Empowerment of cluster and grid load balancing algorithms to support distributed exascale computing systems with high compatibility

Shirin Shahrabi | Elham Adibi | Ehsan Mousavi Khaneghah | Fazeh Mollsalehi

---

URL :   http://research.shahed.ac.ir/WSR/WebPages/Report/PaperView.aspx?PaperID=159543
Date :  2022/05/26
Publish in :    International Journal of Computational Science and Engineering
DOI :  https://doi.org/10.1504/IJCSE.2022.123110

Keywords :distributed exascale computing; load balancing; events with the dynamic and interactive nature; D&I; load balancing algorithms; cluster computing; grid computing.

Abstract :

---

The dynamic and interactive nature of Distributed Exascale Computing System (DECS), leads to a situation in which the load balancing element lacks the proper pattern for the solution. In this article, in addition to analysis and review of dynamic and interactive nature and its effect on load balancing, we introduce a framework for managing load balancing that does not need to analyze dynamic and interactive nature. This framework proposes a mathematical scheme for load balancing element’s functionality and also redefines its functions and consisting elements. The redefinition makes it possible to determine the constituent elements of the framework and their functionality without the need to analyze the dynamic and interactive nature of the system. The proposed framework can manage and control dynamic and interactive events via reviewing changes in the pattern of the functionality of resources, the pattern of data collection to execute processes related to load balancing, and a Scalable tool. In addition to performing the load balancing element’s functionality, our framework is capable of continuing to perform under dynamic and interactive events in distributed exascale systems. This framework has on average 43 of improvement inability to respond to dynamic and interactive requests.