Búsqueda Imágenes Maps Play YouTube Noticias Gmail Drive Más »
Iniciar sesión
Usuarios de lectores de pantalla: deben hacer clic en este enlace para utilizar el modo de accesibilidad. Este modo tiene las mismas funciones esenciales pero funciona mejor con el lector.

Patentes

  1. Búsqueda avanzada de patentes
Número de publicaciónCN1601995 A
Tipo de publicaciónSolicitud
Número de solicitudCN 03135891
Fecha de publicación30 Mar 2005
Fecha de presentación25 Sep 2003
Fecha de prioridad25 Sep 2003
También publicado comoCN1316798C
Número de publicación03135891.8, CN 03135891, CN 1601995 A, CN 1601995A, CN-A-1601995, CN03135891, CN03135891.8, CN1601995 A, CN1601995A
Inventores许都, 王晟, 谭伟
Solicitante电子科技大学
Exportar citaBiBTeX, EndNote, RefMan
Enlaces externos:  SIPO, Espacenet
Exchanging scheduling method of multiple packing collection
CN 1601995 A
Resumen
Using period processing of BHP instead of traditional sequential processing of BHP, the invention provides conditions for BHP queuing, and paroxysmal data in high priority to preempt paroxysmal data in low priority. Different QoS are provided for different types of paroxysmal data through flexible queuing strategy. Within a dispatching cycle, multiple BHP is processed through optimized dispatch algorithm in multiple constraints and multiple targets. BHP dispatch in optimal mode raises utilization efficiency of exchange resources so as to reach purpose of raising switching performance.
Reclamaciones(1)  traducido del chino
1.一种多控制分组汇聚交换调度方法,其特征是采用下面步骤:步骤1确定BHP汇聚时间的长度步骤:根据BHP的到达频度与交换性能要求(主要是突发数据丢失率和交换时延),确定一个BHP汇聚时间的长度,该时间长度同时也是一个交换调度周期;步骤2启动超时计数器步骤:当第一个BHP(来自边缘节点、核心节点或前一个调度周期无法处理的BHP)到达当前核心节点时,启动窗口超时计数器,并记录后续到达的每个BHP的到达时间;步骤3排队步骤:依据BHP中携带的信息,包括服务质量(Quality of Service,QoS)、突发数据长度和偏移时间信息,将汇聚时间内到达的所有BHP排队;所述排队的方法,可以是按照突发数据到达时间进行排队,可以是按照优先级高低进行排队,也可以是按照突发数据长度进行排队,还可以是将以上多种方法综合起来;步骤4出队列步骤:当窗口超时计数器值超过所设定的汇聚时间,则从多个等待队列中按照公平服务算法取出若干个BHP,送入交换调度器进行调度处理;所述的公平服务算法可以是DRR(Dual Round-Robin),也可以是WFQ(WeightedFair Queue),还可以是ERR(Elastic Round-Robin);每次取出BHP的个数,可根据调度算法实现的复杂度、性能提高程度的要求和处理器的处理能力决定;步骤5调度步骤:交换调度器提取步骤4中出队列的BHP所携带的与交换资源占用相关的信息,包括突发数据到达时间、突发数据长度、出端口和波长,搜索当前交换节点可用资源表(主要包括各出端口波长的空闲信息、波长变换器的可用信息、光交叉连接矩阵及光纤延时线配置状态),采用多约束多目标优化算法使尽可能多的BHP能够在这个调度周期内成功预约资源,即使得尽可能多的Burst的光通道得到预约配置;步骤6更新可用资源表步骤:用步骤5已预约的资源更新当前节点的可用资源表;步骤7转发BHP步骤:将已经成功预约资源的BHP转发到下游节点,并返回步骤2等待新的BHP到达;步骤8返回步骤:将未能成功预约资源、且总等待时间还未超过偏移时间的BHP返回服务等待队列,返回步骤2;否则丢弃。 A multi-control packet aggregation switching scheduling method, which is characterized by using the following steps: Step 1 Determine the length of time step convergence BHP: According to BHP's reach and frequency switching performance requirements (mainly burst data loss rate and exchange time extension) to determine the length of time a convergence BHP, the length of time is also an exchange scheduling period; Step 2 Start timeout counter steps: first when BHP BHP (from the edge node, core node or the previous scheduling period can not handle) When it reaches the current core nodes start window timeout counter, and record the arrival time of each subsequent arrival of BHP; Step 3 queuing step: According to BHP carried, including QOS (Quality of Service, QoS), burst length and offset time information of all the convergence BHP queuing time is reached; the queuing method can be carried out according to the arrival time of burst data queue according to the priority level can be queued, the data may be in accordance with the burst length queued, may also be a variety of methods above together; step 4 queues step: When the window timeout counter value exceeds the set convergence time, remove from the plurality of queue BHP in a fair number of service algorithms, send the exchange schedulers treatment; fair service algorithm described may be DRR (Dual Round-Robin), may also be WFQ (WeightedFair Queue), can also be a ERR (Elastic Round-Robin); each removal of a BHP number, according to the complexity of the scheduling algorithm, the performance requirements and to improve the level of processing power decisions; Step 5 Schedule the steps of: switching scheduler extracting step 4 in the queue BHP carried footprint associated with the exchange of information including burst data arrival time, burst length, the port and wavelength, to search for available information on the current switching node list of available resources (including each of the port wavelength free information, a wavelength converter, optical cross-connect matrix and fiber extension When line configuration state), multi-constrained multi-objective optimization algorithm can make as many BHP in this book the resource scheduling period, even with the Burst as many optical channels get subscription configurations; Step 6 steps to update the resource table available : Update the current node in step 5 using the reserved resources available resource table; Step 7 forwarding BHP steps: BHP resource reservation has been successfully transmitted to the downstream node, and return to step 2 wait for the new BHP arrival; Step 8 Return steps: failed to reserve resources, and the total wait time has not been more than offset time BHP return service queue, return to step 2; otherwise discarded.
Descripción  traducido del chino
一种多控制分组汇聚交换调度方法 A multi-control packet aggregation switching scheduling method

