Marin Aranitasi and Levis Jani


Self-tuning is becoming a pressing issue for operating systems. The system class can change when the running applications change. So we need an algorithm to address these problems. In this paper we propose an algorithm, implemented in C language, for self-tuning operating system that is based on a monitor-classify-adjust loop. The idea of this loop is to continuously monitor certain performance metrics, and whenever these metrics changes, the system determines the new system class and dynamically adjusts tuning parameters for this new class. We have created the KernRing, a prototype tool that identifies the system class and improves system performance automatically. To do that it uses a support vector machine (SVM) to identify the system class, and tunes the operating system for that specific system class. It has three components: the Monitor, the Classifier and the Adjustor. We are going to present the design and implementation details for KernRing.


SVM, kernel, operating system, performance