Model-Driven QoS Provisioning for Distributed Real-Time and Embedded Systems

Distributed Real-time and Embedded (DRE) systems are composed of applications with diverse CPU utilization requirements. These applications participate in many end-to-end application flows with diverse network bandwidth requirements. Prior research has advanced the state-of-the-art on bin-packing algorithms, such as First Fit Decreasing (FFD), to allocate per-application CPU resources, and network-layer Quality-of-Service (QoS) mechanisms, such as Differentiated Services (DiffServ), to manage per-flow network resources. Relatively little work, however, has focused on how applications can be deployed and configured to leverage such advances for addressing their end-to-end QoS requirements.