Abstract
In the world of cryptography, stream ciphers are known as primitives used to ensure privacy over a communication channel. One common way to build a stream cipher is to use a keystream generator to produce a pseudorandom sequence of symbols. In such algorithms, the ciphertext is the sum of the keystream and the plaintext, resembling the onetime pad principal. Although the idea behind stream ciphers is simple, serious investigation of these primitives has started only in the late 20th century. Therefore, cryptanalysis and design of stream ciphers are important.
In recent years, many designs of stream ciphers have been proposed in an effort to find a proper candidate to be chosen as a world standard for data encryption. That potential candidate should be proven good by time and by the results of cryptanalysis.
Different methods of analysis, in fact, explain how a stream cipher should be constructed. Thus, techniques for cryptanalysis are also important.
This thesis starts with an overview of cryptography in general, and introduces the reader to modern cryptography. Later, we focus on basic principles of design and analysis of stream ciphers. Since statistical methods are the most important cryptanalysis techniques, they will be described in detail.
The practice of statistical methods reveals several bottlenecks when implementing various analysis algorithms. For example, a common property of a cipher to produce nbit words instead of just bits makes it more natural to perform a multidimensional analysis of such a design. However, in practice, one often has to truncate the words simply because the tools needed for analysis are missing. We propose a set of algorithms and data structures for multidimensional cryptanalysis when distributions over a large probability space have to be constructed.
This thesis also includes results of cryptanalysis for various cryptographic primitives, such as A5/1, Grain, SNOW 2.0, Scream, Dragon, VMPC, RC4, and RC4A. Most of these results were achieved with the help of intensive use of the proposed tools for cryptanalysis.
In recent years, many designs of stream ciphers have been proposed in an effort to find a proper candidate to be chosen as a world standard for data encryption. That potential candidate should be proven good by time and by the results of cryptanalysis.
Different methods of analysis, in fact, explain how a stream cipher should be constructed. Thus, techniques for cryptanalysis are also important.
This thesis starts with an overview of cryptography in general, and introduces the reader to modern cryptography. Later, we focus on basic principles of design and analysis of stream ciphers. Since statistical methods are the most important cryptanalysis techniques, they will be described in detail.
The practice of statistical methods reveals several bottlenecks when implementing various analysis algorithms. For example, a common property of a cipher to produce nbit words instead of just bits makes it more natural to perform a multidimensional analysis of such a design. However, in practice, one often has to truncate the words simply because the tools needed for analysis are missing. We propose a set of algorithms and data structures for multidimensional cryptanalysis when distributions over a large probability space have to be constructed.
This thesis also includes results of cryptanalysis for various cryptographic primitives, such as A5/1, Grain, SNOW 2.0, Scream, Dragon, VMPC, RC4, and RC4A. Most of these results were achieved with the help of intensive use of the proposed tools for cryptanalysis.
Original language  English 

Qualification  Doctor 
Awarding Institution 

Supervisors/Advisors 

Award date  2006 Jun 16 
Publisher  
ISBN (Print)  9171670394 
Publication status  Published  2006 
Bibliographical note
Defence detailsDate: 20060616
Time: 10:15
Place: E:1406, Ehuset, Lunds Tekniska Högskola
External reviewer(s)
Name: Daemen, Joan
Title: Ph.D.
Affiliation: ST Microelectronics, Belgien

Subject classification (UKÄ)
 Electrical Engineering, Electronic Engineering, Information Engineering
Free keywords
 VMPC
 Dragon
 Grain
 A5/1
 SNOW 2.0
 Scream
 large distributions
 correlation attacks
 distinguishing attacks
 algorithms
 Cryptography
 cryptanalysis
 RC4
 RC4A
 Computer science
 numerical analysis
 systems
 control
 Datalogi
 numerisk analys
 system
 kontroll