A mathematical model to compute makespan for application integration processes using graph theory approach. (13th December 2022)

• Record Type: Journal Article
• Authors: Neuckamp, Sandra; Frantz, Rafael Z.; Roos-Frantz, Fabricia; Sawicki, Sandro
• Is Part Of: International journal of computer applications technology. Volume 69:Number 3(2022)
• Page Start: 193
• Page End: 205
• Publication Date: 2022-12-13
• Subjects: enterprise application integration -- runtime system -- mathematical modelling -- integration platforms; Technology -- Data processing -- Periodicals; 620.00285
• Languages: English
• ISSNs: 0952-8091

Abstract

The performance of integration platforms depends on the performance of execution models implemented for execution engines. Integration processes can be divided into two levels: process execution and resource management. In the first level, we considered the processing time of tasks; as for the second level, we only analysed elements of management and allocation of computational resources. Our proposal aims to identify the elements present at these levels that impact the process execution time, so we develop a mathematical model which determines makespan analytically in integration processes using a Directed Acyclic Graph (DAG). The proposed mathematical model can be used not only to compute makespan, but also to identify the optimal number of computational threads to execute an integration process with the lowest possible makespan. This mathematical model is validated by a Coffee Shop integration process, which is a well-known benchmark to evaluate and compare integration platforms.