技术领域 Technical Field

本发明属于通信系统中的交换控制技术领域,它特别涉及光突发交换系统的交换控制技术。 The present invention pertains to a communication system switching control technology, it is particularly involved in the exchange of control technology for optical burst switching system.

背景技术 Background

通信系统一般由接入、传输和交换等部分构成,其中交换主要实现通信网络中不同用户间的连接,如电路交换系统中连接的建立、分组交换系统中业务的转发等。 Communication system generally consists of part access, transport and switching, etc., which connect the main exchange between different users in a communication network, such as circuit switching system connection is established, the packet forwarding and switching systems and other services. 随着网络用户的增加、信息量的增多、网络规模的加大等,传统的基于电交换的技术要达到很高的交换性能,如高交换容量、高端口速率、低交换时延等,所面临的压力越来越大。 With the increase of Internet users, the amount of information increases, the increase of network scale, the traditional exchange-based electricity technology to achieve high switching performance, such as high switching capacity, high port speed, low switching and delay, the facing increasing pressure.

克服电交换瓶颈的最根本方法在于采用全光的分组交换技术,但目前该技术还存在若干难点,如全光存储、光定时/同步等。 The most fundamental way to overcome power exchange bottleneck is the use of all-optical packet switching technology, but the technology, there are a number of difficulties, such as all-optical storage, optical Timing / synchronization. 为此,充分结合并发挥现有光、电技术的优势,构建光电混合的交换系统,是实现大容量交换系统的有效方法,如光突发交换技术(Optical Burst Switching,简称OBS)。 To this end, fully integrate and take advantage of existing light, power technology, to build photovoltaic hybrid switching system is an effective way to achieve high-capacity switching systems, optical burst switching technology (Optical Burst Switching, referred OBS).

在OBS网络中,承载用户业务突发数据(Burst)可以看作是由大量数据分组构成的超长分组,而这个超长分组的分组头就是突发数据的控制分组(BurstHeader Packet,BHP)。 In OBS networks, the burst carrying user traffic data (Burst) can be seen as long packet configured by a large number of data packets, the packet header and the packet is the control packet long burst data (BurstHeader Packet, BHP). 与传统分组交换不同的是,BHP与Burst在物理通道上是分离的:在DWDM传输系统中,可以采用一个(或多个)专门的波长作为控制通道,用于传送BHP,而把其他的波长作为数据通道。 And traditional packet switching difference is, BHP and Burst in the physical channel is separated: in DWDM transmission system, can be used one (or more) specific wavelength as a control channel for transmitting BHP, while the other wavelengths as the data channel. BHP与Burst间一一对应,BHP中包含突发数据的有关信息,如偏移时间、突发长度、数据通道(波长)等。 Burst-one correspondence between the BHP, BHP information contained in the burst data, such as offset time, burst length, data channel (wavelength) and the like. 在源端真正发送用户数据前,设置BHP与突发数据之间的偏移时间(offset time,即源端发送BHP与发送相应突发数据之间的间隔时间),使BHP先于Burst到达OBS的中间节点。 Before actually transmitting user data source, set the offset time (offset time, ie the time between sending BHP source corresponding burst data transmission between) BHP and burst data, thereby making BHP prior to arrival OBS Burst The intermediate node. BHP在中间节点转换为电信号进行处理,包括路由的确定、资源的预约以及交换矩阵的配置等,保证当突发数据到达时相应的数据通道已经配置好,从而实现数据在光域的透明传输。 BHP intermediate node converted to electrical signals for processing, including routing determination, reserve resources and the configuration of the switching matrix, etc., to ensure that when the burst data reach the appropriate data channel has been configured to achieve transparent transmission of data in the optical domain .

