High performance modeling for distributed simulation

  • Authors:
  • Jong Sik Lee

  • Affiliations:
  • School of Computer Science and Engineering, Inha University, Incheon, South Korea

  • Venue:
  • AIS'04 Proceedings of the 13th international conference on AI, Simulation, and Planning in High Autonomy Systems
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a modeling of a distributed simulation system with a data management scheme. The scheme focuses on a distributed simulation concept, which is load balancing, suggests distribution of a different functionality to each distributed component, and assigns various degrees of communication and computation loads in each component. In addition, this paper introduces a design with an inter-federation communication on HLA-compliant distributed simulation. The design focuses on integration among multiple federations and constructs the larger distributed simulation system by suggesting a HLA bridge connection. The integration supports system simulation flexibility and scalability. This paper discusses design issues of a practical system with a HLA bridge for inter-federation communication. This paper analyzes and evaluates performance and scalability of the data management scheme with load balancing on distributed simulation, especially with inter-federation and inside-federation communication configurations. The analytical and empirical results on a heterogeneous OS distributed simulation show improvement of system performance and scalability by using data management and inter-federation communication.