|
Sven Laur
Master of Science, Researcher, Cryptography
| Office: |
Room T-B256
|
| Postal Address: |
Helsinki University of Technology,
Laboratory for Theoretical Computer Science,
P.O.Box 5400,
FI-02015 HUT,
FINLAND
|
| Telephone: |
+358-9-451 5876 |
| Email: |
swen@math.ut.ee |
|
I have an explicit white list spam filtering policy.
All leters not in the white list will be deleted with probability 99%.
To get through the spam into the white list add "Soovin saada kontakti" into the Subject field.
After that I'll read the mail and add your address into the white list.
Sorry, for inconvienience but this is the best way I know how to handle spam.
|
My main research interests
- Provable security with exact success bounds
- Efficient implementations of secure computation protocol
- Cryptographic protocol design and proof methodologies
- Homological classification of cryptographic primitives
Current projects
- Sharemind
C++ implementation of secure multiparty computation.
Aimed to be a blazingly fast assembler level code interpretator.
The main aim is to get a fast prototyping environment for privacy-preserving data mining.
Most of the coding and design is done by my supervised master student Dan Bogdanov.
- My PhD Thesis: Cryptographic Protocol Design
The thesis consists of two parts: published articles and the introductionary part.
The introductionary part is aimed to provide a fresh alternative to complexity theory based protocol design:
- Takes an engineering view on cryptography:
- Security definitions must reflect practical requirements.
- Unnecessary proof complexity must be systematically abstracted away.
- Security assumptions should be stated so that one can intelligently
choose a design from several alternatives.
- Abstracts away complexity theoretic assumptions and uses more natural framework of hypothesis testing to formalise
- required security premises,
- obtainable provable guarantees.
- Provides a well systematised proof methodology in the terms of hypothesis testing.
- Asks and provides answers to non-conventional questions:
- Why do we think that SHA-1 is secure although it is a single function?
- Why do we trust fixed public keys although they cannot be proven to be secure?
- Does the unsolvable conflict between individual and collective properties in
statistics emerge also in cryptography?
- When is it secure to join a cryptographically secure protocol?
- Does an ideal implementation of the functionality preserve security
if an attacker has additional knowledge?
- How to rationally decide whether to join a secure protocol or not?
- Considers composability as an excellent design tool and not the final goal:
- Composability must be otained with a low amortised computational overhead.
- Additional initialisation assumptions are allowed to achive composability.
- We promote extend assumptions, construct a secure protocol and reduce
assumptions by isolation paradigm.
[TCS main]
[Current]
[Contact Info]
[Personnel]
[Research]
[Publications]
[Software]
[Studies]
[Links]
Latest update: 30 April 2007.
|