从完成的功能来看,BHP与传统电路交换网络中的信令非常相似,正是在这个意义上,BHP也被称为信令消息。 From a functional point of view is completed, BHP and traditional circuit-switched network signaling is very similar, it is in this sense, BHP is also known as signaling messages. 但与传统信令不同的是,OBS的信令不必等待目的端的反馈确认,即OBS的资源预约是单向的。 But the difference is that the traditional signaling, OBS signaling without waiting for feedback to confirm the destination, namely OBS resource reservation is unidirectional. 也正是这种“单向预约”机制减小了连接建立延迟,提高了信道利用率。 It is this "one-way reservation" mechanism reduces the connection establishment delay and improve the channel utilization.

一个OBS网络主要由边缘节点、核心节点和DWDM链路构成(见附图1),其中边缘节点负责对数据分组进行缓存和封装,组合成突发数据,然后发送给与之最邻近的OBS核心节点。 A OBS network consists of edge nodes, core nodes and DWDM link configuration (see Figure 1), wherein the edge node is responsible for caching data packets and packages combined into a burst of data, and then sent to the core with the neighboring OBS node. 封装时边缘节点生成描述突发数据特性的BHP分组,先于突发数据在特定的控制通道上发送。 Edge node generates a burst data characteristics described BHP packet is encapsulated, the first sent on the control channel in a particular burst data. 核心节点根据控制通道上收到的BHP,可以得知突发数据的到达时间、持续时间、目的地址(和转发标签)等控制信息,并根据这些信息完成对光路的配置,保证数据的透明通道(见文献Y.Xiong,M Vandenhoute,and H.Cankaya.Control architecture in optical burst-switched WDM networks.IEEE Journal on Selected Areas in Communications,18:1838-1851,October2000)。 Core node based BHP received on the control channel, you can learn the arrival time of burst data, duration, destination address (and forwarding labels) and other control information, and complete the configuration of the optical path based on the information, to ensure transparency channel data (see literature Y.Xiong, M Vandenhoute, and H.Cankaya.Control architecture in optical burst-switched WDM networks.IEEE Journal on Selected Areas in Communications, 18: 1838-1851, October2000).

虽然OBS系统在理论上可以达到很高的性能,如大的交换容量、低的交换时延、高的资源利用率等,但在目前的光技术水平下,需要解决的技术问题还很多,如本发明所针对的交换调度问题。 Although the OBS system can theoretically achieve high performance, such as a large switching capacity, low latency switching, high resource utilization, etc., but in light of the current level of technology, technology is still a lot of problems to be solved, such as The present invention is directed to exchange scheduling problems.

交换调度所要解决的问题可简单描述为:在有限的交换资源下,提供尽可能高的交换性能。 Exchange scheduling problem can be simply described as: the limited exchange of resources to provide the highest possible switching performance. 在此,可用的交换资源有:光交叉连接矩阵、光延时线、波长变换器、单端口(光纤)多个工作波长等;需满足的交换性能有:交换时延、交换容量、交换丢失率等。 In this case, the exchange of resources available: optical cross-connect matrix, optical delay line, wavelength converter, a plurality of working single port (optical) wavelength and the like; the exchange must meet the performance: switching delay, exchange capacity, exchange loss rate.

在现有的关于OBS核心节点的调度算法及相关的性能分析中,均采用了“先来先服务”(First Come First Service,FCFS)的机制,即我们称为顺序到达调度的方法:核心节点每收到一个来自于上游节点的BHP,马上启动调度算法,为该BHP对应的Burst分配可用的交换资源,并将该BHP转发到下游节点。 In the conventional scheduling algorithms on OBS core nodes and related performance analysis, we are used "first come, first served" (First Come First Service, FCFS) mechanism, which we called the order of arrival scheduling methods: the core nodes Each received one from BHP upstream node, immediately start scheduling algorithm, distribution BHP for Burst corresponding exchange resources available, and the BHP forwarded to the downstream node.

