DocumentCode :
2590745
Title :
An improved selection-based parallel range-join algorithm in hypercubes
Author :
Shen, Hong
Author_Institution :
Sch. of Comput. & Inf. Technol., Griffith Univ., Nathan, Qld., Australia
fYear :
1994
fDate :
5-8 Sep 1994
Firstpage :
65
Lastpage :
72
Abstract :
The range-join of two sets R and S is the set that contains all tuples (r, s) satisfying e1⩽|r-s|⩽e2, r∈R and s∈S. For computing the range-join of R and S in a hypercube of p processors, this paper presents an improved selection-based parallel algorithm which reduces the local memory from O(n) repaired in the previous algorithm to O(m+n/p), where |R|=m, |S|=n and p⩽max{m,n}. The new algorithm also reduces the best-case time complexity from O(m/p log2 p+n/p log m) of the previous result to O(m+n/p log2p) when m⩾plog, while maintaining the cost optimality in the worst case. Unlike the previous algorithm, our algorithm works by selecting the median of RUS to evenly partition the whole data set for divide-and-conquer join in the next phase. We present an upper bound of time complexity of the algorithm in the general case and show that the best-case time complexity of the algorithm is better than permutation-based range-join when n⩾plogp+1
Keywords :
computational complexity; database theory; divide and conquer methods; hypercube networks; merging; parallel algorithms; set theory; best-case time complexity; cost optimality; divide-and-conquer join; hypercubes; improved selection-based parallel range-join algorithm; permutation-based range-join; selection-based parallel algorithm; Australia; Costs; Databases; Hypercubes; Information management; Information technology; Memory management; Parallel algorithms; Upper bound; Whales;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
EUROMICRO 94. System Architecture and Integration. Proceedings of the 20th EUROMICRO Conference.
Conference_Location :
Liverpool
Print_ISBN :
0-8186-6430-4
Type :
conf
DOI :
10.1109/EURMIC.1994.390405
Filename :
390405
Link To Document :
بازگشت