ITR: Usable Security: Towards a Trustable Information Infrastructure

点击翻译按钮获取中文摘要

The project is developing an infrastructure for visualizing the real-time state of software system security. Information visualization exploits aspects of the human perceptual system to recognize salient facts, correlations, and features of a complex information space. By presenting visualizations in real time, allows users to relate system security to their own actions, and thereby gain a deeper understanding of how security can be enhanced-or compromised-as a part of their experience of using computer systems. Visualization offers a means by which one can incorporate information about system security into all aspects of computer system use. This is critically important since the security of information in a networked computer system depends on the interrelationships between many different systems, components, and applications; there is no one point of control. Any successful approach must be comprehensive.A concern with security is not with mathematical abstraction but with practical reality. There is a disparity between theoretical security and effective security in day-to-day computing. Theoretical security is the level of information security one can achieve in theory, using strong encryption, zero-knowledge systems, and the current state of the art. On the other hand, effective security is the actual level of information security one can achieve in practice. A system can have high theoretical security but low effective security when the security mechanisms are implemented in ways that confuse or confound users.In his classic article "Why Cryptosystems Fail," Ross Anderson (1993) outlines two paradigms for system security. One is the "automation" paradigm, in which the work of managing information security is automated and embedded in the machinery of the system. The other is the "facilitation" paradigm, in which humans can monitor and managing the security process, adapting it to changing needs and circumstances. He argues that many security failures have their roots in the inherent brittleness of the dominant automation paradigm. However, for regular users, the design of conventional applications and interfaces systematically undermines the facilitation approach. The proposed research addresses this problem with an approach to visualizing system security. This work is not simply applying usability principles to security applications; rather, it is about making security a part of the regular user experience.This work builds on the investigators' existing explorations in system architecture, software system visualization, and end-user understandings of security. The combination of these, however, yields an infrastructure that is not only technological novel but also holds significant promise for making the growing cyber-infrastructure accessible and secure for everyday use. On a scientific level, this research makes novel contributions in four areas: the mental models of Internet users, the use of visualization technologies for software system monitoring, the use of event architectures for interactive systems, and the relationship between user models and system architecture. In addition the project will develop an experimental infrastructure that the research community can exploit for research into interactive visualization and security. At a broader level, this research offers significant benefits to society at large. The impact of the Internet as a cultural and economic phenomenon is hard to overstate, and yet the current mechanisms that support users' secure exchange of information are brittle. A resolution to this problem is pressing, and is the goal of this research.