Escolar Documentos
Profissional Documentos
Cultura Documentos
Infosys
COPYRIGHT NOTICE
© 2010 Infosys Technologies Limited, Bangalore, India. All rights reserved. Infosys believes
the information in this document is accurate as of its publication date; such information is
subject to change without notice. Infosys acknowledges the proprietary rights of other
companies to the trademarks, product names and such other intellectual property rights
mentioned in this document. Except as expressly permitted, neither this document nor any
part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, printing, photocopying, recording or otherwise, without
the prior permission of Infosys Technologies Limited and/or any named intellectual property
rights holders under this document.
Author(s):
Amit Radha Krishna Nigam
Date written (MM/DD/YY):
02/12/2010
Target readers:
Anyone interested in knowing the applications and scope of Information Security and
Cryptography.
Keywords:
Public key cryptosystems, Trapdoor functions, Probabilistic schemes, Unaproximmable trapdoor
predicates, Qubits
Purpose of the document
To invent and study a different way to secure our existing information systems as well to gain
insight into how probabilistic algorithms can achieve it.
Paper Area
Cryptography and Network Security.
Abstract
In this paper, I first enumerate some of the major security concerns with
deterministic public key cryptographic algorithms and then I draw a
© 2010 Infosys Technologies Limited, India
Infosys <Name of the document>
1
It means that messages are Algorithm Input Output
encrypted bit by bit as opposed to Key nЄN pair =
block cipher encryption schemes generator { (p,q) :
2
because of the growing popularity of K p= public
PGP (Pretty good privacy), which key and q
makes use of RSA, many public users = private
have adopted the practice of key}
appending their public keys to Encryption Public key Encrypted
messages that they sent public forums algorithm e, message
example USENET newsgroups. E *plaintext (cipher
bit b Є text)
2) Probabilistic algorithm {0,1} and using
A probabilistic algorithm is an a random Ee(b,r).
algorithm with an additional command string r of
RANDOM that returns “0” or “1”, each length
with probability 1/2. In cryptography p(n) for
literature, these random choices are some
often referred to as 3coin flips. More polynomial
formally, probabilistic algorithms can p
be defined by Turing machines having Decryption Private Original
an additional infinite read-only tape algorithm key d and message
containing random bits. D Cipher using .
The concept of probabilistic encryption text Dd (Ee
schemes replaces the idea of (b,r) ) )
deterministic trapdoor functions with = b.
indeterminist or unapproximable *for quantum systems, this would
4
trapdoor predicates (UTP). be Є {0,1,X}, where X is the
A trapdoor function f is such that it’s superposition of two qubits
easy to evaluate f but it’s very difficult
3
I propose here a model to change these coin flips by dice rolls. This
assumption of coin flipping into the well is due to Blum and Micali [1].
4
Message spaces usually has an associated probability distribution and A
predicate p has probability p to be true wrt this distribution. A Predicate can
be any function or anything like Ex-OR or ‘all bits in message =1 etc. A
predicate B is trapdoor and unapproximable if anyone can select an x such
that B(x) = 0 or y such that B(y) = 1, but only those who know the trapdoor
information can, given z, compute the value of B(z).
Note:The definition of public key cryptosystem defined in section 1 is due
Diffie Hellman[7] in 1970.
The trapdoor used by RSA = c:=me mod n, where m is the message. The
prime factor n (=p*q) can be considered as a trapdoor.
Other disadvantages of encryption schemes based on trapdoor functions are:
1. Inverting may be easy for plain texts of some special form
2. It could be easy to compute at least partial information about the plain
text.
3. It’s easy to detect if the message is sent more than once.
All these features make trapdoor based cryptosystems, 5Semetically or
polynomially unsecure systems. This led to the development of probabilistic
algorithms where the notion of trapdoor functions is replaced by the concept
of unapproximable trapdoor predicates (UTP).
Goldwasser and Micali, was gave this concept and developed first
probabilistic scheme used the predicate function as “is quadratic residue
moduli composite n”. It’s just a concept from so many available in number
theoretic algorithms. But one has to go deep into it.
