For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Discreet Method to Match Safe Site-Pairs in Short Computation Time for Risk-Aware Data Replication
Takaki NAKAMURA Shinya MATSUMOTO Hiroaki MURAOKA
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2015/08/01
Online ISSN: 1745-1361
Type of Manuscript: PAPER
Category: Dependable Computing
remote replication, disaster recovery, distributed storage, integer programming problem, availability,
Full Text: PDF>>
Risk-aware Data Replication (RDR), which replicates data at primary sites to nearby safe backup sites, has been proposed to mitigate service disruption in a disaster area even after a widespread disaster that damages a network and a primary site. RDR assigns a safe backup site to a primary site while considering damage risk for both the primary site and the backup candidate site. To minimize the damage risk of all site-pairs the Integer Programing Problem (IPP), which is a mathematical optimization problem, is applied. A challenge for RDR is to choose safe backup sites within a short computation time even for a huge number of sites. As described in this paper, we propose a Discreet method for RDR to surmount this hurdle. The Discreet method first judges the backup sites of a potentially unsafe primary site and avoids assigning a very safe primary site with a very safe backup site. We evaluated the computation time for site-paring and the data availability in the cases of Earthquake and Tsunami using basic disaster simulations. We confirmed that the computation rate of the proposed method is more than 1000 times faster than the existing method when the number of sites is greater than 1000. We also confirmed the data availability of the proposed method; it provides almost equal rates to existing methods of strict optimization. These results mean that the proposed method makes RDR more practical for massively multiple sites.