Toward absolutely dependable, autonomic personal computing systems

点击翻译按钮获取中文摘要

A variety of technology trends are combining to revolutionize the way that people acquire, organize and manipulate information in digital form.   One important consequence of this revolution is a dramatic increase in the number and variety of computing platforms a person uses on a daily basis.  Currently, for example, a typical user might work on a single task using multiple desktop and/or portable computers in a number of different locations (e.g., at work, at home and on the road).  In the near future, similar interactions will involve an even greater diversity of devices --- television-set-top boxes, game consoles, media players (e.g., an iPod), digital cameras, video recorders and cell phones --- each enabled with massive data-storage capacity and network connectivity (e.g., 802.11, WiFi).  This environment presents considerable data-management, system integration, security, reliability and networking challenges that are not adequately addressed by current approaches.   This research program will develop novel system software to address a variety of issues that arise in this environment.  The goal is to present users with a single, simple and absolutely reliable view of their data, regardless of where it is stored or from which device it is accessed.  To achieve this goal requires tackling several open problems at the core of operating systems and distributed systems research.  For example, two key issues are that current systems are too unreliable and require too much human administration.  This research program will develop techniques that secure valuable data such as family photos from hardware failure, software bugs and malicious attack (e.g., maleware such as viruses).  It will also build mechanisms for automatic configuration, self tuning, re-configuration and recovery that require no human intervention.