ABSTRACT
Blind scheduling policies schedule tasks without knowledge of the tasks' remaining processing times. Existing blind policies, such as FCFS, PS, and LAS, have proven useful in network and operating system applications, but each policy has a separate, vastly differing description, leading to separate and distinct implementations. This paper presents the design and implementation of a configurable blind scheduler that contains a continuous, tunable parameter. By merely changing the value of this parameter, the scheduler's policy exactly emulates or closely approximates several existing standard policies. Other settings enable policies whose behavior is a hybrid of these standards. We demonstrate the practical benefits of such a configurable scheduler by implementing it into the Linux operating system. We show that we can emulate the behavior of Linux's existing, more complex scheduler with a single (hybrid) setting of the parameter. We also show, using synthetic workloads, that the best value for the tunable parameter is not unique, but depends on distribution of the size of tasks arriving to the system. Finally, we use our formulation of the configurable scheduler to contrast the behavior of various blind schedulers by exploring how various properties of the scheduler change as we vary our scheduler's tunable parameter.
- S. Aalto, U. Ayesta, and E. Nyberg-Oksanen. Two-level processor-sharing scheduling disciplines: Mean delay analysis. In Proc. ACM SIGMETRICS '04, pages 97--105, 2004. Google ScholarDigital Library
- J. Abate and W. Whitt. Limits and approximations for the M=G=1 LIFO waiting-time distribution. Operations Research Letters, 20:199--206, 1997. Google ScholarDigital Library
- Apache Software Foundation. Apache Httpd. http://httpd.apache.org/.Google Scholar
- N. Bansal. Achievable sojourn times by non-size based policies in a GI/GI/1 queue. Technical report, IBM Watson Research Center, 2004. http://www.research.ibm.com/people/n/nikhil/papers/blindnew.pdf.Google Scholar
- P. Barford and M. Crovella. Generating representative web workloads for network and server performance evaluation. In Proc. SIGMETRICS/PERFORMANCE '98, pages 151--160, November 1998. Google ScholarDigital Library
- G. Beekmans. Linux from scratch. http://www.linuxfromscratch.org/lfs/.Google Scholar
- B. Caprita, W. C. Chan, J. Nieh, C. Stein, and H. Zheng. Group ratio round-robin: O(1) proportional share scheduling for uniprocessor and multiprocessor systems. In Proc. USENIX '05, 2005. Google ScholarDigital Library
- E. Coffman, R. Muntz, and H. Trotter. Waiting time distribution for processor-sharing systems. Journal of the ACM, 17(1):123--130, 1970. Google ScholarDigital Library
- G. Fayolle, I. Mitrani, and R. Iasnogorodski. Sharing a processor among many job classes. Journal of the ACM, 27(3):519--532, 1980. Google ScholarDigital Library
- H. Feng, V. Misra, and D. Rubenstein. The PBS policy: Some properties and their proofs. Technical Report CUCS-015-07, Dept. of Computer Science, Columbia University, 2007.Google Scholar
- K. P. Gummadi, R. J. Dunn, S. Saroiu, S. D. Gribble, H. M. Levy, and J. Zahorjan. Measurement, modeling, and analysis of a peer-to-peer file-sharing workload. In Proc. SOSP, 2003. Google ScholarDigital Library
- M. Harchol-Balter, B. Schroeder, N. Bansal, and M. Agrawal. Size-based scheduling to improve web performance. ACM Trans. Comput. Syst., 21(2), 2003. Google ScholarDigital Library
- A. Kherani. Sojourn times in (discrete) time shared systems and their continuous time limits. In Proc. Valuetools '06, 2006. Google ScholarDigital Library
- A. Kherani and R. Nùñez-Queija. TCP as an implementation of age-based scheduling: Fairness and performance. In Proc. IEEE Infocom '06, 2006.Google ScholarCross Ref
- L. Kleinrock. Queueing Systems Volume I: Theory, Volume II: Computer Applications. John Wiley&Sons, 1975, 1976. Google ScholarDigital Library
- C. L. Liu and J. W. Layland. Scheduling algorithms for multiprogramming in a hard-real-time environment. Journal of the ACM, 20(1):46--61, 1973. Google ScholarDigital Library
- I. Mitrani and J. H. Hine. Complete parameterized families of job scheduling strategies. Acta Informatica, 8:61--73, 1977.Google ScholarDigital Library
- S. Padhy and A. A. Kherani. Tail equivalence for some time-shared systems. In Proc. Valuetools '06, 2006. Google ScholarDigital Library
- A. K. Parekh and R. G. Gallager. A generalized processor sharing approach to flow control in integrated services networks: The single-node case. IEEE/ACM Transactions on Networking, 1(3):344--357, 1993. Google ScholarDigital Library
- I. A. Rai, G. Urvoy-Keller, M. K. Vernon, and E. W. Biersack. Performance analysis of LAS-based scheduling disciplines in a packet switched network. In Proc. ACM SIGMETRICS '04, pages 106--117, 2004. Google ScholarDigital Library
- D. Raz, H. Levy, and B. Avi-Itzhak. A resource-allocation queueing fairness measure. In Proc. ACM SIGMETRICS '04, pages 130--141, 2004. Google ScholarDigital Library
- R. Righter, J. G. Shanthikumar, and G. Yamazaki. On extremal service disciplines in single-stage queueing systems. Journal of Applied Probability, (2):409--416, 1990.Google ScholarCross Ref
- M. Ruschitzka and R. Fabry. A unified approach to scheduling. Commun. of the ACM, 20(7):469--477, 1977. Google ScholarDigital Library
- L. Schrage. The queue M=G=1 with feedback to lower priority queues. Management Science, 13(7):466--474, 1967.Google ScholarDigital Library
- L. Schrage. A proof of the optimality of the shortest remaining processing time discipline. Operations Research, 16(3):687--690, 1968.Google ScholarDigital Library
- L. E. Schrage and L. W. Miller. The queue M=G=1 with the shortest remaining processing time discipline. Operations Research, 14(4):670--684, 1966.Google ScholarDigital Library
- A. Silberschatz, P. B. Galvin, and G. Gagne. Applied Operating Systems Concepts. John Wiley&Sons, 2000. Google ScholarDigital Library
- A. Wierman and M. Harchol-Balter. Classifying scheduling policies with respect to unfairness in an M=GI=1. In Proc. ACM SIGMETRICS '03, pages 238--249, 2003. Google ScholarDigital Library
- A. Wierman, M. Harchol-Balter, and T. Osogami. Nearly insensitive bounds on smart scheduling. In Proc. ACM SIGMETRICS '05, pages 205--216, 2005. Google ScholarDigital Library
Index Terms
- PBS: a unified priority-based scheduler
Recommendations
PBS: a unified priority-based scheduler
SIGMETRICS '07 Conference ProceedingsBlind scheduling policies schedule tasks without knowledge of the tasks' remaining processing times. Existing blind policies, such as FCFS, PS, and LAS, have proven useful in network and operating system applications, but each policy has a separate, ...
Classifying scheduling policies with respect to unfairness in an M/GI/1
It is common to evaluate scheduling policies based on their mean response times. Another important, but sometimes opposing, performance metric is a scheduling policy's fairness. For example, a policy that biases towards small job sizes so as to minimize ...
Two-level processor-sharing scheduling disciplines: mean delay analysis
Inspired by several recent papers that focus on scheduling disciplines for network flows, we present a mean delay analysis of Multilevel Processor Sharing (MLPS) scheduling disciplines in the context of M/G/1 queues. Such disciplines have been proposed ...
Comments