AN EFFECTIVE WAY TO CONTROL CONGESTION FOR RTI WITH HYBRID ARCHITECTURE

Abstract

Congestions brought by data explosion may suffer system performance or even cause fatal errors in large-scale simulation systems. This paper analyzes the reasons of generating congestion with queuing theory, and proposes a novel congestion control approach for RTI with hybrid architecture, including framework, sampling policy, queue length predicting, and congestion control algorithm. The framework promotes the usage of federate resources, and improves the performance of HLA simulation systems by dynamically distributing loads on RTIG and LRCs (Local RTI Components). At last, experimental results say that the proposed approach is a flexible, simple and efficient way to control congestion in RTI with hybrid architecture.