基于排队论的以太网实时通信技术的研究
|
摘要
在排队论中,M/M/C/∞和M/M/m/k模型描述的是系统中的顾客到达服从泊松分布,服务过程服从指数分布,系统排队模型由多个服务台组成,并且M/M/m/k是M/M/C/∞模型在排队论中的一种特殊情况,即该模型的系统容量是有限的。基于该排队模型,结合顾客到达系统的分布规律以及研究排队系统的运行规律,可以实现系统性能的优化,可以提高以太网的实时通信性能。
在对以太网的应用现状和未来发展情况做出分析后得出,决定以太网实时通信系统实时性的四个重要因素是：以太网中通信通道的个数、容量、通信通道的服务能力及排队延时。结合排队论,提出了用M/M/C/∞模型来对以太网通信系统的实时性进行优化。利用M/M/C/∞模型,将以太网中原有的单输入单输出模型,转变成单输入多输出模型,在以太网的实际生产应用中,其通信通道的容量是有限的,传输的信息所能占用系统的资源也是有限的,信息传输过程中存在优先权的问题,因此,选择排队论中具有优先权的M/M/m/k混合模型,改变以太网原有的传输模式,以达到优化以太网实时通信性能的目的。
基于提出的M/M/C/∞和M/M/m/k模型,结合以太网实时通信系统,在比较各个模型中的各项系统性能指标和系统使用效率的基础上,确定系统的最优队长,以及优化系统排序后系统的实时通信性能,表明了改进模型具有优化以太网实时通信的功能,并且通过对M/M/C/∞和M/M/m/k模型进行实验,验证了改进模型的有效性和准确性。
In queuing theory, M/M/C/∞and M/M/m/k are the models that the reaching of the customers and the service of the system are distributed to Poisson distribution, the service desks in this system are multiple. The M/M/m/k model is a special form of the queuing model. And the capacity of model is limited. This model used to find out the regulation of the customers and the queue system, and based on these regulations, the theory can be used to optimize the systems. Thus, we use the queuing theory model to optimize and improve the performance of the real-time communication of the Ethernet.
Through the studying of the status and future development of the Ethernet, we find out that the number of the channels, the capacity of the channels and the service ability of the real-time communication over Ethernet can determine the performance of the Ethernet. The uncertainty delay of the communication system over Ethernet is the key reason that the system cannot be timely and accurate delivery of information. Therefore, to improve the speed of the communication we have to focus on all these reasons and decide to use the M/M/C/∞model to improve the performance of the Ethernet communication. By using the model M/M/C/∞of queuing theory, we can change the original communication model of single-input and single-output into a single-input and multiple output models. Because in the actual production of the Ethernet, the capacity of the channel is limited, the space for the information in the channel is limited and the information has its own priority, we choose the mixed M/M/m/k model with priority queuing in order to improve the performance of the real-time communication of the Ethernet.
Based on the proposed mathematical model, combining real-time Ethernet communication system, we set the efficiency and the parameters to estimate the system. And we use this information to set the best captain of the queue. Then the validity and the accuracy of the system can be proved by the experiments.
在对以太网的应用现状和未来发展情况做出分析后得出,决定以太网实时通信系统实时性的四个重要因素是：以太网中通信通道的个数、容量、通信通道的服务能力及排队延时。结合排队论,提出了用M/M/C/∞模型来对以太网通信系统的实时性进行优化。利用M/M/C/∞模型,将以太网中原有的单输入单输出模型,转变成单输入多输出模型,在以太网的实际生产应用中,其通信通道的容量是有限的,传输的信息所能占用系统的资源也是有限的,信息传输过程中存在优先权的问题,因此,选择排队论中具有优先权的M/M/m/k混合模型,改变以太网原有的传输模式,以达到优化以太网实时通信性能的目的。
基于提出的M/M/C/∞和M/M/m/k模型,结合以太网实时通信系统,在比较各个模型中的各项系统性能指标和系统使用效率的基础上,确定系统的最优队长,以及优化系统排序后系统的实时通信性能,表明了改进模型具有优化以太网实时通信的功能,并且通过对M/M/C/∞和M/M/m/k模型进行实验,验证了改进模型的有效性和准确性。
In queuing theory, M/M/C/∞and M/M/m/k are the models that the reaching of the customers and the service of the system are distributed to Poisson distribution, the service desks in this system are multiple. The M/M/m/k model is a special form of the queuing model. And the capacity of model is limited. This model used to find out the regulation of the customers and the queue system, and based on these regulations, the theory can be used to optimize the systems. Thus, we use the queuing theory model to optimize and improve the performance of the real-time communication of the Ethernet.
