The Gain of Overbooking

This paper analyzes the effect of overbooking for scheduling systems in a commercial environment. In this scenario each job is associated with a release time and a finishing deadline as well as a fee for a successful execution and a penalty for violating the deadline. The core idea is to exploit overestimations of required job execution times, providing an opportunity to aggressively schedule additional jobs. The proposed probabilistic scheduler is based on histories of job execution times, device failure rates, and penalties for SLA service violations. This paper includes a theoretical background and a mathematical model of the overbooking approach and a simulative evaluation with a synthetic workload on a single-processor system.

[1]  Warren Smith,et al.  Predicting Application Run Times Using Historical Information , 1998, JSSPP.

[2]  Rajkumar Buyya,et al.  Managing Cancellations and No-Shows of Reservations with Overbooking to Increase Resource Revenue , 2008, 2008 Eighth IEEE International Symposium on Cluster Computing and the Grid (CCGRID).

[3]  Marvin Rothstein,et al.  OR Forum - OR and the Airline Overbooking Problem , 1985, Oper. Res..

[4]  Dan Tsafrir,et al.  The Dynamics of Backfilling: Solving the Mystery of Why Increased Inaccuracy May Help , 2006, 2006 IEEE International Symposium on Workload Characterization.

[5]  Stephen A. Jarvis,et al.  Open Issues in Grid Scheduling , 2004 .

[6]  Achim Streit,et al.  Scheduling in HPC Resource Management Systems: Queuing vs. Planning , 2003, JSSPP.

[7]  Dror G. Feitelson,et al.  Utilization, Predictability, Workloads, and User Runtime Estimates in Scheduling the IBM SP2 with Backfilling , 2001, IEEE Trans. Parallel Distributed Syst..

[8]  Janakiram Subramanian,et al.  Airline Yield Management with Overbooking, Cancellations, and No-Shows , 1999, Transp. Sci..

[9]  Dan Tsafrir,et al.  Backfilling Using System-Generated Predictions Rather than User Runtime Estimates , 2007, IEEE Transactions on Parallel and Distributed Systems.

[10]  Richard Gibbons,et al.  A Historical Application Profiler for Use by Parallel Schedulers , 1997, JSSPP.

[11]  Timothy Roscoe,et al.  Resource overbooking and application profiling in shared hosting platforms , 2002, OSDI '02.

[12]  Larry Rudolph,et al.  Job Scheduling Strategies for Parallel Processing: IPPS '97 Workshop, Geneva, Switzerland, April 5, 1997, Proceedings , 1997 .

[13]  M. Siddiqui,et al.  Grid Capacity Planning with Negotiation-based Advance Reservation for Optimized QoS , 2006, ACM/IEEE SC 2006 Conference (SC'06).

[14]  Dmitry N. Zotkin,et al.  Job-length estimation and performance in backfilling schedulers , 1999, Proceedings. The Eighth International Symposium on High Performance Distributed Computing (Cat. No.99TH8469).

[15]  Dominic Battré,et al.  Increasing Fault-tolerance by Introducing Virtual Execution Environments. , 2008 .

[16]  Dan Tsafrir,et al.  Modeling User Runtime Estimates , 2005, JSSPP.

[17]  Uri Yechiali,et al.  On the Hotel Overbooking Problem---An Inventory System with Stochastic Cancellations , 1978 .

[18]  Dror G. Feitelson,et al.  Probabilistic Backfilling , 2007, JSSPP.

[19]  Guangwen Yang,et al.  Efficiently Rationing Resources for Grid and P2P Computing , 2004, NPC.

[20]  Achim Streit,et al.  Self-tuning job scheduling strategies for the resource management of HPC systems and computational grids , 2003 .

[21]  Dror G. Feitelson,et al.  Utilization and Predictability in Scheduling the IBM SP2 with Backfilling , 1998, Proceedings of the First Merged International Parallel Processing Symposium and Symposium on Parallel and Distributed Processing.

[22]  Dror G. Feitelson,et al.  Improved Utilization and Responsiveness with Gang Scheduling , 1997, JSSPP.