但这类交换控制(调度)方案存在的主要问题是,OBS中的任意一个核心节点在处理完一个BHP后,若该BHP预约的资源可用,则该BHP马上转发到下一个节点,使得它所预约的交换资源不可更改;而对于当前节点交换资源无法满足BHP预约要求的,该BHP被丢弃。 But such major issues exchange control (scheduling) programs exist that, OBS any one of the core node after handling a BHP, BHP reservation if the resources are available, then the BHP immediately forwarded to the next node, making it Appointment of exchange resources can not be changed; and for the exchange of resources can not meet the current node BHP reserve requirements, which BHP is discarded. 这样导致的问题有:1、突发数据的交换优先级无法充分、灵活保证(见文献M.Yoo and C.Qiao.Supporting multiple classes of services in IP over WDM networks.In proceeding ofGLOBECOM,volume lb,pages 1023-1027,1999.和文献Mei Yang,SQZheng,and D.Verchere,“A QoS supporting scheduling algorithm for Optical Burst Switching DWDMNetworks,”Proc.of IEEE GLOBECOM 2001,Vol.1,pp.86-91,11/01.); Such problems caused by: 1, the priority of the burst data exchange can not be sufficiently flexible to ensure that (see document M.Yoo and C.Qiao.Supporting multiple classes of services in IP over WDM networks.In proceeding ofGLOBECOM, volume lb, pages 1023-1027,1999. and literature Mei Yang, SQZheng, and D.Verchere, "A QoS supporting scheduling algorithm for Optical Burst Switching DWDMNetworks," Proc.of IEEE GLOBECOM 2001, Vol.1, pp.86-91,11 / 01.);

2、从时间角度看,交换资源的使用效率低,存在大量的“碎片”,即在某个时间段内虽然部分交换资源是空闲的,但这些资源却无法为后续到达的突发数据所利用(见文献M.Iizuka,M.Sakuta,Yoshiyuki,“A SchedulingAlgorithm Minimizing Voids Generated by Arriving Bursts in Optical Burst Switched WDMNetwork,”Proceedings,IEEE Globecom 2002,November 2002.);3、在一次调度过程中无法处理的BHP被丢弃,使得系统的交换丢失率高(见文献Jinhui Xu,C.Qiao,J.Li,and G.Xu.″Efficient Channel Scheduling Algorithmsin Optical Burst Switched Networks″,IEEE INFOCOM 2003,22nd Annual JointConference of the IEEE Computer and Communications Societies,San Francisco,March2003)。 2. From the time perspective, inefficient use of resources exchange, there is a lot of "debris" that in a period of time, although part of the free exchange of resources, but these resources can not for the sudden arrival of the subsequent use of the data (see literature M.Iizuka, M.Sakuta, Yoshiyuki, "A SchedulingAlgorithm Minimizing Voids Generated by Arriving Bursts in Optical Burst Switched WDMNetwork," Proceedings, IEEE Globecom 2002, November 2002.); 3, in a scheduling process can not handle BHP is discarded, so that the exchange rate system is lost (see literature Jinhui Xu, C.Qiao, J.Li, and G.Xu. "Efficient Channel Scheduling Algorithmsin Optical Burst Switched Networks", IEEE INFOCOM 2003,22nd Annual JointConference of the IEEE Computer and Communications Societies, San Francisco, March2003).

发明内容 DISCLOSURE

针对现有OBS系统中核心交换调度方法的缺陷,本发明的目的是提供一种多控制分组汇聚交换调度方法,该方法在相同的交换资源下(如相同的光交叉连接矩阵、光延时线、波长变换器等,及这些光路相关器件的连接关系),可提高OBS系统的交换性能,如降低交换丢失率,提高对多优先级业务的支持等。 Defects in the existing system, the core switching OBS scheduling method, object of the present invention is to provide a multi-control packet aggregation switching scheduling method (such as the same optical cross-connect matrix at the same exchange resources, optical delay line wavelength converter, etc., and the connection between these optical paths associated devices), can improve the performance of OBS switching system, such as reducing the rate of exchange loss and improve support for multi-priority traffic and so on.