2. Quantitative comparison of different schemes as against
Quantum computers.
All these algorithms, mentioned above are widely used to achieve web
security in a highly vulnerable internet world, to encrypt mail
communications, and many other types of data.
It is obvious that the Breaking of all or any of these infact would bring
serious immediate concern for electronic commerce as well as for mobile
commerce industry.
Definitely, A way out of this dilemma would be to use some kind of quantum
cryptography only.
Besides factorization and discrete logarithms, quantum algorithms offering a
more than polynomial speedup over the best known classical algorithm have
been found for several problems. Due to [11], it is a genuine optimism to
draw mechanism that could secure us for the future quantum computers era.
Uniqueness in my approach
The novelty of this work lies in the fact that I have replaced all concepts of
numbers by problems themselves. Because a product of two numbers,
however exponential in time, can be reduced to polynomial time using
quantum systems, but a product of two problems that generate a third kind
of problem, which is unsolvable, can’t be broken unless both the constituting
set of problems of problems are known.
This paper however, does not elaborately define any Unapproximable
trapdoor Predicate (UTP) as such due to limited research done on number
theoretic implementations of a public key cryptosystem, up to this point.
1) Idea
To encrypt a message, I usually make use of a flip of a fair coin. The
encryption of each message then depends on two things. First, the message
itself, and Second, the result of the sequence of coin flips.
1. I propose here that this randomness due to coin flips can be increased
if instead of flipping a coin, I roll two fair dice simultaneously and then
encrypt the message using the results of the product of the numbers
occurring on each faces.
A dice roll can be logically realized using two random number generators
with sample space limited to the set S={1,2,3,4,5,6}. Consequently, there
are many possibilities of message encryption. But, they will always be
uniquely decodable.
2. While stating that probability makes it much difficult for a message to
be decoded, I do not impose any restrictions on the message space like in
Rabin Scheme, it is said that decoding of message is hard for an adversary if
the set of possible messages has some property.
6
(q / n) is called Jacobi symbol and (q / N) = (q / p1) (q /p2)
On second dice, the face 1 is printed with problem Q and all other faces
printed with instances of this same problem q1, q2, q3, q4, q5 all subset of
Q.
P and Q which may belong to any of the complexity classes, are solvable
problems but their product, Letit be the 7Product N=P*Q is unsolvable.
Since rolling of n dice can’t be shown in the paper, so I take the case of
rolling 2 dice, then generalize it to the result to rolling n dice.
1. Unique possible combinations. When two dice are rolled. Note that
both the dice are equivalent so 1 on dice 1 and 1 on dice 2 means the same.
Number of possible combinations D = {(1,1), (1,2)…., (2,2), (2,3), (), (),
…},,,
Number of possible roles is | D | = 62
Rolling two dice
Based on this model, a random line will be selected and all the numbers that
RLG Model
lie on this line will be multiplied to get the number.
This would ensure that public key itself is chosen from undeterministic way,
so that one who calculates the right product will be an intended user. This
would surely augment the above model too.
7
insolvability, as this combination or product would generate, means no
solution exists or I don’t know whether a solution exists or not, not even in
exponential time. Because in quantum systems, time will not be a resource
constraint.Unsolvability concept can be burrowed from any sets of problems
in complexity theory where combination of two problems is unsolvable,
although single problems themselves can be solvable. The product can mean
anything, not only mathematical product. It can be using ones problem
space in other etc.
Previous methods
Most public key system are based on the 8provably hardness of one or the
other problems from number theoretic algebra or complexity theories such
as, RSA cryptanalysis is difficult due to infeasibility of factoring the given
product of a large prime numbers. The trapdoor function used for RSA is
Let M = set of all possible messages i.e. sample space that can be
countable, uncountable, set of points such as pair of points in Elliptic curve
cryptosystems or can be a probability distribution function.
Let f be a function defined such that
fm = Vm:
For all, m Є M, let pm be the probability that m is sent and
pm (max ) = Pm, maximum probability that m is sent i.e. probability of
finding Vm
Game Analogy is as follows:
M = area to search in for the key.
