Size-based scheduling to improve web performance

362 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Hopper: Decentralized Speculation-aware Cluster Scheduling at Scale (2015, background): “Despite being intuitively unfair to large job sizes, it in fact improves the average response time of every job size (when job sizes are heavytailed) compared to fair schedulers [28, 43, 44].”
Performance Modeling and Design of Computer Systems: Queueing Theory in Action (2013, methodology): “In [92, 162], SRPT scheduling is implemented for an Apache web server running on Linux by modifying the Linux kernel to schedule outgoing bandwidth so as to favor HTTP requests with small remaining file size.”
Sharing the Data Center Network (2011, background): “If the provider has perfect knowledge about demands, scheduling the shortest remaining transfer ﬁrst will minimize the activity period [18].”
Open Versus Closed: A Cautionary Tale (2006, methodology): “While Processor-Sharing scheduling (PS) and First-Come-First-Served (FCFS) are most commonly used in computer systems, many system designs give preference to short jobs (requests with small service demands), applying policies like Non-Preemptive-ShortestJob-First (SJF) or Preemptive-Shortest-Job-Fir”