A two-stage dispatching architecture with an optimized job-size threshold improves mean response time over single-stage classical policies in large clusters.
Network analysis without exponen tiality assump- tions,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.DC 2years
2025 2verdicts
UNVERDICTED 2representative citing papers
Simulations on real Google traces show JIQ and LWL dispatching have optimal cluster sizes for fixed utilization while RR worsens monotonically; task decomposition reveals simpler JIQ outperforming size-aware LWL.
citing papers explorer
-
"Two-Stagification": Job Dispatching in Large-Scale Clusters via a Two-Stage Architecture
A two-stage dispatching architecture with an optimized job-size threshold improves mean response time over single-stage classical policies in large clusters.
-
Dispatching Odyssey: Exploring Performance in Computing Clusters under Real-world Workloads
Simulations on real Google traces show JIQ and LWL dispatching have optimal cluster sizes for fixed utilization while RR worsens monotonically; task decomposition reveals simpler JIQ outperforming size-aware LWL.