Distributed computing systems are designed to solve computationally intensive problems with the help of convergence of computing resources scattered across the network.But the selection of most efficient distributed computing object middleware technology on the basis of different performance metrics is an important research issue.In this paper,the performance evaluation and comparison of distributed computing object middleware technologies were presented,including CORBA,ICE,HORB and TCP based Dot NET Remoting.The results show that ICE demonstrates better performance in terms of overhead generation,while HORB shows reduced round trip latency.%分布式计算系统旨在借助对分散在网络上的计算资源的整合来解决大量计算问题.对基于不同的性能指标,选择最有效的分布式对象中间件技术是一项重要的研究课题.本文在额外开销和往返延迟等方面对CORBA、ICE、HORB和基于TCP的Dot NET Remoting这几种分布式计算对象中间件技术做了性能评价和实验测试.结果表明,ICE额外开销的产生较低,而相比于其他中间件,HORB往返延迟较小.
展开▼