Analysis of LAS scheduling for job size distributions with high variance

Abstract

Recent studies of Internet traffic have shown that flow size distributions often exhibit a high variability property in the sense that most of the flows are short and more than half of the total load is constituted by a small percentage of the largest flows. In the light of this observation, it is interesting to revisit scheduling policies that are known to favor small jobs in order to quantify the benefit for small and the penalty for large jobs. Among all scheduling policies that do not require knowledge of job size, the least attained service (LAS) scheduling policy is known to favor small jobs the most. We investigate the M/G/1/LAS queue for both, load < 1 and 1.Our analysis shows that for job size distributions with a high variability property, LAS favors short jobs with a negligible penalty to the few largest jobs, and that LAS achieves a mean response time over all jobs that is close to the mean response time achieved by SRPT.Finally, we implement LAS in the ns-2 network simulator to study its performance benefits for TCP flows. When LAS is used to schedule packets over the bottleneck link, more than 99% of the shortest flows experience smaller mean response times under LAS than under FIFO and only the largest jobs observe a negligible increase in response time. The benefit of using LAS as compared to FIFO is most pronounced at high load.Size-based scheduling, least attained service, high variability property, Web objects response time.

Batch Arrival Processor-Sharing with Application to Multi-Level Processor-Sharing Scheduling

“…Hence, researchers have recently analyzed and proposed different size-based scheduling disciplines that aim to improve the performance that the current network architecture provides [2,[8][9][10]17]. In the next section, we describe these schedulers as particular examples of the set of MLPS disciplines introduced by Kleinrock [12, Section 4.7] and we show how the results presented for the BPS queue can be applied.…”

“…Size-based scheduling has recently received a fairly big attention in connection with the differentiation of Short and Long flows in the Internet [2,[8][9][10]17]. Kleinrock et al [12][13][14] introduced a quite general set of size-based scheduling termed as Multi-Level Processor-Sharing (MLPS).…”

When Mice devour the Elephants: A DDoS attack against size-based scheduling schemes in the internet

“…In this work we used the ns2 simulator (The network simulator) since it is widely used to study DDoS attacks in the Internet (e.g., see Mondal and Kuzmanovic, 2010;Maci a-Fern andez et al, 2009;Wang and Shin, 2003;Kuzmanovic and Knightly, 2006;Shevtekar and Ansari, 2008). Our decision to use ns2 was further supported by the fact that the scheduling algorithms being evaluated in this work have very widely been studied using the same simulation environment (e.g., see LAS in Rai et al, 2003;Rai et al, 2004). These factors convinced us that ns2 was a credible environment for the investigation of DDoS activity against LAS, and we hence used it for all our experiments.…”

“…2 shows the topology used in the study. This kind of dumbbell topology represents the "atomic unit" of larger and much more complex networks and is widely used in simulation-based performance evaluation of networking algorithms and protocols (e.g., see Kuzmanovic and Knightly, 2006;Rai et al, 2003;Kuzmanovic and Knightly, 2003;Rai and Serwadda, 2010;Kuzmanovic, 2005). In our experiments, legitimate TCP and UDP flows traversed a bottleneck link, R1ÀR2, which was also bombarded by Mice DDoS traffic.…”

Scheduling in polling systems

“…For instance, variants of Shortest-Remaining-Processing-Time (SRPT) and Foreground-Background (FB) have been applied in many computer applications, e.g. web servers [8,17] and routers [14,15]. This growing focus on priority-based policies has led to an increasing number of theoretical studies of such policies, e.g.…”