|
Type of Document Dissertation Author Henry, Peter Thomas URN etd-07022007-184611 Title Toward Usable, Robust Memometric Authentication: An Evaluation of Selected Password Generation Assistance Degree Doctor of Philosophy Department Information Studies, College of Advisory Committee
Advisor Name Title Charles R. McClure Committee Chair Gary Burnett Committee Member John Carlo Bertot Committee Member Michael Burmester Committee Member Keywords
- usability
- memometrics
- authentication
- information security
- security
- password
Date of Defense 2007-05-21 Availability unrestricted Abstract This dissertation explored the effects of various types of assistance on the generation, recall, and input of robust passwords containing at least twenty characters. Passwords are desirable memometric authentication secrets for many reasons, but their continued effectiveness depends on increasing their resistance to emerging attacks. Resistance to attacks is increasingly a function of length. Although previous password research revealed widespread use of short, weak passwords and conventional wisdom considers users incapable of reliably generating, recalling, and accurately inputting strong passwords, this study investigated ways to assist users in meeting the specific challenges of robust password management.Interventions in the password-generation stage of this study introduced participants to five password generation schemes, supplied various numbers of example passwords, and required reentry of passwords immediately after generation to explore possible benefits on subsequent authentication performance. Key findings of this research were that:
· Twenty-character passwords can be as strong as their corresponding 128-bit hashes;
· Acrostic password-generation schemes produced strong passwords;
· Confessional and Unexpected Nonsense schemes produced memorable passwords;
· Supplying example passwords led to stronger passwords;
· All participants easily generated 20-character passwords and most experienced few problems in the vague recall of them;
· 30% of participants generated and used very strong passwords without failure for seven weeks;
· The input of the precise formulation of robust passwords was the greatest single cause of authentication failure;
· Exposure to 5 or 10 additional password examples during the generation stage did not improve subsequent password performance;
· Reentry of passwords four times during the generation stage did not improve subsequent password performance;
· Although education and training are beneficial, the actual study treatments were not universally effective; and
· The population of password users and the reasons for password failure are complex, and users who experience difficulties require additional attention and resources on a contingency basis.
Files
Filename Size Approximate Download Time (Hours:Minutes:Seconds)
28.8 Modem 56K Modem ISDN (64 Kb) ISDN (128 Kb) Higher-speed Access pth_dissertation.pdf 2.34 Mb 00:10:50 00:05:34 00:04:52 00:02:26 00:00:12