本发明的一种多控制分组汇聚交换调度方法是设业务量强度为ρ,BHP的到达频度为λ,且初始时核心节点没有未处理的BHP,其特征是采用下面步骤:步骤1确定BHP汇聚时间的长度步骤:根据BHP的到达频度与交换性能要求(主要是突发数据丢失率和交换时延),确定一个BHP汇聚时间的长度,该时间长度同时也是一个交换调度周期;步骤2启动超时计数器步骤:当第一个BHP(来自边缘节点、核心节点或前一个调度周期无法处理的BHP)到达当前核心节点时,启动窗口超时计数器,并记录后续到达的每个BHP的到达时间;步骤3排队步骤:依据BHP中携带的信息,包括服务质量(Quality of Service,QoS)、突发数据长度和偏移时间信息,将汇聚时间内到达的所有BHP排队;所述排队的方法,可以是按照突发数据到达时间进行排队,可以是按照优先级高低进行排队,也可以是按照突发数据长度进行排队,还可以是将以上多种方法综合起来; The present invention is a multi-control packet aggregation switching scheduling method is set up traffic intensity ρ, BHP reach the frequency of λ, and the initial core node when there are no pending BHP, which is characterized by using the following steps: Step 1 Determine BHP the length of time step convergence: According to BHP's reach and frequency switching performance requirements (mainly burst data rate and exchange loss delay), to determine a length BHP convergence of time, the length of time is also an exchange scheduling period; Step 2 Start out counter step: When the first BHP (from the edge node, core node or the previous scheduling period can not handle BHP) to reach the core of the current node, the startup window timeout counter, and record the arrival time of each subsequent arrival of BHP; Step 3 line up the steps of: based BHP carried, including QOS (Quality of Service, QoS), burst length and offset time information, all the time to reach convergence within BHP queuing; the queuing method, in accordance with the arrival time of burst data queues may be queued according to priority level, may be carried out in accordance with the burst length of data queue, also may be a variety of the above methods together;

步骤4出队列步骤:当窗口超时计数器值超过所设定的汇聚时间,则从多个等待队列中按照公平服务算法取出若干个BHP,送入交换调度器进行调度处理;所述的公平服务算法可以是DRR(Dual Round-Robin),也可以是WFQ(Weighted Fair Queue),还可以是ERR(Elastic Round-Robin);每次取出BHP的个数,可根据调度算法实现的复杂度、性能提高程度的要求和处理器的处理能力决定;步骤5调度步骤:交换调度器提取步骤4中出队列的BHP所携带的与交换资源占用相关的信息,包括突发数据到达时间、突发数据长度、出端口和波长,搜索当前交换节点可用资源表(主要包括各出端口波长的空闲信息、波长变换器的可用信息、光交叉连接矩阵及光纤延时线配置状态),采用多约束多目标优化算法使尽可能多的BHP能够在这个调度周期内成功预约资源,即使得尽可能多的Burst的光通道得到预约配置;步骤6更新可用资源表步骤:用步骤5已预约的资源更新当前节点的可用资源表;步骤7转发BHP步骤:将已经成功预约资源的BHP转发到下游节点,并返回步骤2等待新的BHP到达;步骤8返回步骤:将未能成功预约资源、且总等待时间还未超过偏移时间的BHP返回服务等待队列,返回步骤2;否则丢弃。 Step 4 queues step: When the window timeout counter value exceeds the set convergence time, remove from the plurality of queue BHP in a fair number of service algorithms into exchange schedulers treatment; fair service algorithm described can be DRR (Dual Round-Robin), may also be WFQ (Weighted Fair Queue), may also be ERR (Elastic Round-Robin); each number BHP withdrawn, according to the complexity of the scheduling algorithm, the performance increase degree requirements and processing power decisions; Step 5 Schedule the steps of: switching scheduler extracting step 4 in the queue BHP carried footprint associated with the exchange of information, including burst data arrival time, burst length, the port and wavelength switching node searches the current list of available resources (including each of the port wavelength free information, a wavelength converter available information, optical cross-connect matrix and fiber optic delay line configuration status), multi-constrained multi-objective optimization algorithm BHP can make as much in this book the resource scheduling period even get as much as possible to get an appointment Burst of light channel configuration; Step 6 update available resource table step: updating the current node in step 5 with the reserved resources available resource table; Step 7 BHP forwarding steps of: BHP resource reservation has been successfully transmitted to the downstream node, and returns to step 2 to wait for the new BHP arrives; Step 8 returns to step: the failed resource reservation, and the total wait time has not been exceeded BHP offset time of return to the service queue, return to step 2; otherwise discarded.

本发明方法的流程图如图2所示。 A flowchart of a method of the present invention is shown in Figure 2. 实现功能框图见附图3。 Achieve a functional block diagram see Figure 3.

