Energy-efficient soft real-time CPU scheduling for mobile multimedia systems
SOSP '03 Proceedings of the nineteenth ACM symposium on Operating systems principles
Rotational-Position-Aware Real-Time Disk Scheduling Using a Dynamic Active Subset (DAS)
RTSS '03 Proceedings of the 24th IEEE International Real-Time Systems Symposium
QoS Control Strategies for High-Quality Video Processing
Real-Time Systems
Energy-efficient CPU scheduling for multimedia applications
ACM Transactions on Computer Systems (TOCS)
Enforceable component-based realtime contracts
Real-Time Systems
Adaptive Resource Allocation Control for Fair QoS Management
IEEE Transactions on Computers
A real-time programmer's tour of general-purpose L4 microkernels
EURASIP Journal on Embedded Systems - Operating System Support for Embedded Real-Time Applications
A real-time programmer's tour of general-purpose L4 microkernels
EURASIP Journal on Embedded Systems - Operating System Support for Embedded Real-Time Applications
Efficient and robust probabilistic guarantees for real-time tasks
Journal of Systems and Software
Hi-index | 0.02 |
We present a unified model for admission and scheduling,applicable for various active resources such as CPU or diskto assure a requested quality in situations of temporaryoverload. The model allows us to predict and control thebehavior of applications based on given quality requirements.It uses the variations in the execution time, i.e., the time anyactive resource is needed.We split resource requirements into a mandatory partwhich must be available and an optional part which shouldbe available as often as possible but at least with a certainpercentage.In combination with a given distributionfor the execution time we can move away from worst-casereservations and drastically reduce the amount of reservedresources for applications which can tolerate occasionaldeadline misses. This increases the number of admittableapplications.For example, with negligible loss of qualityour system can admit more than two times the disk band-widththan a system based on the worst-case.Finally, we validated the predictions of our model byMeasurements using a prototype real-time system and observeda high accuracy between predicted and measured values.