In this paper, Sterritt and Chung present a proof-of-concept self-healing tool for a personal computer (PC) environment. They begin by outlining the goals of their field of research, autonomic computing, which is a computing environment able to monitor its own health, and react when it begins to fail.
The core components of such an environment are heartbeat and beacon monitoring; the former looks after the state of each computer, and the latter looks after the state of neighboring computers. In this way, the system as a whole achieves the four key properties of an autonomic computing environment: self-configuration, self-healing, self-optimization, and self-protection. The authors then explain that these concepts were born in a very technical context, the National Aeronautics and Space Administration’s (NASA’s) Deep Space 1 (DS1) mission, and outline the problems, as well as the challenges, involved in extending the field to a less technical environment (such as a PC network).
Then comes the most technical, important, and interesting part of the paper, in which the authors describe how they designed and implemented their tool, the technological and architectural alternatives they faced, and the reasons for the choices they finally made. We learn that each PC participating in this proof-of-concept environment has a heartbeat monitor, used to record its state of health; a pulse broadcast component, to send its state to other PCs; a pulse monitor, to receive the state of health of other PCs; and the heart of the system, an adaptive and configurable knowledge base, used to decide the course of action to take in response to the state of the environment. The technologies involved range from the C language (used to implement the heartbeat monitor), to the Java language (used to send pulses to and receive pulses from other PCs), to the user datagram protocol (UDP) (for message handling). The historical perspective used by the authors to frame their wor!k, from early concepts developed at NASA to the future of their own tool, is also interesting.
For these reasons, this paper makes for smooth and pleasant reading; it is also well suited for an audience not familiar with the specific field. A general knowledge of computer science or information technology is, however, recommended.