针对OPNET网络仿真规模很大时系统资源消耗增加、仿真效率降低，严重时会导致仿真无法进行甚至自动退出等问题，本文以话音网络为例，从构建网络拓扑、提高仿真效率等两个方面对OPNET在仿真大规模网络性能时可能出现的问题及解决方法进行了举例验证。结果表明当网络规模很大且分布不规律时，可以利用EMA技术生成网络节点，根据实际需求可以择优选取混合式仿真、汇聚网络流量并简化网络模型、OPNET与其他编程软件相结合、分布式仿真等方法用于提高仿真效率，相关结论对利用OPNET仿真大规模网络性能具有一定的借鉴意义。
Aiming at the large scale of OPNET network simulation, the system resource consumption increases and the simulation efficiency decreases. Seriously, the simulation cannot be carried out or even automatically quit. This paper took voice network as an example to illustrate the possible problems and solutions of OPNET in simulating large-scale network performance from two aspects: building network topology and improving simulation efficiency. The results showed that when the network scale was large and the distribution was irregular, EMA technology could be used to generate net-work nodes, hybrid simulation, aggregation of network traffic and simplification of network model, combination of OPNET and other programming software; distributed simulation could be selected according to actual needs to improve simulation efficiency. The conclusion can be used for reference in simulating large-scale network performance with OPNET.
