,  0.0){\line(0,1){42}}
+%\put( 90.0,  0.0){\line(0,1){42}}
+% Transmitter
+\put( 18.0, 37.0){\textsf{Alice's}}
+\put( 16.2, 34.0){\textsf{public key}}
+\put( 16.8, 28.0){$\overbrace{\phantom{XXXX}}$}
+\put( 14.0, 11.0){\usebox{\bigblock}}
+\put( 18.0, 16.5){\large{\texttt{modex}}}
+\put(  8.0, 18.0){$m$}
+\put(  6.0, 17.0){\vector(1,0){8}}
+\put( 17.3, 27.0){$n$}
+\put( 18.0, 26.0){\vector(0,-1){3}}
+\put( 25.3, 27.0){$e$}
+\put( 26.0, 26.0){\vector(0,-1){3}}
+\put( 33.5, 18.0){$c$}
+\put( 15.5,  7.0){\textsf{transmitter}}
+\put( 18.8,  4.0){\textsf{(Bob)}}
+% Channel
+\put( 30.0, 17.0){\vector(1,0){28}}
+\put( 39.0, 37.0){\textsf{insecure}}
+\put( 39.0, 34.0){\textsf{channel}}
+\put( 36.4, 28.0){$\overbrace{\phantom{XXXXXX}}$}
+\put( 44.0, 17.0){\circle{2}}
+\put( 44.0, 16.0){\line(0,-1){5.5}}
+\put( 40.5,  7.0){\textsf{sniffer}}
+\put( 41.0,  4.0){\textsf{(Eve)}}
+% Receiver
+\put( 62.0, 37.0){\textsf{Alice's}}
+\put( 59.7, 34.0){\textsf{private key}}
+\put( 60.8, 28.0){$\overbrace{\phantom{XXXX}}$}
+\put( 58.0, 11.0){\usebox{\bigblock}}
+\put( 62.0, 16.5){\large{\texttt{modex}}}
+\put( 53.0, 18.0){$c$}
+\put( 61.3, 27.0){$n$}
+\put( 62.0, 26.0){\vector(0,-1){3}}
+\put( 69.3, 27.0){$d$}
+\put( 70.0, 26.0){\vector(0,-1){3}}
+\put( 76.5, 18.0){$y$}
+\put( 74.0, 17.0){\vector(1,0){8}}
+\put( 62.0,  7.0){\textsf{receiver}}
+\put( 62.5,  4.0){\textsf{(Alice)}}
+\end{picture}

conclusions.tex

@@ -0,0 +1,30 @@
+%%%% why-RSA-works/conclusions.tex
+%%%% Copyright 2012 Peter Franusic.
+%%%% All rights reserved.
+%%%%
+
+So why does RSA work?
+Why is it that we can take some message $m$,
+run it through two modex operations, and come out with the same $m$?
+There are several reasons.
+First of all, RSA computations are done in a commutative ring
+and the multiplicative association property holds in commutative rings.
+This property tells us that 
+the two exponentiations $(m^e)^d$ are the same as the one exponentiation $m^{ed}$.
+
+A second reason is that exponents $e$ and $d$ are chosen
+such that they satisfy the multiples-plus-one condition $ed = k\lambda + 1$.
+This insures that $ed$ is one of the identity columns
+in the exponential table of ring $\mathcal{R}_n$.
+
+A third reason is that the exponential table contains 
+repeating blocks of columns where $m^a=m^{k\lambda+a}$.
+This is the wallpaper pattern that we saw in Table \ref{modex-33}.
+This pattern is the reason for the multiples-plus-one condition.
+
+Finally, RSA works because it relies on the intractability of the factoring problem.
+A huge RSA modulus $n$ cannot be factored expeditiously.
+Given that $n$ is the product of two distinct huge random primes,
+it is virtually impossible to factor $n$ in any reasonable amount of time,
+even if the factoring effort is distributed across thousands of computers.
+