Through the studying of the status and future development of the Ethernet, we find out that the number of the channels, the capacity of the channels and the service ability of the real-time communication over Ethernet can determine the performance of the Ethernet. The uncertainty delay of the communication system over Ethernet is the key reason that the system cannot be timely and accurate delivery of information. Therefore, to improve the speed of the communication we have to focus on all these reasons and decide to use the M/M/C/∞model to improve the performance of the Ethernet communication. By using the model M/M/C/∞of queuing theory, we can change the original communication model of single-input and single-output into a single-input and multiple output models. Because in the actual production of the Ethernet, the capacity of the channel is limited, the space for the information in the channel is limited and the information has its own priority, we choose the mixed M/M/m/k model with priority queuing in order to improve the performance of the real-time communication of the Ethernet.
Based on the proposed mathematical model, combining real-time Ethernet communication system, we set the efficiency and the parameters to estimate the system. And we use this information to set the best captain of the queue. Then the validity and the accuracy of the system can be proved by the experiments.
引文
[1]Paolo F, Alessandra F. A Distributed Instrument for Performance Analysis of Real-Time Ethernet Networks[J]. Transactions on industrial informatics. IEEE.2008,4(1):16-25.
[2]Magnus J, Xing F, Guaranteed real-time communication in packet-switched networks with FCFS queuing[J]. Computer Networks. ELSEVIER.2008,53(3):400-417.
[3]Jay Cheng, Po-Kai Huang. Recursive Constructions of Parallel FIFO and LIFO Queues With Switched Delay Lines[J]. IEEE. Transactions on industrial informatics. 2007,35(5):1778-1798.
[4]Seyed H, Maysam H. The Monitoring of the Network Traffic Based On Queuing Theory and Simulation in Heterogeneous Network Environment[C]. International conference on computer technology and development. IEEE.2009.30(1):322-326.
[5]刘昌锦,童利标等.数据链时隙动态分配算法建模分析[J].仪器仪表学报,2007,28(4)：365-368.
[6]宋宇鲲,王锐等.使用排队论模型对FIFO深度的研究[J].仪器仪表学报,2006,27(6)：2485-2487
[7]王平,谢吴飞等.工业以太网技术[M].北京：科学出版社,2007.
[8]吴大鹏,武穆清,甄岩.有效提高WLAN吞吐量的数据帧发送策略[J].通信学报,2010,31(2)：10-16.
[9]韩伯棠.管理运筹学[M].北京：高等教育出版社,2005.
[10]陈鑫林.现代通信中的排队论[M].北京：电子工业出版社,1999.
[11]缪学勤.20种类现场总线进入IEC61158第四版国际标准[J].自动化仪表,2007,28：25-29.
[12]Matthias G, Christian S. System Q:A Queuing-Based Approach to Architecture Performance Evaluation with System C[J].IEEE Computer.2004,36:78-80.
[13]William H. A Multi-Class Dynamic User Equilibrium Model for Queuing Networks with Advanced Traveler Information Systems[J]. Journal of Mathematical Modeling and Algorithms 2003,2:349-377.
[14]F. Baccelli. Inverse problems in queueing theory and Internet probing [J]. Queue ing Syst, 2009,63:59-107.
[15]Novak, A. Taylor, P. The distribution of the number of arrivals in a subinterval of a busy period in a single server queue[J]. Queueing Syst.2006,53(3):105-114.
[16]Aditya Mahajan. Optimal Design of Sequential Real-Time Communication Systems[J]. IEEE TRANSACTIONS on Information Theory,2009,55(11):3120-3123.
[17]D. Teneketzis. On the structure of optimal real-time encoders and decoders in noisy communication[J]. IEEE Trans. Inf. Theory.2006,34:4017-4035.
[18]V. S. Borkar, S. Tatikonda. Optimal sequential vector quantization of Markov sources[J]. SIAM H. Opt. Contr. Jan,2001,40(8):135-148.
[19]M. Fiore, C. Casetti. Concurrent multipath communication for real-time traffic[J].Elsevier Computer Communications 2007,3307-3320.
[20]Seyed H,Maysam H. The Monitoring of the Network Traffic Based On Queuing Theory and Simulation in Heterogenneous Network Environment[J]. IEEE. Computer Society.2009,19(6):322-326.
[21]李立忠,李乐民.ATM在无线衰落信道上的前向纠错技术及其性能分析.通信学报,2005,20(5)：293-298.
[22]戴锦友.电信级以太网的传输技术和资源优化管理算法研究[D].武汉：华中科技大学,2009.
[23]孟坤,朱翼隽.动态优先级队列的离散时间排队分析[J].科学技术与工程,2008,24：56-59.
[24]禹海波.离散时间排队系统分析[D].西安：西安交通大学,2002.
