Size-based scheduling to improve web performance

Abstract

Is it possible to reduce the expected response time of every request at a web server, simply by changing the order in which we schedule the requests? That is the question we ask in this paper.This paper proposes a method for improving the performance of web servers servicing static HTTP requests. The idea is to give preference to requests for small files or requests with short remaining file size, in accordance with the SRPT (Shortest Remaining Processing Time) scheduling policy.The implementation is at the kernel level and involves controlling the order in which socket buffers are drained into the network. Experiments are executed both in a LAN and a WAN environment. We use the Linux operating system and the Apache and Flash web servers.Results indicate that SRPT-based scheduling of connections yields significant reductions in delay at the web server. These result in a substantial reduction in mean response time and mean slowdown for both the LAN and WAN environments. Significantly, and counter to intuition, the requests for large files are only negligibly penalized or not at all penalized as a result of SRPT-based scheduling.

Stability of size-based scheduling disciplines in resource-sharing networks

“…Note that in case of infinite variance the mean delay under the First-Come First-Served discipline is infinite for any load, whereas the SRPT discipline yields a finite mean delay as long as the system is stable [33]. For the latter reasons, the SRPT discipline has been widely adopted as an effective mechanism for improving the response time performance in web servers [11,18]. To some extent, the huge variability in job sizes also alleviates the long-standing concerns that have surrounded SRPT regarding the perceived unfairness towards extremely long jobs [7,16,35].…”

Decoupled speed scaling: Analysis and evaluation

“…With this approach, the mean waiting time and the mean response time are minimized [15], [16], [17], [18]. SRPT scheduling has generated significant research interest, particularly in the context of request scheduling in Web servers [19], [20], [21], [22], [23]. Under certain job size distributions, SRPT even has a counter-intuitive "all can win" property, where all jobs prefer SRPT to PS [19].…”

Decentralized Preemptive Scheduling Across Heterogeneous Multi-core Grid Resources

“…Shortest remaining time [14] is a scheduling method that is a preemptive version of shortest job next [15] scheduling. In this scheduling algorithm, the process with the smallest amount of time remaining until completion is selected to execute.…”