Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberCN1183190 A
Publication typeApplication
Application numberCN 96193597
PCT numberPCT/GB1996/000536
Publication date27 May 1998
Filing date8 Mar 1996
Priority date8 Mar 1995
Also published asCA2214839A1, CA2214839C, EP0813783A1, US6452905, WO1996027966A1
Publication number96193597.9, CN 1183190 A, CN 1183190A, CN 96193597, CN-A-1183190, CN1183190 A, CN1183190A, CN96193597, CN96193597.9, PCT/1996/536, PCT/GB/1996/000536, PCT/GB/1996/00536, PCT/GB/96/000536, PCT/GB/96/00536, PCT/GB1996/000536, PCT/GB1996/00536, PCT/GB1996000536, PCT/GB199600536, PCT/GB96/000536, PCT/GB96/00536, PCT/GB96000536, PCT/GB9600536
InventorsAJ史密斯, JL阿达姆斯
Applicant牛津布鲁克斯大学, 英国电讯公司
Export CitationBiBTeX, EndNote, RefMan
External Links: SIPO, Espacenet
Broadband switching system
CN 1183190 A
Abstract  translated from Chinese
在交换异步传递数据信元的宽带交换系统中,动态带宽控制器(DBC)控制要送到系统输入端口的数据信元的申请,及若干传输终端系统提供的数据信元。 In exchange asynchronous data cell passing data cells broadband switching system, dynamic bandwidth controller (DBC) control system to be sent to the input port of the application data cells, and a number of transmission terminal system provides. 当终端系统开始传输数据信元时,DBC检测输入信元存在与否,并且向系统的连接允许控制(CAC)申请带宽。 When the terminal system begins transmitting data cells when, DBC detects the presence or absence of input cells, and to the system of connection admission control (CAC) application bandwidth. 交换系统保存一个与输入端口相连的若干信源及其预定最长传输时间和最大传输速率的表。 Switching system saves several sources connected to an input port and a predetermined maximum transmission time and a maximum transfer rate tables. 当在输入端口检测到来自若干信源之一的信元时,DBC与CAC合作在交换系统中分配足够的带宽以使相应长度的消息能在相应最长传输时间内传完。 When the input port is detected from one of several sources of cells, DBC and CAC cooperation allocate sufficient bandwidth switching system so that the corresponding length of the message can pass completely within the corresponding maximum transmission time. 信源也确保有一个最小传输速率。 Source also ensure that there is a minimum rate.
Claims(10)  translated from Chinese
1.一个宽带交换系统至少有一个输入端口连到相应的信源,至少有一个输出端口连到选定的信号接收系统,该交换系统至少有一个交换机把携带信息的异步传递数据信元从输入端口传到输出端口,有一个通过上述交换机接受并建立输入端口和输出端口之间连接的系统控制装置,有一个带宽控制装置,检测输入端口收到的信元,并自动响应信元检测,以使系统控制装置分配足够的带宽以便能在预定时间内传完预定数目的信元。 1. A broadband switching system having at least one input port connected to a respective source, at least one output port connected to a selected signal receiving system, the switching system having at least one switch of the information-carrying asynchronously transferred data cells from the input Port reached the output port, and having a receiving system via the switch and establish the connection between the input port and an output port of the control device, there is a bandwidth control means for detecting an input port to receive cell, cell detection and automatic response to the system control means to allocate sufficient bandwidth in order to complete a predetermined number of cells within a predetermined time pass.
2.根据权利要求1的系统,其特征在于保存预定时间周期与信源的关系表的表存贮装置,时间周期是预定的传输一组预定数目信元的最长时间;带宽控制装置,用于识别被检测到信元的源,从而确定传送被检测到信元的最长时间。 Bandwidth control means; 2. The system of claim 1, wherein the memory means to save the table with the source for a predetermined time period of relational tables, the predetermined time period is the longest time a group of a predetermined number of transmission cells to identify the source of the detected cells to determine the cell transfer is detected to the maximum time.
