Abstract
Service demand burstiness, or serial correlations in resource service demands, has previously been shown to have an adverse impact on system performance metrics such as response time. This article proposes AMIR, an analytic framework to characterize burstiness and identify strategies to reduce its impact on performance. AMIR considers an overtake-free system model consisting of multiple queues that service multiple classes of sessions, i.e., sequences of requests. Given the per-class service demand distributions and number of sessions belonging to each class, AMIR can identify an ordering of sessions, i.e., a schedule, that minimizes burstiness at the bottleneck. Hence, it is likely to improve system responsiveness metrics, including mean session wait time and total schedule processing time. To characterize burstiness, the technique uses an order O schedule burstiness metric βO representing the mean probability that O + 1 consecutive sessions in the schedule have resource demands at the bottleneck greater than the mean bottleneck demand of the schedule. For a given O, AMIR uses Integer Linear Programming (ILP) to identify schedules that progressively minimize βi ∀i ∈ {1, … O}. We conduct an extensive simulation study to provide insights on the conditions under which such schedules can improve system responsiveness. These results show that schedules derived from AMIR can significantly outperform those derived from baseline policies such as Shortest Job First (SJF) and random scheduling when session classes are dissimilar from one another in terms of their service demand distributions. Furthermore, minimizing for higher orders of schedule burstiness is most critical when the bottleneck is heavily utilized and when the service demands of a workload are highly variable. For the system model that we consider, we are not aware of other techniques that are designed to analytically derive insights on the performance impact of high-order service demand burstiness.
- Kenneth R. Baker and Dominik Altheimer. 2012. Heuristic solution methods for the stochastic flow shop problem. European Journal of Operational Research 216, 1, 172--177. DOI:https://doi.org/10.1016/j.ejor.2011.07.021Google ScholarCross Ref
- Dimitris Bertsimas and Georgia Mourtzinou. 1996. A unified method to analyze overtake free queueing systems. Advances in Applied Probability 28, 02, 588--625. http://journals.cambridge.org/abstract_S0001867800048631.Google ScholarCross Ref
- Giuliano Casale, Amir Kalbasi, Diwakar Krishnamurthy, and Jerry Rolia. 2010a. Automatically generating bursty benchmarks for multitier systems. ACM SIGMETRICS Performance Evaluation Review 37, 3, 32--37. DOI:https://doi.org/10.1145/1710115.1710122Google ScholarDigital Library
- Giuliano Casale, Amir Kalbasi, Diwakar Krishnamurthy, and Jerry Rolia. 2012. BURN: Enabling Workload Burstiness in Customized Service Benchmarks. IEEE Transactions on Software Engineering 38, 4, 778--793. DOI:https://doi.org/10.1109/TSE.2011.58Google ScholarDigital Library
- Giuliano Casale, Ningfang Mi, and Evgenia Smirni. 2010b. Model-driven system capacity planning under workload burstiness. IEEE Transactions on Computers 59, 1, 66--80. DOI:https://doi.org/10.1109/TC.2009.135Google ScholarDigital Library
- Tai Chiu Edwin Cheng and Helmut G. Kahlbacher. 1991. A proof for the longest-job-first policy in one-machine scheduling. Naval Research Logistics 38, 5, 715--720. DOI:https://doi.org/10.1002/1520-6750(199110)38:5<715::AID-NAV3220380506>3.0.CO;2-6Google ScholarCross Ref
- Gianfranco Ciardo, Alma Riska, and Evgenia Smirni. 2001. EQUILOAD: A load balancing policy for clustered web servers. Performance Evaluation 46, 2, 101--124. http://www.sciencedirect.com/science/article/pii/S0166531601000499.Google ScholarDigital Library
- Alan Cobham. 1954. Priority assignment in waiting line problems. Journal of the Operations Research Society of America 2, 1, 70--76. http://www.jstor.org/stable/166539.Google ScholarCross Ref
- Hans Daduna. 1982. Passage times for overtake-free paths in Gordon--Newell networks. Advances in Applied Probability 14, 03, 672--686. http://journals.cambridge.org/abstract_S000186780002070X.Google ScholarCross Ref
- Emanuel Falkenauer and S. Bouffouix. 1991. A genetic algorithm for job shop. In International Conference on Robotics and Automation Proceedings. IEEE, 824--829, vol. 1. DOI:https://doi.org/10.1109/ROBOT.1991.131689Google Scholar
- Michael R. Garey, David S. Johnson, and Ravi Sethi. 1976. The complexity of flowshop and jobshop scheduling. Mathematics of Operations Research 1, 2, 117--129. http://pubsonline.informs.org/doi/abs/10.1287/moor.1.2.117.Google ScholarDigital Library
- Michel Gourgand, Nathalie Grangeon, and Sylvie Norre. 2003. A contribution to the stochastic flow shop scheduling problem. European Journal of Operational Research 151, 2, 415--433. DOI:https://doi.org/10.1016/S0377-2217(02)00835-4Google ScholarCross Ref
- Michael Grant and Stephen Boyd. 2008. Graph implementations for nonsmooth convex programs. In Recent Advances in Learning and Control, V. Blondel, S. Boyd, and H. Kimura (Eds.). Springer-Verlag Limited, 95--110. http://stanford.edu/boyd/graph_dcp.html.Google Scholar
- Michael Grant and Stephen Boyd. 2014. CVX: MATLAB Software for Disciplined Convex Programming, version 2.1. Retrieved October 28, 2019 from http://cvxr.com/cvx.Google Scholar
