Abstract
Join procedures are cost intensive tasks even for in-memory databases. As today’s systems introduce more and more parallelism, intra-operator parallelization moves into focus. This chapter discusses possible schemes to parallelize a hash-join algorithm, as described in Chap. 19.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
C. Kim, T. Kaldewey, V.W. Lee, E. Sedlar, A.D. Nguyen, N. Satish, J. Chhugani, A. Di Blas, P. Dubey, Sort vs. Hash revisited: fast join implementation on modern multi-core CPUs, in VLDB, 2009
S. Manegold, P. Boncz, M. Kersten, Optimizing main-memory join on modern hardware. IEEE Trans. Knowl. Data Eng. 19, 412–426 (2008)
Author information
Authors and Affiliations
Corresponding author
Self Test Questions
Self Test Questions
-
1.
Parallelizing Hash-Join Phases
What is the disadvantage when the probing phase of a join algorithm is parallelized and the hashing phase is performed sequentially?
-
(a)
Sequentially performing the hashing phase introduces inconsistencies in the produced hash values
-
(b)
The algorithm still has a large sequential part that limits its potential to scale
-
(c)
The sequential hashing phase will run slower due to the large resource utilization of the parallel probing phase
-
(d)
The table has to be split into smaller parts, so that every core, which performs the probing, can finish.
-
(a)
Rights and permissions
Copyright information
© 2013 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Plattner, H. (2013). Parallel Join. In: A Course in In-Memory Data Management. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-36524-9_23
Download citation
DOI: https://doi.org/10.1007/978-3-642-36524-9_23
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-36523-2
Online ISBN: 978-3-642-36524-9
eBook Packages: Business and EconomicsBusiness and Management (R0)