F = clue to find the key
fm = Vm: is the key
Case 1 : Show him the area (message sample space) M and ask for the key,
fm without giving him the clue.
To Win :He can use a quantum mechanism like an exhaustive search for the
whole day and night and get the key.
The best he can do is always guessing Vm with probability Pm.
Case 2: Give him the clue but don’t give him the sample space i.e. keep M
infinite and them ask for the key.
The best he can do is always guessing Vm with probability Pm.
Case 3: Give him the WRONG clue and don’t give him the sample space M
i.e. keep M infinite and them ask for the key.
To Win: with the wrong clue, it’s infeasible to find the key.
The probability reduces by factor of multiplication if you give him wrong
clues again and again. THIS IS WHAT THE MODEL TRIES TO PROPOSE / DO.
Suppose that user A want to end a single qubit to user B with complete
security. The qubit is equally likely to be – 1, 0, or X(a superposition of
both).
A want that no adversary to guess even 1% of the message. For that I make
use of the model where it’s impossible to find N (=p*q).
No other classical scheme, such as difficulty arising due to factorization or
due to quadratic residuo problem, would work here as I am dealing with
much powerful systems now.
CONCLUSION
Public key cryptosystems such as RSA and the ones that are used for secret
key exchange in a pubic channel are not semantically and or polynomially
secure. Limitations they posses are of two types. Limitations due to the
underlying trapdoor functions and their inability to service any cryptanalysis
on Quantum systems. However using
(.) Probabilistic models (I have proposed here two models : ‘n dice roll’
model for building a private key and a ‘Random Line generator’ to be used
[1] Probabilistic Encryption & How To Play Mental Poker Keeping Secret,
Shaft Goldwasser and Silvio Micali, Computer Science Department,
University of California – Berkeley.
[2] An Introduction to Probabilistic Encryption, Georg J. Fuchsbauer.
[3] A new probabilistic encryption scheme, Hi Jingmin and Lu Kaicheng,
Dept. of Computer Science , Tsinghua University, China.
[4] Efficient Protocols based on Probabilistic Encryption using Composite
Degree Residue Classes, Ivan B. Damgard and Mads J. Jurik, BRICS Report
Series, March 2000.
[5] A probabilistic symmetric encryption scheme for very fast secure
communication on chaotic systems of different equations. Stergios
Papadimitrou, Tassos Bountis, Seferina Mavraoudi and Anastassios
Bezeianos, University of Patras, Grecee.
[6] Probabilistic Encryption, Journal of Computer & System Sciences 28, pp.
270–299.
[7] Diffie and Hellman public key cryptosystems.
[8] SECOQC White Paper on Quantum Key Distribution and Cryptography.
[9] Blum, Blum and Shub “ A simple unpredictable pseudo random number
generator”, SIAM. J Computing.
[10] Security of Signed ElGamal Encryption, Claus Peter Schnorr1 and
Markus Jakobsson.
[11] Experimental Demonstration of Free-Space Decoy-State Quantum Key
Distribution over 144 km,T. Schmitt-Manderbach,1,2,HenningWeier,2 Martin
Fu¨rst,2 Rupert Ursin,3 Felix Tiefenbacher,4,3 Thomas Scheidl,4,3 Josep
Perdigues,5 Zoran Sodnik,5 Christian Kurtsiefer,6 John G. Rarity,7 Anton
Zeilinger,4,3 and Harald Weinfurter1,2.
[12] EPOC3: E_cient Probabilistic Public-Key Encryption V3, Tatsuaki
Okamoto1 and David Pointcheval.
Bibliography
1. An Introduction to Probabilistic Encryption, Georg J. Fuchsbauer.
2. Probabilistic and Statistical Methods in Cryptology, Daniel
Neuenschwander.
3. Primality and Cryptography, E. Kranakis.
4. New Directions in Cryptography, Whitfield Diffie and Martin E. Hellman.
5. Cryptography and Network security, William Stalliongs, edition 2003.
6. Martin Pelikan, Kumara Sastry, Erick Cantú-Paz (Eds.)Scalable
Optimization via Probabilistic Modeling.