对比传统的方案,可见本文所阐述方案的创新性在于:设置BHP汇聚时间进行周期性处理;可以利用灵活的排队和出队列策略为高优先级的突发数据提供QoS保证,并可降低总的突发数据丢失率;采用多约束多目标优化调度算法进一步优化调度,提高系统性能。 Compared to traditional solutions, innovative programs described in this paper shows that: Set BHP convergence time for periodic treatment; can use flexible policies for queuing and the queue bursts of high-priority data provide QoS guarantee, and reduce the total Burst data loss; multi-constrained multi-objective optimization scheduling algorithm to further optimize the scheduling, improve system performance. 特别地,步骤3和4用于实现系统对多QoS突发数据的支持,与现有的OBS系统中的QoS机制相比,该方法在公平性、灵活性等方面有很大提高。 In particular, steps 3 and 4 for implementing the system supports multiple QoS bursts of data, compared with the existing QoS mechanisms OBS system, which has greatly improved in terms of fairness and flexibility. 传统的调度方案是基于BHP顺序到达的,调度器每次处理一个BHP,由BHP信息确定它所要求的交换资源是否可用:能够预约配置光路,则占用交换资源,转发BHP,并当它对应的Burst到达核心交换节点时让该Burst通过;否则丢弃该BHP,丢弃它对应的Burst分组。 Traditional scheduling scheme is based BHP order of arrival, a scheduler for each treatment BHP, BHP is determined by whether the information it has requested exchange resources available: the ability to configure the optical path reservation, the occupancy exchange resources, forwarding BHP, and when it corresponds to Let arrives when Burst Burst through the core switching node; otherwise discard the BHP, which corresponds to Burst discarded packets.

本发明的多控制分组汇聚交换调度方法的实质是:利用对BHP的周期性处理这一创新来打破传统的对BHP的顺序处理,为BHP排队和高优先级突发数据抢占低优先级突发数据提供条件;并利用灵活的排队和出队列策略为不同类的突发数据提供不同的QoS;同时利用多约束多目标优化调度算法在一个调度周期内同时处理多个BHP,以更优化的方式对BHP进行调度。 More substance control packet aggregation switching scheduling method of the present invention are: use of BHP periodically deal with this innovation to break the traditional sequential processing to BHP for BHP and high-priority queue preempt lower priority burst burst Data provided conditions; and with flexible queuing and a queue policy to provide different QoS for different types of data bursts; while taking advantage of multi-constrained multi-objective optimization scheduling algorithm in a scheduling period at the same time to handle multiple BHP, in a more optimal way scheduled for BHP. 从而提高了交换资源的利用效率,达到了提高交换性能的目的。 Thereby improving the utilization efficiency of the exchange of resources, to achieve the purpose of improving the switching performance.

本发明的多控制分组汇聚交换调度方法的特点在于利用OBS系统中BHP先于突发数据到达核心交换节点的特点,执行每次调度时不是依据一个BHP所携带的信息,而是利用某个时间段内所到达的多个BHP,基于交换资源优化利用的原则,同时完成多个BHP的处理,进而达到在相同的交换结构(资源)下提高交换性能的目的。 More control of the present invention, packet aggregation switching scheduling method is characterized by the use of OBS system BHP first burst data arrives in the core switching node characteristics, not based on the implementation of each scheduling carried a BHP information, but to use a time a plurality of inner segments reached by BHP, based on the principle of exchange optimal use of resources, as well as complete a number of BHP process, and then to improve the exchange performance at the same exchange structure (resources) purposes.

光突发交换系统中涉及大量的光、电技术,目前虽然电处理的技术手段较成熟、实现成本也较低,但光器件的性能仍然不很理想,成本还很高,如高速半导体光开关、可调协激光器等。 Optical burst switching system involves a large number of optical, electrical technology, electrical processing technology, although currently more mature and achieve cost is low, but the performance of optical devices is still not very good, cost is also high, such as high-speed semiconductor optical switch , tunable lasers Association. 因此在OBS系统中,若采用较为复杂的电处理技术,能够在一定的光交换结构下带来更高的交换性能,具有明确的工程应用价值。 Therefore, OBS systems, the use of more complex electrical processing technology, can lead to higher switching performance under certain optical switching structure, with a clear value in engineering.

本发明提出了一种多控制分组汇聚交换调度方法,就是以有限程度的电处理复杂度的提高,换取光交换系统整体性能的提高。 The present invention provides a multi-control packet aggregation switching scheduling method, is to a limited degree of electrical treatment increased complexity of optical switching exchange to improve the overall system performance. 比较OBS核心交换系统与传统电交换系统,可发现OBS系统在很多方面受限于光器件的性能,如无法时隙化操作,采用内部交换加速的代价过高,只能使用光延时而不是真正意义上的光缓存等;但OBS也有其固有的优势,如端到端业务的全光传输/交换,每个端口(或每根光纤)有多个可用波长,控制分组先于数据分组到达而完成对交换资源的预约等。 Comparison OBS core switching systems and conventional electrical switching systems, OBS systems can be found in many ways limited by the performance of optical devices, such as slot operation can not be used to exchange high cost of accelerated internal use only delay rather than light Light caching in the true sense; but OBS also has its inherent advantages, such as all-optical transmission-end business / exchange, for each port (or each optical fiber) has a plurality of available wavelengths, the control packet before the data packet arrives The completion of the exchange resources of appointments. 本发明所提出的调度方案就充分利用了OBS的这些优势,在我们已进行了的计算机仿真中,采用多控制分组汇聚交换调度方案的OBS系统,在相同的光交换网络配置、相同的交换负载率下,系统交换丢失率比采用顺序调度方案的系统可下降一个数量级,极大的提高了OBS系统的交换性能。 The present invention proposes a scheduling program to take full advantage of these advantages of OBS in computer simulations we have conducted, the multi-control packet aggregation switching scheduling scheme OBS system, in the same optical switching network configuration, the same switching load rates, exchange loss rates than systems using system order scheduling scheme can be reduced an order of magnitude, which greatly improves the switching performance OBS system.

