University of PardubiceFaculty of Economics and Administration

University

Employees
|
Students
Česky

Publication detail

Bezpečnost informací - autentizace

Authors: Hub Miloslav

Year: 2005

Type of publication: odborná kniha

Publisher name: Univerzita Pardubice

Place: Pardubice

Page from-to: nestránkováno

Titles:

Language	Name	Abstract	Keywords
cze	Bezpečnost informací - autentizace	V práci je navržen model vícefaktorové autentizace založené na kombinaci dvou nezávislých druhů identifikačních znaků ? znalosti a biometrické charakteristiky autentizované osoby. Znalost je reprezentována příslušným heslem, jako biometrická charakteristika je zvolena dynamika psaní příslušného hesla na klávesnici. Jako kvantitativních kritérií ohodnocení navrženého modelu dvoufaktorové autentizace je použito podmíněné pravděpodobnosti chybného přijetí neoprávněného narušitele a podmíněné pravděpodobnosti chybného odmítnutí oprávněného uživatele, na základě kterých je sestrojen DET graf.	Autentizace, identifikace, bezpečnost informací, hesla
eng	Information security - authentication	In this paper the model of multi-factor authentication based on combination of two independent varieties of identification attributes ? the knowledge and the biometric characteristic of the authentication person ? is proposed. The knowledge is represented by the appropriate password, as a biometric characteristic is chosen a dynamics of typing the appropriate password on the keyboard. As the quantitative criteria of the evaluation of the presented model of the two-factor authentication it is used the conditional probability of the impostor error acceptance and the conditional probability of the error rejection of a valid user, in the terms of which the DET graph is designed. For the purpose of specification of this model there are tested different measures of dissimilarity between the identification vector, which is submitted by the authentication subject, and the appropriate template of the identification vector. The best results from the aspect of chosen quantitative criteria are accomplished through a suggested measure of dissimilarity-residual dispersion with utilization of Laplace probability density function. For determination of the suitable boundary of the dissimilarity measure presented in the paper it is analyzed the use of non-Bayes tasks of the statistic recognition. At the same time the model of the multi-factor authentication is also validated for the case of special passwords that are only generated by random numbers and random alphanumeric symbols. For comparison of the model of multi-factor authentication and the model of authentication based only on passwords there were simulated two types of attack on the really used passwords ? the dictionary attack and the brute force attack. The proposed model of the multi-factor authentication is evaluated not only in the quantitative way, but also the qualitative point of view is considered using the propositional calculus.	authentication, data security, identification, biometric, passwords