A Heuristic Approach to Computer System Tuning

摘要

We describe a program that uses heuristics to the diagnosis of bottlenecks in computer systems. The program provides the necessary remedies to achieve acceptable computer performance. The technique relies on a high level functional model of the interaction between application workloads, the UNIX operating system, and system hardware. While the system is running, the program predicts the values of observable system counters available from the UNIX performance-monitoring tools. During diagnostic inference, we use observed performance monitor values to find the most probable assignment to the 晈orkload parameters. We consider the tuning problem as two interrelated activities: self-tuning and learning. We provide some background on automated bottleneck detection, describe the structure of the system model, and discuss empirical procedures for implementing these activities.