Jorvik: A framework for effective scheduling

This paper presents a collection of mechanisms that together form a framework for the support of flexible scheduling. By choosing a mixture of established and novel techniques, a robust scheme is developed that can provide real support to a wide range of applications that mix hard and soft tasks, periodic and aperiodic load and complex intertask relationships. Key aspects of the framework are the low inherent implementational overheads - no worse that $O(logn)$, and the fact that it can be implemented with commercial programming languages and operating system standards. A novel aspect of the framework is the use of a $O(logn)$ algorithm for a threshold based acceptance test for firm tasks. Simulations are provided which illustrate the effectiveness of the proposed scheme. In general it performs as well as the Best-effort scheme which has considerable higher complexity and run-time overhead.