附图及附图说明 Brief description and drawings

图1是OBS网络的构成示意图图2是本发明方法的流程图图3多控制分组汇聚交换调度功能框图图4一种基本的核心节点OBS光交换结构图5采用顺序交换调度方案的示意图图6采用多控制分组汇聚交换调度方案的示意图图7高优先级交换丢失率的仿真对比具体实施方式本发明所提出的交换调度方案可用于不同的光交换结构。 Figure 1 is a schematic diagram OBS network 2 is a flow diagram of a method of the present invention is more than three control packet aggregation switching scheduling a functional block diagram of a basic core node 4 optical switching structure OBS 5 sequential switching scheduling scheme schematic diagram 6 Simulation of multi-control packet aggregation switching scheduling scheme schematic diagram 7 high-priority exchange loss rate comparison embodiment of the present invention proposed scheduling scheme can be used to exchange different light switching fabric. 以下以图4所示的光交换结构给出本发明的一个实施例。 The following optical switching structure shown in Figure 4 shows an embodiment of the present invention.

由图4可知,OBS核心节点的结构主要包括光交叉连接矩阵、波长变换器和光延迟线。 Figure 4 shows the structure OBS core node includes optical cross-connect matrix, wavelength converters and optical delay line. 这种系统结构中,有4根输入光纤,每根输入光纤有9个波长,用于承载8路业务的突发数据和一个BHP信道。 Such a system configuration, there are four input optical fibers, each input fiber 9 has a wavelength of, for carrying eight burst data traffic channel and a BHP. 同样的有4根输出光纤,因此系统支持4组共32路突发数据的交换。 The same has four output fibers, the system supports four groups of 32-way exchange of data bursts. 图中每路仅给出了8个数据波长,它们在到达光交换结构时通过分波器连接到8个波长变换器;波长变化器的输出通过8个独立的可控光延迟线连接到公共的32×32无阻塞光交叉连接矩阵;矩阵的4组(每组8个)输出,要求组内输出波长不重叠,通过合波器连接到一根输出光纤。 Each figure gives only eight data wavelengths, when they are reaching the light switch fabric connected to eight by demultiplexer wavelength converters; wavelength varying output device connected to the public through eight separate controllable optical delay line The 32 × 32 nonblocking optical cross-connect matrix; matrix 4 groups (n = 8) outputs, the group requested output wavelength do not overlap, the optical fiber is connected to an output through the multiplexer.

在这种光交换结构下,假设当前输出光纤(端口)只有一个波长可用,因此所有发往该光纤的突发数据都必须由32×32光交叉连接矩阵的一个端口输出。 In this optical switching structure, assuming that the current output fiber (port) only one wavelength is available, so all burst data sent to the optical fiber must be a port of the optical cross-connect matrix output by 32 × 32. 此时采用现有的顺序交换调度方案时,由于发生冲突而导致部分突发分组无法转发的情况示于图5,其中假设Burst1(对应于BHP1)的优先级低于Burst2(对应于BHP2)。 In this case the use of existing sequential exchange scheduling program, which led to the case because of conflicts burst packet can not be forwarded part shown in Figure 5, where it is assumed Burst1 (corresponding to BHP1) lower priority than Burst2 (corresponding to BHP2). 由图5可见,BHP1(假设来自输入光纤1)先到达,因此预约了交换资源(交换矩阵的输出端口);BHP2(假设来自输入光纤2)到达时,查找资源表无空闲(交换矩阵的输出端口被Burst1占用、输入端口光延迟线达最大值),因此Burst2将被丢弃。 It is seen from Figure 5, BHP1 (assume from the input fiber 1) first arrive, so a reservation exchange resources (switching matrix output port); BHP2 (assume from the input fiber 2) arrives, look no idle resource table (switching matrix output Burst1 port is occupied, the input port optical delay line up to the maximum), so Burst2 will be discarded. 此时不仅高优先级的分组(Burst2)没有得到交换,而且被丢弃了。 In this case not only the high priority packet (Burst2) have not been exchanged, and was discarded.

