The growing trend in computer systems towards using scheduling policies that prioritize jobs with small service requirements has resulted in a new focus on the fairness of such policies. In particular, researchers have been interested in whether prioritizing small job sizes results in large jobs being treated “unfairly.” However, fairness is an amorphous concept and thus difficult to define and study. This article provides a short survey of recent work in this area.

Recently, size-based policies such as SRPT and FSP have been proposed for scheduling requests in web servers. SRPT and FSP are superior to policies that ignore request size, such as PS, in both efficiency and fairness given heavy-tailed service times. However, a central assumption that is usually made in implementing size-based policies in a web server is that the service time of a request is strongly correlated with the size of the file it serves. This paper shows how the performance of SRPT and FSP are affected by the degree of this correlation. We developed a simulator that supports both M/G/1/m and G/G/n/m queuing models. The simulator can be driven with trace data, which can be taken from the logs of modified Apache servers, or which can be produced by a workload generator we have developed that allows us to control the correlation. Using both trace data and generated data, we find that the degree of correlation has a dramatic effect on the performance of SRPT and FSP. In response, we propose and evaluate domain-based scheduling, a simple technique that better estimates connection times by making use of the source IP address of the request. Domain-based scheduling improves SRPT and FSP performance on web servers, particularly in regimes where correlation is low, thus making size-based policies such as these more broadly deployable.

Motivated by the optimality of Shortest Remaining Processing Time (SRPT) for mean response time, in recent years many computer systems have used the heuristic of “favoring small jobs” in order to dramatically reduce user response times. However, rarely do computer systems have knowledge of exact remaining sizes. In this paper, we introduce the class of ɛ-SMART policies, which formalizes the heuristic of “favoring small jobs” in a way that includes a wide range of policies that schedule using inexact job-size information. Examples of ɛ-SMART policies include (i) policies that use exact size information, e.g., SRPT and PSJF, (ii) policies that use job-size estimates, and (iii) policies that use a finite number of size-based priority levels. For many ɛ-SMART policies, e.g., SRPT with inexact jobsize information, there are no analytic results available in the literature. In this work, we prove four main results: we derive upper and lower bounds on the mean response time, the mean slowdown, the response-time tail, and the conditional response time of ɛ-SMART policies. In each case, the results explicitly characterize the tradeoff between the accuracy of the job-size information used to prioritize and the performance of the resulting policy. Thus, the results provide designers an understanding of how accurate job-size information must be in order to achieve desired performance guarantees.

This paper uses trace-driven simulation to study the unfairness properties of Web server scheduling strategies, such as Processor Sharing (PS) and Shortest Remaining Processing Time (SRPT). We use a general-purpose probe-based sampling approach to estimate the mean and variance of the job response time for different job sizes, for arbitrary arrival processes and service time distributions. The results illustrate two different aspects of unfairness called endogenous unfairness and exogenous unfairness. We quantify each, focusing on the mean and variance of slowdown conditioned on job size, for a range of system loads. Our work confirms recent theoretical results regarding the asymptotic convergence of scheduling policies with respect to slowdown, and illustrates typical performance results for a practical range of job sizes in an empirical workload. Finally, we show the sensitivities of SRPT and PS scheduling to selected characteristics of the arrival process and job size distribution.

Recently, there have been a number of scheduling success stories in computer applications. Across a wide array of applications, the simple heuristic of “prioritizing small jobs ” has been used to reduce user response times with enormous success. As a result of the attention given to size based policies by computer systems

Abstract—Scheduling decisions can have a pronounced impact on the performance of multi-radio wireless systems. In this paper, we study the effects of dispatch policies and queue scheduling strategies on the user-perceived performance for Internet traffic flows in a multi-channel WLAN. Our work is carried out using simulation and an empirical Web workload trace, with mean response time as the primary performance metric. The simulation results demonstrate the good/bad combination of the dispach policy with queue scheduling strategy, the advantages of deferred dispatch over immediate dispatch, and the sensitivity of dispatch policies to heavy-tailed workload characteristics. The results also highlight the pros and cons of a simple lookahead scheduling policy, particularly in the presence of high variability workloads on a heterogeneous multi-channel system with random losses. Our results provide insights into efficient and robust scheduling policies for multi-channel WLANs. Keywords: Multi-channel wireless networks, Dispatch policy, Scheduling, Simulation I.

Traditionally, the study of scheduling policies has focused on performance metrics such as response time, queue length, and throughput. However, the more vague notion of ‘fairness ’ is often equally or more important than these traditional performance metrics. But, the concept of fairness is difficult to define and so it has been studied only sporadically. This has changed over the last decade and a growing literature providing an analytic framework for studying fairness has emerged. This article surveys recent developments, which include a rich variety of fairness metrics as well as a growing understanding of the fairness of common scheduling policies. Keywords: Queueing theory, Scheduling theory, Fairness

Abstract—Requests scheduling in Web servers is a hot research topic. Many works aim at providing optimal algorithms according to various metrics. Most of these works are based on classical scheduling metrics, considering jobs completion times, but ignoring intermediate states. We claim that this choice conduces to the design of algorithm that do not efficiently share the system resources. Indeed, Web servers have some properties that make them different than the system considered in usual scheduling theory. The classical round-robin policy, used in most production Web servers, has intrinsic qualities: it shares equally the system resources and avoids any job starvation. We introduce a novel parameterizable algorithm proposing a compromise between the benefits of the round-robin and the policies that provide the best performances. Then, we discuss the appropriate choice of the parameter depending in the requirements and the context of the Web server. I.