TC: Small: Mobile Phone Password Managers: An Evaluation and a Re-Design based on Human-Perceptible Communication

点击翻译按钮获取中文摘要

Textual passwords and personal identification numbers (PINs) are the most dominant means of authentication used currently; and this trend is very likely to continue in the future. However, passwords are either difficult to use (if they are long and randomly generated), or insecure (if users are given the choice of their own passwords). Password Managers are one promising approach aimed to improve the usability and security of passwords, by having a computing device, rather than the user herself, store (and optionally, generate) passwords, and then later deliver or recall them to the user whenever access is needed. A number of password management schemes have been proposed and are employed currently by many affected users. This project concentrates on password management using mobile devices (e.g., cell phones), whose ubiquity makes them an appealing authentication aid. It embarks upon two research directions vis-a-vis such phone managers. First, the project aims at a systematic and formal evaluation, via usability studies and surveys, of currently deployed phone managers. Second, it proposes a redesign that exploits the many different capabilities and characteristics of modern phones (such as on-board sensors and computational resources), in order to address several of the drawbacks with current phone managers. Specifically, a general-purpose password management approach -- called proxy-based authentication -- is introduced. As the name suggests, this approach uses the phone as an authentication proxy between the user and the device to be authenticated to. The project explores how proxy-based authentication can be used to strongly authenticate to: (1) critical online services -- that continue using passwords or PINs -- without incorporating any service-specific modifications, (2) local devices (such as desktops, laptops, ATMs), and (3) ubiquitous but constrained devices (such as personal RFID tags) for user-controlled privacy and enhanced security.The technical merit of this project lies in two aspects. First, it will arrive at a better understanding of current phone managers in terms of usability, efficiency, and security. The goal is to gain insights into users' mental models when using these password managers. Second, the project will pursue the realization of usable proxy-based authentication primitives. To this end, this work is able to simply reuse wealth of existing research on usable user-phone authentication. Instead, the main thrust is on exploring the design and evaluation of usable authentication methods between the phone and the service that ultimately requires authentication. In particular, for phone-service authentication, the project investigates novel short-range human-perceptible (HP) communication that is commonly and cheaply available, fast, robust, least intrusive, and low-power. Notably, the research investigates how to use HP communication to create authenticated channels, and authenticated and eavesdropping resilient channels. Based on the principle of extrinsically motivated design, the project also explores playful HP channels. These channels make the task of manual transmission a fun and entertaining activity for the users.The anticipated impacts of the project include: (1) enhanced interaction among several disciplines including security and cryptography, computer and electrical engineering, networking, and usability and HCI; (2) increased awareness among students and users regarding security practices vis-a-vis one of the most important security problems (authentication); (3) integration of PI's research with educational activities, enabling students taking part in the project to acquire currently uncommon skills at the cusp of Human-Computer Interaction and Trustworthy Computing; (4) emphasis on technology transfer by working with manufacturers and industrial consortia. Another long-term impact of this work is the development of security technologies that can eventually be put to use by general population, i.e., are usable in the true sense. Furthermore, the work is expected to be instrumental in stimulating research on usable security technologies for the blind or visually impaired users who are usually at a high risk for various security vulnerabilities and attacks, perhaps more so in the context of authentication.