exponent-arithmetic.tex

@@ -0,0 +1,56 @@
+%%%% why-RSA-works/exponent-arithmetic.tex
+%%%% Copyright 2012 Peter Franusic.
+%%%% All rights reserved.
+%%%%
+
+RSA uses exponential notation in the ring $\mathcal{R}_n$.
+Exponential notation is simply a mathematical shorthand for writing 
+a series of multiplications using the $\otimes$ operator.
+The multiplicative association property allows us to derive 
+two rules for doing arithmetic with exponents.
+
+Consider the set of three equations below.
+The left side of the first equation is the expression $m^2 \otimes m^3$.
+We can replace the $m^2$ with $(m \otimes m)$.
+We can also replace the $m^3$ with $(m \otimes m \otimes m)$.
+The right side of the first equation shows this.
+The multiplicative association property says that we can 
+ignore the parentheses and simply count the number of $m$'s that are multiplied.
+There are 5 and we show this in the second equation.
+Note that 5 is the sum of 2 plus 3.
+So instead of expanding the expression $m^2 \otimes m^3$
+we can simply add 2 and 3, as shown in the third equation.
+\begin{eqnarray*}
+  m^2 \otimes m^3 &=&  (m \otimes m) \otimes (m \otimes m \otimes m)  \\
+  &=&  m^5  \\
+  &=&  m^{2 + 3}
+\end{eqnarray*}
+
+\paragraph{Exponent addition rule:}  In general, when we have an expression of the form 
+$m^e \otimes m^d$ in the ring $\mathcal{R}_n$ we can simply add the exponents.
+\[  m^e \otimes m^d  =  m^{e + d}  \]
+
+Consider the set of four equations below.
+The left side of the first equation is the expression $(m^2)^3$.
+This means three copies of $m^2$ are multiplied using the $\otimes$ operator.
+The right side of the first equation shows this.
+In the second equation, we replace each $m^2$ with $(m \otimes m)$.
+The multiplicative association property says that we can 
+ignore the parentheses and simply count the number of $m$'s that are multiplied.
+There are 6 and we show this in the third equation.
+Note that 6 is the product of 2 times 3.
+Instead of expanding the expression $(m^2)^3$
+we can simply multiply 2 and 3, as shown in the fourth equation.
+\begin{eqnarray*}
+  (m^2)^3  &=&  m^2 \otimes m^2 \otimes m^2  \\
+  &=&  (m \otimes m) \otimes (m \otimes m) \otimes (m \otimes m)  \\
+  &=&  m^6 \\
+  &=&  m^{2 \times 3}
+\end{eqnarray*}
+
+\paragraph{Exponent multiplication rule:}  In general, when we have an expression of the form 
+$(m^e)^d$ in the ring $\mathcal{R}_n$ we can simply multiply the exponents.
+\begin{equation} \label{eq:expo-mult}
+  (m^e)^d  =  m^{ed} 
+\end{equation}
+

exponential-notation.tex

@@ -0,0 +1,49 @@
+%%%% why-RSA-works/exponential-notation.tex
+%%%% Copyright 2012 Peter Franusic.
+%%%% All rights reserved.
+%%%%
+
+Let's say we're given three elements $a,b,c$ which are members of the set $Z_n$.
+We're also given the expression $a \otimes b \otimes c$.
+The question is this:  How do we compute this expression?
+Do we first multiply $a$ and $b$ and then multiply $c$?
+Or do we multiply $b$ and $c$ and then multiply $a$?
+The answer is that either way is correct.
+It doesn't matter what order we multiply the elements.
+This is because the ring $\mathcal{R}_n$ has the property of \emph{multiplicative association}.
+The multiplicative association property says that
+when we have a series of $\otimes$ operations,
+we can do the operations in whatever order we want.
+The answer will be the same.
+\begin{eqnarray*}
+  a \otimes b \otimes c  &=&  (a \otimes b) \otimes c \\
+                         &=&  a \otimes (b \otimes c)
+\end{eqnarray*}
+
+The modex function is represented mathematically using \emph{exponential notation}.
+Exponential notation is an efficient way to describe a series of multiplications of the same value.
+For example, the value $m$ can be multiplied by itself any number of times.
+We use exponential notation to describe this.
+Remember that it doesn't matter in what order the $m$'s are multiplied together.
+\begin{eqnarray*}
+  \overbrace{m}^1  &=&  m^1  \\
+  \overbrace{m \otimes m}^2  &=&  m^2  \\
+  \overbrace{m \otimes m \otimes m}^3 &=&  m^3  \\
+  \overbrace{m \otimes m \otimes m \otimes m}^4 &=&  m^4  \\
+  &\vdots&
+\end{eqnarray*}
+
+RSA uses the exponential notation $m^e$.
+The value $m$ is the \emph{message} integer.
+The value $e$ is the \emph{encryptor} exponent.
+The exponential notation $m^e$ means that $e$ copies of $m$ are multiplied together
+using the $\otimes$ operator in the ring $\mathcal{R}_n$.
+\[ m^e \quad = \quad \overbrace{m \otimes m \otimes m \, \cdots \otimes m \otimes m}^e \]
+
+RSA also uses the exponential notation $c^d$.
+The value $c$ is the \emph{ciphertext} integer.
+The value $d$ is the \emph{decryptor} exponent.
+The exponential notation $c^d$ means that $d$ copies of $c$ are multiplied together
+using the $\otimes$ operator in the ring $\mathcal{R}_n$.
+\[ c^d \quad = \quad \overbrace{c \otimes c \otimes c \, \cdots \otimes c \otimes c}^d \]
+

