Large volumes of data processing for resource optimization using 3D printers in distributed environments

Abstract

As time goes, it is expected the amount of data will increase explosively, which needs better data processing technology for resource optimization as well as data processing algorithm. 3D printing is a good example to need to process heavy data. To generate 3D objects by a 3D printer, 3D model is first generated, and then converted to operations which the printer can interpret. 3D object surface is represented by facets and the number highly depends on the height or precision of the output, which much more conversion needs time from 3D model to printer operations. In this paper, we first devise a conversion algorithm on distributed systems and then adjust it on hadoop systems. The motivation of our distributed algorithm or hadoop-based algorithm is that the conversion from a facet does not affect the other facets. In case of distributed algorithm, the conversion process is divided into two phase: first, grouping facets according to z-axis value, and second, generating g-code from facets which are ordered by z-value. Those phases will be done in distributed manner. To do so, there are n+1 processing nodes and the first one n0 plays a role of coordinator. Apache hadoop is a software framework to support distributed processing for large data set and its application range gets widening. The second, hadoop-based algorithm proceeds to step 4: pre-processing, mapping, mapping, shuffling, and reducing. Finally, through the simulation works, we come to know the transformation can be done well in proposed environments. © 2020Alpha Publishers. All rights reserved.