The question
Can a scheduler improve completion times and deadline attainment without starving large jobs or violating resource limits?
This is a proposed research direction. No novel algorithm, completed benchmark or performance gain is claimed.
Why it matters
Compute platforms, query engines and build systems could serve mixed workloads more predictably.
Search space
Queue ordering, aging, resource estimates, admission and preemption policies.
What would count as progress?
- Completion-time distributions
- deadline misses
- throughput
- fairness
- starvation
- scheduling overhead.
A useful result improves a measured trade-off against strong baselines on held-out workloads, with reproducible code and documented limitations.
Baselines and boundaries
Compare FIFO, fair sharing and size-aware policies with aging. Define fairness explicitly, include inaccurate size estimates and preemption costs, and validate trace-replay winners on real execution.
Where you could contribute
Help define representative, legally shareable workloads; critique the correctness contract; implement baselines; propose candidates; or reproduce measurements. Negative results and failure cases are useful contributions.
A detailed research brief is being prepared. Access requests will become available when that brief and the request service are ready.