ABSTRACT
In the decision support queries which manipulate large data volumes, it is frequent that a query constituted by several joins can not be computed completely in memory. In this paper, we propose three strategies allowing to assign the memory of a shared-nothing parallel architecture to operation clones of a query. The performance evaluation of the three strategies shows that the strategies which favor the operation clones using a lot of memory obtain a better response time than the strategy which favors the clones using little memory. The main contribution of this paper is to take into account the available memory sizes on every processor and to avoid allotting the same processor to two operation clones that must run in parallel.
- Bonneau S., Hameurlain A. Hybrid Simultaneous Scheduling and Mapping in SQL Multi-Query Parallelization. Proc. of the 10 th intl. Conf. on Database and Expert Systems Applications, DEXA, LNCS 1677, Florence, Sept. 1999, pp. 88-99.]] Google ScholarDigital Library
- Bouganim L. et al. Memory-Adaptive Scheduling for Large Query Execution. Proc. of the 8th intl. Conf. CIKM, Bethesda, Maryland, Nov. 1998, pp. 105-115.]] Google ScholarDigital Library
- Chen M. S. et al. Scheduling and Processor Allocation for Parallel Execution of Multi-join Queries. Proc. 8th Intl. Conf. Data Eng., Tempe, Feb. 1992, pp. 58-67.]] Google ScholarDigital Library
- Chen M. S. et al. Using Segmented Right-Deep Trees for the Execution of Pipelined Hash Joins. Proc. of the 18th VLDB Conf., Vancouver, Aug. 1992, pp. 15-26.]] Google ScholarDigital Library
- Chou H., DeWitt D. An Evaluation of Buffer Management Stratégies for Relational Database Systems. Proc. of the 11th Intl VLDB Conf., Stockholm, Aug. 1985, pp. 127-141.]]Google Scholar
- Conway R. W. et al. The Theory of Scheduling. Ed. Addition-Wesley, 1967.]]Google Scholar
- Ganguly S., and al. Efficient and Accurate Cost Models for Parallel Query Optimization. Symposium in Principles of Database Systems PODS'96 , june 1996, pp. 172-182.]] Google ScholarDigital Library
- Garofalakis M. N., Ioannidis Y. E. Parallel Query Scheduling and Optimization with Time- and Space- Shared Resources, Proc. of the 23rd VLDB Conf., Athens, 1997, pp. 296-305.]] Google ScholarDigital Library
- Hameurlain A., Morvan F. Scheduling and Mapping for Parallel Execution of Extended SQL Queries. Proc. of the 4th intl. Conf. CIKM, Baltimore, Nov-Dec 1995, pp. 197-204.]] Google ScholarDigital Library
- Helal A. et al. Dynamic Data Reallocation for Skew Management in Shared-Nothing Parallel Databases, Distributed and Parallel Databases, Vol. 5, 1997. pp. 271-288.]] Google ScholarDigital Library
- Mehta M., DeWitt D. Dynamic Memory Allocation for Multiple-Query Workload. Proc. of the 19th Intl VLDB Conf., Dublin, Aug. 1993, pp. 354-367.]] Google ScholarDigital Library
- Mehta M., DeWitt D. Data Placement in Shared-Nothing Parallel Database Systems. The VLDB Journal, 1997, N°6, pp. 53-72.]] Google ScholarDigital Library
- Morvan F., Hameurlain A. Dynamic Memory Allocation Strategies for Parallel Query Execution. Technical report, IRIT/2001-14-R, June 2001.]]Google Scholar
- Nag B., DeWitt D. Memory Allocation Strategies for Complex Decision Support Queries. Proc. of the 8th intl. Conf. CIKM, Bethesda, Nov. 1998, pp. 116-123.]] Google ScholarDigital Library
- Ng R. et al. Flexible Buffer Allocation Based on Marginal Gains. Proc. of the ACM SIGMOD'91 Conf. on Management of Data, Denver, May 1991, pp. 387-396.]] Google ScholarDigital Library
- Schneider D., DeWitt D. A performance Evaluation of Four Parallel Join Algorithms in a Shared-Nothing Multiprocessor Environment. Proc. ACM SIGMOD Conf. on Management of Data, Portland, June 1989, pp. 110-121.]] Google ScholarDigital Library
- Schneider D., DeWitt D. Tradeoffs in Processing Complex Join Queries via Hashing in Multiprcessor Database Machines. Proc. of the 16th VLDB Conf., Brisbane, Aug. 1990, pp. 469-480.]] Google ScholarDigital Library
- Shekita E. J. et al. Multi-Join Optimization for Symmetric Multiprocessors. 19th Intl. Conf. on VLDB, Dublin, Aug. 1993, pp. 479-492.]] Google ScholarDigital Library
Index Terms
- Dynamic memory allocation strategies for parallel query execution
Recommendations
Load Balancing for Parallel Query Execution on NUMA Multiprocessors
To scale up to high-end configurations, shared-memory multiprocessors are evolving towards Non Uniform Memory Access (NUMA) architectures. In this paper, we address the central problem of load balancing during parallel query execution in NUMA ...
Performance analysis of "Groupby-After-Join" query processing in parallel database systems
Queries containing aggregate functions often combine multiple tables through join operations. This query is subsequently called "Groupby-Join". There is a special category of this query whereby the group-by operation can only be performed after the join ...
Fast dynamic memory allocator for massively parallel architectures
GPGPU-6: Proceedings of the 6th Workshop on General Purpose Processor Using Graphics Processing UnitsDynamic memory allocation in massively parallel systems often suffers from drastic performance decreases due to the required global synchronization. This is especially true when many allocation or deallocation requests occur in parallel. We propose a ...
Comments