Strong authentication methods (EWI.4536)
Project nummer:
ewi4536
Omschrijving van het onderzoek
Paper documents normally carry a signature as legal proof that the signer agrees with the contents of the document. In case of dispute, the holder of a document will use this signature in court. If the document is forged, this becomes clear from a fake signature or from demonstrable tampering with the text.
In the last ten years more and more communication takes places in digital form over telephone lines(e.g. fax) and internet (E-mail, WWW). Inside multinationals, traditional systems like letters and telex are no longer the main method of communication.
It is common knowledge that telephone and internet lines are quite accessible for unauthorised people and, even worse, that they are not too difficult to tamper with (because the data consist of bits this can be done without leaving any trace). Privacy of data is an important issue in this context, but much bigger security riscs for the legitimate users are:
- proof that the indicated sender is the real sender (proof that convinces the receiver and that, preferably, holds in court),
- proof that not a single letter in the text has been changed by others (think of a change of account number in a bank transmission).
The above issues are in a similar way relevant when data are stored in a computer for later retrieval. There are technical solutions that solve these problems. They are called digital signature schemes (subproject A) or message authentication codes (subproject B), depending on the underlying techniques. These solutions are not completely satisfactory. There are four aspects that seem to interfere with each other:
- the desired level of security (of importance in case of a dispute),
- the complexity of the used techniques (necessary computerpower, speed),
- the practicality of the used methods (cumbersome update, key distribution),
- the assumption underlying the solution.
The practical goal of this project will be to develop authentication methods with a better overall performance.
The theoretical goal will be to achieve a better understanding of how the above four points interact with each other.
Resultaten van het onderzoek
Twee nieuwe families van priemgetallen zijn gevonden. Aanwijzingen zijn gegeven voor een veiliger toepassing van het McEliece cryptosysteem. Het genereren van pseudorandom getallen voor sleutelgeneratie is verbeterd. Aangetoond is dat digitale handtekeningen gezet volgens het Alabaddi and Wicker schema niet veilig zijn. Een versnelling is bereikt van XTR en een aanverwant cryptosysteem. Zodra XTR serieus geïmplementeerd wordt, zullen de ontwikkelde algoritmes daar zeker een rol in spelen. Een bepaalde variant van elliptische krommen is versneld. Elliptische krommen worden op grote schaal toegepast, bijvoorbeeld in smartcards. Toepassing van deze resultaten verloopt via publicaties en kan pas op de langere termijn verwacht worden.
Gebruikers
Seven companies are involved in this project.
Projectleider
Prof.dr.ir. H.C.A. van Tilborg
Technische Universiteit Eindhoven
Wiskunde en informatica
Postbus 513
5600 MB Eindhoven
Status van het project
| Gestart |
: 01-12-1998 |
| Einddatum |
: 01-08-2003 |
Trefwoorden
Authenticatie; Betalingsverkeer; Chipkaart; Cryptografie; Digitale handtekening