采用本发明的多控制分组汇聚交换调度方法时,先将BHP按其Burst到达的时间排队,如果Burst有重叠,则再按优先级排队,当优先级相同时再按突发数据的长度排队。 According to the present invention, a multi-exchange aggregation control packet scheduling method, the first Burst BHP their arrival time queue, if Burst overlap, then priority queuing, when the same priority then burst length of the data queue. 从图6中可以发现,由于BHP1、BHP2....BHPn被同时处理(在同一个调度窗口内),此时先到的BHP1并没有马上预约交换矩阵出端口,而是BHP2预约到了该端口,使得Burst2首先得到交换;BHP1对应的突发数据分组预约到的是其入端口的光延迟线,而经延迟的Burst1与Burst2在交换矩阵出端口的竞争得到消除,两个Burst都得到顺利交换,提高了交换性能。 From Figure 6 can be found, due BHP1, BHP2 .... BHPn be processed simultaneously (in the same scheduling window), then the first of the BHP1 did not immediately make an appointment switching matrix output port, but BHP2 appointment to the port so Burst2 was first exchange; BHP1 corresponding burst data packet reservation to its ingress port optical delay line, and the delayed Burst1 and Burst2 in exchange matrix ports to get rid of the competition, two Burst have been successfully exchange , improve the switching performance.

基于图4所示的光交换结构,在采用现有的顺序交换调度方案和本发明的多控制分组汇聚交换调度方案时,高优先级业务的交换丢失率仿真结果见图7(仿真工具为Visual C/C++开发平台)。 Based optical switching structure shown in Figure 4, in order to exchange scheduling the use of existing programs and the present invention is a multi-control packet aggregation switching scheduling scheme, the exchange rate loss simulation result of the high priority traffic is shown in Figure 7 (simulation tool for Visual C / C ++ development platform). 在采用现有的顺序调度时,如果没有采用其他QoS机制,则所有业务其实都是平等地处理;采用本发明方法后,高优先级业务的丢失率大大下降,特别是在业务量较大时。 When using the existing scheduling order, if no use other QoS mechanisms, all businesses are in fact equal treatment; after the method of the present invention, the loss rate of high priority traffic declined significantly, especially in the larger volume of business . 另外,从图7中还可看出,随着汇聚时间的增加,丢失率也随之下降。 In addition, can also be seen from Fig. 7, with the increasing convergence of time, the loss rate is lower.

明显的,本发明提出的调度方案在系统负载率较大的情况下,交换丢失率可比现有的顺序交换调度方案下降近10倍。 Obviously, the present invention proposes a scheduling scheme in system load factor is larger, exchange loss rate comparable exchange existing order scheduling scheme has dropped by nearly 10 times. 这充分表明了本发明的良好效果。 This fully demonstrates the favorable effect of the present invention.

Citada por
Patente citante Fecha de presentación Fecha de publicación Solicitante Título
CN101060392B7 Sep 20068 Dic 2010华为技术有限公司A transmission service data dispatching multiplex method and system
CN101212389B30 Dic 200623 Jun 2010华为技术有限公司;电子科技大学Outburst convergence control method, and corresponding device and communication device
CN101212819B27 Dic 20068 Dic 2010华为技术有限公司;电子科技大学Flow forecast based periodical and adaptive convergence method and system
CN101370314B10 Oct 200810 Ago 2011江苏大学Resource reservation method based on intra-domain network state perception in optical burst exchange
CN101473236B *15 May 200718 May 2016贺利实公司用于基于入站内容的QoS的方法和系统
CN101998194A *5 Nov 201030 Mar 2011电子科技大学Implementation method for optical burst switching network (OBS) resource reservation protocol for supporting multi-service fusion transmission
CN103716721B *30 Sep 201317 May 2017株式会社Ntt都科摩向分组光网络的目的地节点提供外部光数据分组的方法和系统
US805028412 May 20091 Nov 2011Huawei Technologies Co., Ltd.Method, device and communication equipment for controlling a burst assembly
US805953520 May 200915 Nov 2011Huawei Technologies Co., Ltd.Method and core router for delaying burst
US943836730 Sep 20136 Sep 2016Ntt Docomo, Inc.Method and system for providing an external optical data packet to a destination node of a packet optical network
WO2008092325A1 *30 Oct 20077 Ago 2008Huawei Technologies Co., Ltd.Method and core node for realizing burst packet delay
Clasificaciones
Clasificación internacionalH04L12/751, H04L12/861, H04Q3/52, H04L12/24, H04J14/02, H04L12/26
Eventos legales
FechaCódigoEventoDescripción
30 Mar 2005C06Publication
1 Jun 2005C10Request of examination as to substance
16 May 2007C14Granted
16 Mar 2011C17Cessation of patent right