Thomas Sandholm
Kevin Lai
ABSTRACT
We present a system for allocating resources in shared data and compute clusters that improves MapReduce job scheduling in three ways. First, the system uses regulated and user-assigned priorities to offer different service levels to jobs and users over time. Second, the system dynamically adjusts resource allocations to fit the requirements of different job stages. Finally, the system automatically detects and eliminates bottlenecks within a job. We show experimentally using real applications that users can optimize not only job execution time but also the cost-benefit ratio or prioritization efficiency of a job using these three strategies. Our approach relies on a proportional share mechanism that continuously allocates virtual machine resources. Our experimental results show a 11-31% improvement in completion time and 4-187% improvement in prioritization efficiency for different classes of MapReduce jobs. We further show that delay intolerant users gain even more from our system.
- K. Arrow. Aspects of the theory of risk-bearing. Helsinki: Yrjo Jahnsson Lectures, 1965.Google Scholar
- A. AuYoung, L. Grit, J. Wiener, and J. Wilkes. Service contracts and aggregate utility functions. In Proceedings of the IEEE International Symposium on High Performance Distributed Computing (HPDC), June 2006.Google ScholarCross Ref
- R. Avnur and J. M. Hellerstein. Eddies: Continuously adaptive query processing. In ACM SIGMOD: International Conference on Management of Data, 2007. Google ScholarDigital Library
- P. Barham, B. Dragovic, K. Fraser, S. Hand, T. Harris, A. Ho, R. Neugebauer, I. Pratt, and A. Warfield. Xen and the art of virtualization. In Proceedings of the ACM Symposium on Operating Systems Principles, 2003. Google ScholarDigital Library
- R.E. Bryant. Data-intensive supercomputing: The case for DISC. Technical Report CMU-CS-07-128, Carnegie Mellon University, 2007.Google Scholar
- K. Cardona, J. Secretan, M. Georgiopoulos, and G. Anagnostopoulos. A grid based system for data mining using MapReduce. Technical Report TR-2007-02, AMALTHEA, 2007.Google Scholar
- B.N. Chun, P. Buonadonna, A. AuYoung, C. Ng, D.C. Parkes, J. Shneidman, A.C. Snoeren, and A. Vahdat. Mirage: A microeconomic resource allocation system for SensorNet testbeds. In Proceedings of the 2nd IEEE Workshop on Embedded Networked Sensors, 2005. Google ScholarDigital Library
- B.N. Chun and D.E. Culler. Market-based proportional resource sharing for clusters. Technical Report CSD-1092, University of California at Berkeley, Computer Science Division, January 2000. Google ScholarDigital Library
- B.N. Chun and D.E. Culler. User-centric performance analysis of market-based cluster batch schedulers. In Proceedings of the 2nd IEEE International Symposium on Cluster Computing and the Grid, 2002. Google ScholarDigital Library
- J. Dean and S. Ghemawat. MapReduce: Simplified data processing on large clusters. In Symposium on Operating System Design and Implementation, 2004. Google ScholarDigital Library
- M. Feldman, K. Lai, and L. Zhang. A price-anticipating resource allocation mechanism for distributed shared clusters. In Proceedings of the ACM Conference on Electronic Commerce, 2005. Google ScholarDigital Library
- http://hadoop.apache.org/core, 2008.Google Scholar
- G. Hardin. The tragedy of the commons. Science, 162:1243--1248, 1968.Google ScholarCross Ref
- B. He, W. Fang, Q. Luo, N.K. Govindaraju, and T. Wang. Mars: a MapReduce framework on graphics processors. In PACT '08: Proceedings of the 17th international conference on Parallel architectures and compilation techniques, pages 260--269, New York, NY, USA, 2008. ACM. Google ScholarDigital Library
- D. Irwin, J. Chase, and L. Grit. Balancing risk and reward in market-based task scheduling. In International Symposium on High Performance Distributed Computing, 2004. Google ScholarDigital Library
- M. Isard, M. Budiu, Y. Yu, A. Birrell, and D. Fetterly. Dryad: distributed data-parallel programs from sequential building blocks. In EuroSys '07: Proceedings of the 2nd ACM SIGOPS/EuroSys European Conference on Computer Systems 2007, pages 59--72, New York, NY, USA, 2007. ACM. Google ScholarDigital Library
- P.R. Jelenkovic, X. Kang, and J. Tan. Adaptive and scalable comparison scheduling. In ACM SIGMETRICS'07: International Conference on Measurement and Modeling of Computer Systems, pages 215--226, 2007. Google ScholarDigital Library
- E. Jensen, C. Locke, and H. Tokuda. A time-driven scheduling model for real-time operating systems. In IEEE Real-Time Systems Symposium , pages 112--122, 1985.Google Scholar
- K. Lai, L. Rasmusson, E. Adar, S. Sorkin, L. Zhang, and B.A. Huberman. Tycoon: an implemention of a distributed market-based resource allocation system. Multiagent and Grid Systems, 1(3):169--182, Aug. 2005. Google ScholarDigital Library
- D. Logothetis and K. Yocum. Ad-hoc data processing in the cloud. Proc. VLDB Endow., 1(2):1472--1475, 2008. Google ScholarDigital Library
- N. Moroney, P. Obrador, and G. Beretta. Lexical image processing. In Proceedings of the 16th IS&T/SID Color Imaging Conference, pages 268--273, 2008.Google Scholar
- C. Olston. Pig: Web-scale processing. http://www.cs.cmu.edu/~olston/pig.ppt, 2008.Google Scholar
- C. Olston, B. Reed, A. Silberstein, and U. Srivastava. Automatic optimization of parallel dataflow programs. In USENIX Annual Technical Conference, 2008. Google ScholarDigital Library
