A Distributed Join Algorithm between Two Fragmented Relations in Distributed Database Systems

Jae Moon LEE  Jong Soo PARK  Myunghwan KIM  

IEICE TRANSACTIONS (1976-1990)   Vol.E73   No.7   pp.1225-1232
Publication Date: 1990/07/25
Online ISSN: 
Print ISSN: 0000-0000
Type of Manuscript: PAPER
Category: Databases

Full Text: PDF>>
Buy this Article

Minimizing intersite data transmissions is important for the join operation between two fragmented relations in distributed databases. In this paper, we examine communication costs of distributed joins when data redundancy and semantic information associated with fragments are not considered. We use a bit vector filtering technique to minimize unnecessary data transmissions in a computer network. The procedure of a distributed join algorithm is proposed and its communication cost is analyzed with filtering error in hashing. We performed computational experiments to show effect of the communication cost of the proposed algorithm in the number of sites and the semijoin selectivities. The experiments also include performance comparison between the proposed algorithm and another semijoin method which recently appeared in the literature. The results show that the communication cost of the proposed algorithm is smaller than that of the semijoin method the fragmented relations.