An approach to parallelization of legacy software

Abstract

The issue of parallelization of large and complex software systems has been widely researched, but no definite solution has been published yet. In the paper we introduce the concept of parallelization based on the identification of mutually independent data sets and refactoring legacy software for the purpose of adjusting it to execution as a task tree. Then we present implementation of the concept on existing distribution management system. Finally, we present the results of the performance evaluation based on a series of experiments made on the dual-core symmetric multiprocessor. © 2009 IEEE.