Abstract— One challenge in data mining come from data sorting problem which needs enormous computation power when it comes to big data. While many machines and methods have been developed to tackle this problem, most solutions are either too expensive or too complex to be implemented in typical companies, non-profit organization, and individual researchers. One technique to tackle these difficulties is using volunteer computing model. This paper suggests one adaptable sorting model that can be used in such volunteer computing environment. The model is generic so it can be adapted directly in the existing application without the need to install additional agent or server. The model was tested using 750,000 random integer data. The tests were run 300 times each employing four nodes and three nodes with different number of buckets and number of data in a bucket, apparently those are faster than sequential merge sort.

Index Terms— Data mining, parallel, sorting, volunteer computing.

Y. E. Soelistio is with the Faculty of ICT, Universitas Multimedia Nusantara, Tangerang, Indonesia (e-mail: yustinus.eko@ umn.ac.id).
P. Yugopuspito is with Research and Development Computer Lab., Universitas Pelita Harapan (e-mail: pujianto.yugopuspito@uph.edu).

Cite: Yustinus Eko Soelistio and Pujianto Yugopuspito, " Simple Parallel Sorting Model for Embarrassingly Parallelism Environment," International Journal of e-Education, e-Business, e-Management and e-Learning vol. 4, no. 1, pp. 52-54, 2014.