摘要
为了定量分析系统的吞吐量、时延、缓存平均利用率等长期样本均值(long-run sample average)性能指标,应用再生周期技术(the technique of regenerative cycles)扩展了形式化建模语言POOSL的构件库。并针对在嵌入式系统中应用广泛的滑动窗口协议,使用扩展后的POOSL为选择重传滑动窗口协议建立了性能分析模型。根据模型模拟执行结果,定量地分析了窗口大小、包大小、超时间隔等参数对协议性能的影响。分析表明,扩展后的POOSL能够有效地应用于对系统的长期样本均值性能指标进行定量分析。
To analyze the long-run sample average performance metrics such as throughput, latency and average buffer occupancy etc., the component library of Parallel Object-Oriented Specification Language (POOSL) was extended with the technique of regenerative cycles. Based on the extended POOSL, the performance analysis model for the Selective Repeat Sliding Window Protocol was built. Results of the quantitative performance analysis on the influence of the parameters such as the window size, the packet size and the timeout period were given. The experimental results show that the extended POOSL is effective for the quantitative analysis on the long-run sample average performance metrics.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2007年第13期2908-2912,共5页
Journal of System Simulation
基金
国家自然科学基金(90407017)
北京市教育委员会科技发展计划重点项目(KZ200510005006)
关键词
模型
性能分析
模拟
POOSL
再生周期
models
performance evaluation
simulation
POOSL
regenerative cycles