A feature model and development approach for schedulers

Abstract

Schedulers decide when to execute what in a system. They often work in constrained environments, where these decisions have high impact on performance. Since schedulers should be designed according to a system's needs, it is imperative that scheduling requirements are well defined. Building a scheduler that satisfies these requirements is not a trivial task. In this position paper we present our initial work and ideas on a domain-specific framework for scheduling. We introduce our feature model of the problem space of scheduling and explain how it fits within a framework. Furthermore we present our ideas on how to make the scheduler framework modular by using an aspect oriented approach.

Managing software complexity of adaptive systems

“…Future printing systems are going to implement more flexible scheduling algorithms [56]. Such algorithms can reduce or eliminate the delay to change speed, lowering the need to take margins into account.…”