Simulations driven by Google cloud traces show mean response time minimized by varying server count at fixed budget, simple policies like Join Idle Queue matching size-based ones at high parallelism, and multi-stage clusters with Round Robin outperforming complex policies.
Optimal scheduling in general multi-queue system by combining simulation and neural network techniques,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
The Merit of Simple Policies: Buying Performance With Parallelism and System Architecture
Simulations driven by Google cloud traces show mean response time minimized by varying server count at fixed budget, simple policies like Join Idle Queue matching size-based ones at high parallelism, and multi-stage clusters with Round Robin outperforming complex policies.