[25]William stallings著,何军等译.无线通信基础(第二版)[M].北京：清华大学出版社,2005.
[26]顾小洪,吴秋峰.无线网络在工业企业中应用的一些探讨[J].工业控制计机,2006,19(11)：1-2.
[27]蒋放鸣.马尔可夫骨架过程及其应用[D].长沙：中南大学,2004,20(2)：13-25.
[28]刘宴兵,李秉智,幸云辉.高速信元优先级交换调度策略定性研究[J].计算机工程与应用,2002,7(15)：143-145.
[29]藏玉卫,王同胜.ATM网络中输入排队信元调度研究[J].计算机工程与应用,2004,11(11)：142-145.
[30]杨玉海,宾雪莲,郑玉墙.支持优先级的高速交换开关缓冲队列分析与设计[J].计算机工程与科学,2003,38(1)：128-131.
[31]龚文斌,甘仲民.多移动通信系统中的呼叫接入控制[J].通信学报,2004,25(3)：123-125.
[32]王博文.IP报文优先级分配的一种方案[J].计算机工程与应用,2002,38(18)：186-187.
[33]侯爽,宋颖慧.一种实时系统中的多任务可预测调度算法[J].计算机工程,2004,30(16)：67-69.
[34]邹勇.开放式实时系统的调度方法的研究[D].北京,中国科学院研究生院(软件研究所),2003.
[35]杨军.基于LAN的分布式系统实时性研究与应用[D].哈尔滨：哈尔滨工程大学,2003.
[36]沈钢,李勇男等.新型实时以太网介质访问控制协议的仿真实现[J].系统仿真学报,2002,14(07)：890-893.
[37]饶运涛,邹继军,郑勇芸著.现场总线CAN原理与应用技术[M].北京：北京航空航天大学出版社,2003.
[38]王永炎,王强,王宏安等.基于优先级表的实时调度算法及其实现[J].软件学报,2004,15(03)：360-370.
[39]王梓坤,杨向群.生灭过程与马尔可夫链[M].科学出版社,2005,1.
[40]孙荣恒.随机过程及其应用(M).北京：清华大学出版社,2004.
[41]候玉梅,田乃硕.M/M/n休假排队系统综述[J].运筹学杂志,2000,5(2)：88-94.
[42]王平,谢昊飞,肖琼等.工业以太网技术.北京：科学出版社,2007.
[43]孙攀,王平,谢昊飞.以太网工厂自动化协议中确定性调度的研究与实现[J].计算机集成制造系统,2007,13(3)：563-567.
[44]王忠锋,于海斌,王宏等.工业以太网实时通信中的调度表构建策略研究[J].信息与控制,2007,36(3)：165-261.
[45]吴爱国.工业以太网的发展现状[J].信息与控制,2003,32(5)：459-461.
[46]阳宪惠.工业数据通信与控制网络[M].北京：清华大学出版社,2003.
[2]Magnus J, Xing F, Guaranteed real-time communication in packet-switched networks with FCFS queuing[J]. Computer Networks. ELSEVIER.2008,53(3):400-417.
[3]Jay Cheng, Po-Kai Huang. Recursive Constructions of Parallel FIFO and LIFO Queues With Switched Delay Lines[J]. IEEE. Transactions on industrial informatics. 2007,35(5):1778-1798.
[4]Seyed H, Maysam H. The Monitoring of the Network Traffic Based On Queuing Theory and Simulation in Heterogeneous Network Environment[C]. International conference on computer technology and development. IEEE.2009.30(1):322-326.
[5]刘昌锦,童利标等.数据链时隙动态分配算法建模分析[J].仪器仪表学报,2007,28(4)：365-368.
[6]宋宇鲲,王锐等.使用排队论模型对FIFO深度的研究[J].仪器仪表学报,2006,27(6)：2485-2487
[7]王平,谢吴飞等.工业以太网技术[M].北京：科学出版社,2007.
[8]吴大鹏,武穆清,甄岩.有效提高WLAN吞吐量的数据帧发送策略[J].通信学报,2010,31(2)：10-16.
[9]韩伯棠.管理运筹学[M].北京：高等教育出版社,2005.
[10]陈鑫林.现代通信中的排队论[M].北京：电子工业出版社,1999.
[11]缪学勤.20种类现场总线进入IEC61158第四版国际标准[J].自动化仪表,2007,28：25-29.
[12]Matthias G, Christian S. System Q:A Queuing-Based Approach to Architecture Performance Evaluation with System C[J].IEEE Computer.2004,36:78-80.
[13]William H. A Multi-Class Dynamic User Equilibrium Model for Queuing Networks with Advanced Traveler Information Systems[J]. Journal of Mathematical Modeling and Algorithms 2003,2:349-377.
