DefinePK

Abstract

In the world where collaboration among multidisciplinary services is necessary for continuous improvements, scientific computing, unfortunately, lacks the infrastructure that allows for this type of collaboration. As such, this paper recommends guidelines of service-oriented, multi-layered designs through the exemplifications of models, service-oriented architecture, architectural layers, service compositions, and computational stages. The objective is to apply these services-computing principles to the designs of scientific-computing applications. To establish the validity of these suggestions, further benchmarks that run on 32 computers have demonstrated consistent speedups of at least 20 times with negligible latencies in service orchestrations (under 1% of the total simulated time). This study provides sciences with procedures that coalesce with technologies through services computing platforms.