- Varun Gupta, Michelle Burroughs, and Mor Harchol-Balter. 2010. Analysis of scheduling policies under correlated job sizes. Performance Evaluation 67, 11, 996--1013. DOI:https://doi.org/10.1016/j.peva.2010.08.010Google ScholarDigital Library
- Gurobi Optimization. 2016. Gurobi Optimizer Reference Manual. Retrieved October 28, 2019 from https://www.gurobi.com/wpcontent/plugins/hddocumentations/content/pdf/referencemanual.pdf.Google Scholar
- Mor Harchol-Balter, Mark E. Crovella, and Cristina D. Murta. 1999. On choosing a task assignment policy for a distributed server system. Journal of Parallel and Distributed Computing 59, 2, 204--228. DOI:https://doi.org/10.1006/jpdc.1999.1577Google ScholarDigital Library
- Intel. 2016. Intel QuickSync Video and FFmpeg. Retrieved October 28, 2019 from https://www.intel.com/content/dam/www/public/us/en/documents/white-papers/cloud-computing-quicksync-video-ffmpeg-white-paper.pdf.Google Scholar
- S. M. Johnson. 1954. Optimal two- and three-stage production schedules with setup times included. Naval Research Logistics Quarterly 1, 1, 61--68. DOI:https://doi.org/10.1002/nav.3800010110Google ScholarCross Ref
- Amir Kalbasi, Diwakar Krishnamurthy, Jerry Rolia, and Sharad Singhal. 2014. Simulation by example for complex systems. In Proceedings of the 2014 Winter Simulation Conference (WSC’14). IEEE Press, Piscataway, NJ, 974--985. http://dl.acm.org/citation.cfm?id=2693848.2693977Google ScholarDigital Library
- Hui Li and Michael Muskulus. 2007. Analysis and modeling of job arrivals in a production grid. ACM SIGMETRICS Performance Evaluation Review 34, 4, 59--70. http://dl.acm.org/citation.cfm?id=1243402Google ScholarDigital Library
- Ningfang Mi, Giuliano Casale, Ludmila Cherkasova, and Evgenia Smirni. 2008. Burstiness in multi-tier applications: Symptoms, causes, and new models. In Middleware 2008, Lecture Notes in Computer Science, Valrie Issarny and Richard Schantz (Eds.). Springer, Berlin, 265--286. http://link.springer.com.ezproxy.lib.ucalgary.ca/chapter/10.1007/978-3-540-89856-6_14 DOI:10.1007/978-3-540-89856-6_14.Google ScholarDigital Library
- Ningfang Mi, Giuliano Casale, and Evgenia Smirni. 2012. ASIdE: Using autocorrelation-based size estimation for scheduling bursty workloads. IEEE Transactions on Network and Service Management 9, 2, 198--212. DOI:https://doi.org/10.1109/TNSM.2012.041712.100073Google ScholarCross Ref
- Ningfang Mi, Qi Zhang, Alma Riska, Evgenia Smirni, and Erik Riedel. 2007. Performance impacts of autocorrelated flows in multi-tiered systems. Performance Evaluation 64, 9--12, 1082--1101. DOI:https://doi.org/10.1016/j.peva.2007.06.016Google ScholarDigital Library
- MOSEK ApS. 2015. The MOSEK Optimization Toolbox for MATLAB Manual. Version 7.1 (Revision 28). Retrieved October 28, 2019 from http://docs.mosek.com/7.1/toolbox/index.html.Google Scholar
- Tadahiko Murata, Hisao Ishibuchi, and Hideo Tanaka. 1996a. Genetic algorithms for flowshop scheduling problems. Computers 8 Industrial Engineering 30, 4, 1061--1071. DOI:https://doi.org/10.1016/0360-8352(96)00053-8Google Scholar
- Tadahiko Murata, Hisao Ishibuchi, and Hideo Tanaka. 1996b. Multi-objective genetic algorithm and its applications to flowshop scheduling. Computers 8 Industrial Engineering 30, 4, 957--968. DOI:https://doi.org/10.1016/0360-8352(96)00045-9Google Scholar
- Shrikant S. Panwalkar and Wafik Iskander. 1977. A survey of scheduling rules. Operations Research 25, 1, 45--61. http://www.jstor.org/stable/169546.Google ScholarDigital Library
- Thomas E. Phipps and W. R. Van Voorhis. 1956. Machine repair as a priority waiting-line problem. Operations Research 4, 1, 76--86. http://www.jstor.org/stable/167518.Google ScholarDigital Library
- Vili Podgorelec and Peter Kokol. 1997. Genetic algorithm based system for patient scheduling in highly constrained situations. Journal of Medical Systems 21, 6, 417--427. DOI:https://doi.org/10.1023/A:1022828414460Google ScholarDigital Library
- Alma Riska and Erik Riedel. 2006. Long-range dependence at the disk drive level. In Third International Conference on the Quantitative Evaluation of Systems (QEST’06). IEEE, 41--50. http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1703988.Google ScholarDigital Library
- Bianca Schroeder and Mor Harchol-Balter. 2006. Web servers under overload: How scheduling can help. ACM Transactions on Internet Technolology 6, 1, 20--52. DOI:https://doi.org/10.1145/1125274.1125276Google ScholarDigital Library
Index Terms
- AMIR: Analytic Method for Improving Responsiveness by Reducing Burstiness
Recommendations
Scheduling of deteriorating jobs with release dates to minimize the maximum lateness
In this paper, we consider the problem of scheduling n deteriorating jobs with release dates on a single (batching) machine. Each job's processing time is a simple linear function of its starting time. The objective is to minimize the maximum lateness. ...
Primary-secondary bicriteria scheduling on identical machines to minimize the total completion time of all jobs and the maximum T-time of all machines
In this paper, we study a new primary-secondary bicriteria scheduling problem on identical machines. The primary objective is to minimize the total completion time of all jobs and the secondary objective is to minimize the maximum T-time of all machines,...
Comments