[14]F. Baccelli. Inverse problems in queueing theory and Internet probing [J]. Queue ing Syst, 2009,63:59-107.
[15]Novak, A. Taylor, P. The distribution of the number of arrivals in a subinterval of a busy period in a single server queue[J]. Queueing Syst.2006,53(3):105-114.
[16]Aditya Mahajan. Optimal Design of Sequential Real-Time Communication Systems[J]. IEEE TRANSACTIONS on Information Theory,2009,55(11):3120-3123.
[17]D. Teneketzis. On the structure of optimal real-time encoders and decoders in noisy communication[J]. IEEE Trans. Inf. Theory.2006,34:4017-4035.
[18]V. S. Borkar, S. Tatikonda. Optimal sequential vector quantization of Markov sources[J]. SIAM H. Opt. Contr. Jan,2001,40(8):135-148.
[19]M. Fiore, C. Casetti. Concurrent multipath communication for real-time traffic[J].Elsevier Computer Communications 2007,3307-3320.
[20]Seyed H,Maysam H. The Monitoring of the Network Traffic Based On Queuing Theory and Simulation in Heterogenneous Network Environment[J]. IEEE. Computer Society.2009,19(6):322-326.
[21]李立忠,李乐民.ATM在无线衰落信道上的前向纠错技术及其性能分析.通信学报,2005,20(5)：293-298.
[22]戴锦友.电信级以太网的传输技术和资源优化管理算法研究[D].武汉：华中科技大学,2009.
[23]孟坤,朱翼隽.动态优先级队列的离散时间排队分析[J].科学技术与工程,2008,24：56-59.
[24]禹海波.离散时间排队系统分析[D].西安：西安交通大学,2002.
[25]William stallings著,何军等译.无线通信基础(第二版)[M].北京：清华大学出版社,2005.
[26]顾小洪,吴秋峰.无线网络在工业企业中应用的一些探讨[J].工业控制计机,2006,19(11)：1-2.
[27]蒋放鸣.马尔可夫骨架过程及其应用[D].长沙：中南大学,2004,20(2)：13-25.
[28]刘宴兵,李秉智,幸云辉.高速信元优先级交换调度策略定性研究[J].计算机工程与应用,2002,7(15)：143-145.
[29]藏玉卫,王同胜.ATM网络中输入排队信元调度研究[J].计算机工程与应用,2004,11(11)：142-145.
[30]杨玉海,宾雪莲,郑玉墙.支持优先级的高速交换开关缓冲队列分析与设计[J].计算机工程与科学,2003,38(1)：128-131.
[31]龚文斌,甘仲民.多移动通信系统中的呼叫接入控制[J].通信学报,2004,25(3)：123-125.
[32]王博文.IP报文优先级分配的一种方案[J].计算机工程与应用,2002,38(18)：186-187.
[33]侯爽,宋颖慧.一种实时系统中的多任务可预测调度算法[J].计算机工程,2004,30(16)：67-69.
[34]邹勇.开放式实时系统的调度方法的研究[D].北京,中国科学院研究生院(软件研究所),2003.
[35]杨军.基于LAN的分布式系统实时性研究与应用[D].哈尔滨：哈尔滨工程大学,2003.
[36]沈钢,李勇男等.新型实时以太网介质访问控制协议的仿真实现[J].系统仿真学报,2002,14(07)：890-893.
[37]饶运涛,邹继军,郑勇芸著.现场总线CAN原理与应用技术[M].北京：北京航空航天大学出版社,2003.
[38]王永炎,王强,王宏安等.基于优先级表的实时调度算法及其实现[J].软件学报,2004,15(03)：360-370.
[39]王梓坤,杨向群.生灭过程与马尔可夫链[M].科学出版社,2005,1.
[40]孙荣恒.随机过程及其应用(M).北京：清华大学出版社,2004.
[41]候玉梅,田乃硕.M/M/n休假排队系统综述[J].运筹学杂志,2000,5(2)：88-94.
[42]王平,谢昊飞,肖琼等.工业以太网技术.北京：科学出版社,2007.
[43]孙攀,王平,谢昊飞.以太网工厂自动化协议中确定性调度的研究与实现[J].计算机集成制造系统,2007,13(3)：563-567.
[44]王忠锋,于海斌,王宏等.工业以太网实时通信中的调度表构建策略研究[J].信息与控制,2007,36(3)：165-261.
[45]吴爱国.工业以太网的发展现状[J].信息与控制,2003,32(5)：459-461.
[46]阳宪惠.工业数据通信与控制网络[M].北京：清华大学出版社,2003.