Preventing Large Sojourn Times Using SMART Scheduling

Recently, the so-called class of SMART scheduling policies has been introduced to formalize the common heuristic of “biasing toward small jobs.” We study the tail of the sojourn-time (response-time) distribution under both SMART policies and the foreground-background policy (FB) in the GI/GI/1 queue. We prove that these policies behave very well under heavy-tailed service times. Specifically, we show that the sojourn-time tail under all SMART policies and FB is similar to that of the service-time tail, up to a constant, which makes the SMART class superior to first-come-first-served (FCFS). In contrast, for light-tailed service times, we prove that the sojourn-time tail under FB and SMART is larger than that under FCFS. However, we show that the sojourn-time tail for a job of size y under FB and all SMART policies still outperforms FCFS as long as y is not too large.

[1]  Fabrice Guillemin,et al.  Tail asymptotics for processor-sharing queues , 2004, Advances in Applied Probability.

[2]  Ludmila Cherkasova.  Scheduling Strategy to improve Response Time for Web Applications , 1998, HPCN Europe.

[3]  Sem C. Borst,et al.  The impact of the service discipline on delay asymptotics , 2003, Perform. Evaluation.

[4]  Mor Harchol-Balter,et al.  Priority mechanisms for OLTP and transactional Web applications , 2004, Proceedings. 20th International Conference on Data Engineering.

[5]  Adam Wierman,et al.  Classifying scheduling policies with respect to higher moments of conditional response time , 2005, SIGMETRICS '05.

[6]  Bert Zwart,et al.  A large-deviations analysis of the GI/GI/1 SRPT queue , 2005, Queueing Syst. Theory Appl..

[7]  Ajay D. Kshemkalyani,et al.  SWIFT: scheduling in web servers for fast response time , 2003, Second IEEE International Symposium on Network Computing and Applications, 2003. NCA 2003..

[8]  Predrag R. Jelenkovic,et al.  Large Deviation Analysis of Subexponential Waiting Times in a Processor-Sharing Queue , 2003, Math. Oper. Res..

[9]  Ward Whitt,et al.  Asymptotics for M/G/1 low-priority waiting-time tail probabilities , 1997, Queueing Syst. Theory Appl..

[10]  Ronald W. Wolff,et al.  Stochastic Modeling and the Theory of Queues , 1989 .

[11]  M.F.M. Nuijens.  The Foreground-Background Queue , 2004 .

[12]  Murad S. Taqqu,et al.  On the Self-Similar Nature of Ethernet Traffic , 1993, SIGCOMM.

[13]  David L. Peterson.  Data Center I/O Patterns And Power Laws , 1996, Int. CMG Conference.

[14]  Sem C. Borst,et al.  Sojourn time asymptotics in processor-sharing queues , 2006, Queueing Syst. Theory Appl..

[15]  Walter Willinger,et al.  On the Self-Similar Nature of Ethernet Traffic ( extended version ) , 1995 .

[16]  Urtzi Ayesta,et al.  Two-level processor-sharing scheduling disciplines: mean delay analysis , 2004, SIGMETRICS '04/Performance '04.

[17]  Adam Wierman,et al.  Nearly insensitive bounds on SMART scheduling , 2005, SIGMETRICS '05.

[18]  J. George Shanthikumar,et al.  Scheduling Multiclass Single Server Queueing Systems to Stochastically Maximize the Number of Successful Departures , 1989, Probability in the Engineering and Informational Sciences.

[19]  Sلأren Asmussen,et al.  Applied Probability and Queues , 1989 .

[20]  J. Shanthikumar,et al.  Extremal properties of the FIFO discipline in queueing networks , 1992, Journal of Applied Probability.

[21]  Allen B. Downey,et al.  Evidence for long-tailed distributions in the internet , 2001, IMW '01.

[22]  Michel Mandjes,et al.  Large deviations of sojourn times in processor sharing queues , 2006, Queueing Syst. Theory Appl..

[23]  Mor Harchol-Balter,et al.  Implementation of SRPT Scheduling in Web Servers , 2000 .

[24]  Walter Willinger,et al.  On the self-similar nature of Ethernet traffic , 1993, SIGCOMM '93.

[25]  Søren Asmussen,et al.  Applied probability and queues, Second Edition , 2003, Applications of mathematics.

[26]  Guillaume Urvoy-Keller,et al.  Performance Models for LAS-based Scheduling Disciplines in a Packet Switched Network , 2004 .

[27]  Rudesindo Núñez-Queija,et al.  Queues with Equally Heavy Sojourn Time and Service Requirement Distributions , 2002 .

[28]  Gustavo de Veciana,et al.  Enhancing both network and user performance for networks supporting best effort traffic , 2004, IEEE/ACM Transactions on Networking.

[29]  Gennady Samorodnitsky,et al.  Activity periods of an infinite server queue and performance of certain heavy tailed fluid queues , 1999, Queueing Syst. Theory Appl..

[30]  Mor Harchol-Balter,et al.  Size-based scheduling to improve web performance , 2003, TOCS.

[31]  Adam Wierman,et al.  Classifying scheduling policies with respect to unfairness in an M/GI/1 , 2003, SIGMETRICS '03.

[32]  Paul Barford,et al.  Generating representative Web workloads for network and server performance evaluation , 1998, SIGMETRICS '98/PERFORMANCE '98.

[33]  M. Mandjes,et al.  SOJOURN TIMES IN THE M/G/1 FB QUEUE WITH LIGHT-TAILED SERVICE TIMES , 2003, Probability in the Engineering and Informational Sciences.