exponential-tables.tex

@@ -0,0 +1,46 @@
+%%%% why-RSA-works/exponential-tables.tex
+%%%% Copyright 2012 Peter Franusic.
+%%%% All rights reserved.
+%%%%
+
+%% Define an exponential product and give an example.
+We now take a closer look at exponential products $m^a$ in the ring $\mathcal{R}_n$.
+When $n$ is very small we can compute exponential products by hand.
+As an example we compute $7^3$ in the ring $\mathcal{R}_{15}$ using Table \ref{otimes-15}.
+\[  7^3 \quad = \quad 7 \otimes 7 \otimes 7 \quad = \quad (7 \otimes 7) \otimes 7 \quad 
+= \quad 4 \otimes 7 \quad = \quad 13  \]
+
+%% Define an exponential table and give an example.
+We can go on to calculate the exponential product 
+of each pair of elements in $Z_{15}$ and put them all in a table.
+Table \ref{modex-15} specifies the exponential products $m^a$ in the ring $\mathcal{R}_{15}$.
+The product of $7^3$ is at the intersection of row 7 and column 3.
+\vspace{2ex}
+%%%% modex table
+\begin{table}[!h]
+  \begin{center}
+    \input{modex-15.tex}
+    \caption{$m^a \quad (\mathcal{R}_{15})$}
+    \label{modex-15}
+  \end{center}
+\end{table}
+
+%% Define a cycle and point out examples in the table.
+Now consider the product sequence in row 3 (shown below).
+Notice how the sequence starts at 1 and then repeats itself.
+The shortest repetitive part of a sequence is called a \emph{cycle}.
+The cycle in row 3 is (3, 9, 12, 6).
+The \emph{period} of this cycle is 4.
+\[ 1 \quad \overbrace{3 \quad 9 \quad 12 \quad 6}
+     \quad \overbrace{3 \quad 9 \quad 12 \quad 6} \quad \cdots \]
+
+%% Define an identity column and point out examples in the table.
+Each row in Table \ref{modex-15} is a sequence that starts with 1 followed by a series of cycles.
+Each cycle in the various rows has a period of either 1 or 2 or 4.
+Remarkably, all of the cycles line up vertically in such a way
+as to provide what may be called \emph{identity columns}.
+Consider columns 1, 5, 9, and 13.  These are the identity columns in the table.
+Each is identical to the row number column on the left side of the table.
+So for any $m \in Z_{15}$ we have
+\[  m^1 =  m^5  =  m^9  =  m^{13}  \]
+