- C. Olston, B. Reed, U. Srivastava, R. Kumar, and A. Tomkins. Pig latin: A not-so-foreign language for data processing. In SIGMOD 2008: ACM SIGMOD/PODS Conference, 2008. Google ScholarDigital Library
- C.H. Papadimitriou. Algorithms, games, and the Internet. In Symposium on Theory of Computing, 2001. Google ScholarDigital Library
- L. Peterson, T. Anderson, D. Culler, and T. Roscoe. Blueprint for Introducing Disruptive Technology into the Internet. In First Workshop on Hot Topics in Networking, 2002.Google Scholar
- R. Pike, S. Dorward, R. Griesemer, and S. Quinlan. Interpreting the data: Parallel analysis with Sawzall. Scientific Programming Journal Special Issue on Grids and Worldwide Computing Programming Models and Infrastructure, 13(4):227--298, 2003. Google ScholarDigital Library
- F.I. Popovici and J. Wilkes. Profitable services in an uncertain world. In SC05: Proceedings of Supercomputing, 2005. Google ScholarDigital Library
- J. Pratt. Risk aversion in the small and in the large. Econometrica, 32:122--136, 1964.Google ScholarCross Ref
- C. Ranger, R. Raghuraman, A. Penmetsa, G. Bradski, and C. Kozyrakis. Evaluating MapReduce for multi-core and multiprocessor systems. In HPCA'07: IEEE 13th International Symposium on High Performance Computer Architecture , pages 13--24, 2007. Google ScholarDigital Library
- T. Sandholm. Statistical methods for computational markets.Doctoral Thesis ISRN SU-KTH/DSV/R-08/6-SE. Royal Institute of Technology, Stockholm, 2008.Google Scholar
- T. Sandholm and K. Lai. A statistical approach to risk mitigation in computational markets. In Proceedings of the ACM International Symposium on High Performance Distributed Computing (HPDC), June 2007. Google ScholarDigital Library
- T. Sandholm, K. Lai, J. Andrade, and J. Odeberg. Market-based resource allocation using price prediction in a high performance computing grid for scientific applications. In Proceedings of the IEEE International Symposium on High Performance Distributed Computing (HPDC), June 2006.Google ScholarCross Ref
- P.G. Selinger, M.M. Astrahan, D.D. Chamberlin, R.A. Lorie, and T.G. Price. Access path selection in a relational database management system. In SIGMOD 1979: ACM SIGMOD International Conference on the Management of Data, 1979. Google ScholarDigital Library
- M.A. Shah, J.M. Hellerstein, S. Chandrasekaran, and M.J. Franklin. Flux: An adaptive partitioning operator for continuous query systems. Technical Report UCB/CSD-2-1205, U. C. Berkley, 2002. Google ScholarDigital Library
- M. Stonebraker, P.M. Aoki, W. Litwin, A. Pfeffer, A. Sah, J. Sidell, C. Staelin, and A. Yu. Mariposa: a wide-area distributed database system. The VLDB Journal, 5(1):048--063, 1996. Google ScholarDigital Library
- S.V. Valvag and D. Johansen. Oivos: Simple and efficient distributed data processing. High Performance Computing and Communications, 2008. HPCC '08. 10th IEEE International Conference on, pages 113--122, Sept. 2008. Google ScholarDigital Library
- M. Wachs, M. Abd-El-Malek, E. Thereska, and G.R. Ganger. Argon: performance insulation for shared storage servers. In FAST'07: 5th USENIX Conference on File and Storage Technologies, 2007. Google ScholarDigital Library
- C.A. Waldspurger and W.E. Weihl. Lottery scheduling: Flexible proportional-share resource management. In Operating Systems Design and Implementation, pages 1--11, 1994. Google ScholarDigital Library
- A. Wierman and M. Nuyens. Scheduling despite inexact job-size information. In ACM SIGMETRICS'08: International Conference on Measurement and Modeling of Computer Systems, pages 25--36, 2008. Google ScholarDigital Library
- J. Wolfe, A. Haghighi, and D. Klein. Fully distributed EM for very large datasets. In ICML '08: Proceedings of the 25th international conference on Machine learning, pages 1184--1191, New York, NY, USA, 2008. ACM. Google ScholarDigital Library
- M. Zaharia, A. Konwinski, A.D. Joseph, R. Katz, and I. Stoica. Improving MapReduce performance in heterogeneous environments. In OSDI'08: 8th USENIX Symposium on Operating Systems Design and Implementation, 2008. Google ScholarDigital Library
- L. Zhang. The efficiency and fairness of a fixed budget resource allocation game. In International Colloquium on Automata, Languages and Programming, pages 485--496, 2005. Google ScholarDigital Library
Index Terms
- MapReduce optimization using regulated dynamic prioritization
Recommendations
MapReduce optimization using regulated dynamic prioritization
SIGMETRICS '09We present a system for allocating resources in shared data and compute clusters that improves MapReduce job scheduling in three ways. First, the system uses regulated and user-assigned priorities to offer different service levels to jobs and users over ...
Multi-policy-aware MapReduce resource allocation and scheduling for smart computing cluster
When a user submit a MapReduce job in the smart computing cluster, we first need to allocate cluster resource for the job. It is widely concerned that how to save time and resource costs to provide users with computing capacity and services. Here, we ...
ARIA: automatic resource inference and allocation for mapreduce environments
ICAC '11: Proceedings of the 8th ACM international conference on Autonomic computingMapReduce and Hadoop represent an economically compelling alternative for efficient large scale data processing and advanced analytics in the enterprise. A key challenge in shared MapReduce clusters is the ability to automatically tailor and control ...
Comments