factor-ops.tex

@@ -0,0 +1,88 @@
+%%%% why-RSA-works/factor-ops.tex
+%%%% Copyright 2012 Peter Franusic.
+%%%% All rights reserved.
+%%%%
+%% This is LaTeX source code for a figure that contains four curves.
+%% The curves are specified by LaTeX and Lisp expressions shown below.
+%% The labels of the curves are TD, QS, NFS, and Shor.
+%% TD = Trial Division factoring algorithm.
+%% QS = Quadratic Sieve factoring algorithm.
+%% NFS = Number Field Sieve factoring algorithm.
+%% Shor = Peter Shor's factoring algorithm for quantum computers.
+%% The curves are overlayed on a 64 by 30 grid pattern.
+%% The x-axis has lines every 4 grids, with labels {128,256,384,...,1204}.
+%% The y-axis has lines every 3 grids, with labels {6,12,18,24,30}.
+%%
+%% In the Lisp code below, the expt function will accept integer exponents
+%% greater than 128 but not floating-point exponents.
+%% E.g., (expt 2 129) returns 680564733841876926926749214863536422912,
+%% but (expt 2 129.0) causes an error message to be printed.
+%% 
+%% TD curve:
+%%   $y = \log \left( \sqrt{2^x} \right)$
+%%   (setf y (log (sqrt (expt 2 x)) 10))
+%% 
+%% QS curve:
+%%   $y = \log \left( e^{\left( \left( \ln \; 2^{x} \right)^{\frac{1}{2}}\; \cdot \; 
+%%        \left( \ln \; \left( \ln \; 2^{x} \right) \right)^{\frac{1}{2}} \right)} \right)$
+%%   (setf y (log (exp (* (expt (log (expt 2 x)) 1/2) (expt (log (log (expt 2 x))) 1/2))) 10))
+%% 
+%% NFS curve:
+%%   $y = \log \left( e^{\left( \left( \ln \; 2^{x} \right)^{\frac{1}{3}}\; \cdot \; 
+%%        \left( \ln \; \left( \ln \; 2^{x} \right) \right)^{\frac{2}{3}} \right)} \right)$
+%%   (setf y (log (exp (* (expt (log (expt 2 x)) 1/3) (expt (log (log (expt 2 x))) 2/3))) 10))
+%% 
+%% Shor curve:
+%%   $y = \log \left( \left( \ln \left( 2^{x} \right) \right)^{3} \right)$
+%%   (setf y (log (expt (log (expt 2 x)) 3) 10))
+%% 
+
+\setlength{\unitlength}{1.6mm}
+\begin{picture}(64,30) 
+\linethickness{0.075mm} 
+
+%% grid pattern
+%% \multiput (x,y) (dx,dy) {n} {object}
+\multiput (0,0) (8,0) {9} {\line(0,1){30}} % x divisions
+\multiput (0,0) (0,6) {6} {\line(1,0){64}} % y divisions
+
+%% y-axis labels
+%% 6 12 18 24 30 
+\put (-2.5, 29.5){\scriptsize\textsf{30}}
+\put (-2.5, 23.5){\scriptsize\textsf{24}}
+\put (-2.5, 17.5){\scriptsize\textsf{18}}
+\put (-2.5, 11.5){\scriptsize\textsf{12}}
+\put (-2.2,  5.5){\scriptsize\textsf{ 6}}
+
+%% x-axis labels
+%% 128 256 384 512 640 768 896 1024
+\put( 6.8,-2.0){\scriptsize\textsf{128}}
+\put(14.8,-2.0){\scriptsize\textsf{256}}
+\put(22.8,-2.0){\scriptsize\textsf{384}}
+\put(30.8,-2.0){\scriptsize\textsf{512}}
+\put(38.8,-2.0){\scriptsize\textsf{640}}
+\put(46.8,-2.0){\scriptsize\textsf{768}}
+\put(54.8,-2.0){\scriptsize\textsf{896}}
+\put(62.0,-2.0){\scriptsize\textsf{1024}}
+
+\thicklines
+
+%% TD curve
+%% \qbezier (start-x,start-y) (pull-x,pull-y) (stop-x,stop-y)
+\put (7.0, 26.0) {\textsf{TD}}
+\qbezier (0.00, 0.00) (6.25, 15.00) (12.50, 30.00)
+
+%% QS curve
+\put (42.5, 25.0) {\textsf{QS}}
+\qbezier (0.00, 0.00) (7.00, 15.00) (64.00,29.65)
+
+% NFS curve
+\put (50.0, 13.25) {\textsf{NFS}}
+\qbezier (0.00,0.00) (5.50, 9.00) (64.00,13.58)
+
+%% Shor curve
+\put (51.0, 9.25) {\textsf{Shor}}
+\qbezier (0.00,0.00) ( 0.50,4.50) ( 8.00,5.84)
+\qbezier (8.00,5.84) (24.00,8.00) (64.00,8.55)
+
+\end{picture}