3.根据权利要求2的系统,其特征在于,存贮装置设法保存包括消息长度的表格,该长度与相应的信源及构成消息的预定最大信元系数相关。 3. The system of claim 2, wherein the memory means includes a table trying to save the message length, the length of the corresponding source and a predetermined maximum cell factor constituted related news.
4.根据权利要求2或3的系统,其特征在于存贮装置设法保存包括带宽值的表,该带宽值是相应信源的预定最大带宽分配值。 4. The system of claim 2 or 3, wherein the memory means to try to save the table include bandwidth value The bandwidth value is a predetermined maximum bandwidth corresponding source allocation value.
5.根据权利要求2至4任一条的系统,其特征在于最长传输时间是根据将一条消息传送到输出端口的预定结束时间计算出来的。 5. any of 2 to 4, a system according to claim wherein the maximum transmission time is calculated based on the transmission of a message to the scheduled ending time of the output port.
6.根据权利要求2至5任一条的系统,其特征在于带宽控制装置周期性地监视已开始发送的消息的信元数,并将信元计数通知系统控制装置。 6. 2 to 5, a system according to any one of the preceding claims, characterized in that the bandwidth control means periodically monitor the message has been sent starting cell number and the cell count notifies the system control unit.
7.根据权利要求2至6任一条的系统,其特征在于系统控制装置设法计算出与已识别信源相关的时间周期内传输完一条消息所要求的带宽值。 7. 2 to 6, a system according to any one of the preceding claims, wherein the system control means trying to figure out the bandwidth value in the identified source associated with the time period required for the transmission of a message to finish.
8.根据权利要求2至7任一条的系统,其特征在于带宽控制装置设法检测信源的非活动周期,测量非活动周期的长度,并且,当非活动周期超过了预定的超时长度后,收到了信元时,就认为该信元是新消息的开始。 8. to either 2 7 A system according to claim, characterized in that the bandwidth control means trying to detect the source of the inactive period, measuring the length of the inactive period, and when the period of inactivity exceeds a predetermined length of time-out after close to the cell when the cell is considered the beginning of a new message.
9.上述任一条权利要求的系统中使用的带宽控制装置。 Bandwidth used in any of the preceding claims strip system control devices.
10.操作将携带信息的异步传递数据信元从系统输入端口经过至少一个交换机送到系统输出端口的宽带交换系统的方法,方法包括:保存与能连到输入端口的信源相关的预定传输时间和预定消息长度的信息,检测从上述连在输入端口的源送来的信元,并且自动响应该检测,分配该信元一个足够的带宽,允许与上述消息长度相应的信元数在上述预定传输时间内传完。 10. The operation of the information carrying asynchronously transferred data cells from the cell system through the at least one input port to the broadband switching system switches the output port of the system, the method comprising: saving and can be connected to the input port of the source associated with the predetermined transmission time and a predetermined message length information detection from the source connected to the input port of the cell sent and automatically in response to the detection, assign the cell a sufficient bandwidth to allow the above message length corresponding to the predetermined number of cells in End transmission within the transmission time.
Description  translated from Chinese
宽带交换系统 Broadband switching system

本发明涉及一个用于交换异步传递的数据信元的宽带交换系统,及一种交换异步传递的数据信元的方法。 The present invention relates to a broadband switching system for the switching of asynchronously transmitted data cells of, and a method for transmitting asynchronous data cell exchange.

在交换异步传递数据信元的宽带交换网中,一个预定的带宽被分配给连接两个用户的传输信道。 In exchange asynchronous transfer data cells broadband switching network, a predetermined bandwidth is allocated to two users connected to the transmission channel. 在一些已知的这类系统中,信道在一段特定的时间内被占用,事实上是以租用专线的方式被占用,而且是用人工方式根据特定终端和传输流量大小建立或更改一个连线。 In some known such systems, the channel is occupied for a specified period of time, in fact, it is a way of leased lines are occupied, and is used to create or change a connection manually depending on the specific size of the terminal and transfer traffic. 结果是,用户通常要付一个固定比率的费用来作为这条连接的总费用的一部分,而这费用与该连接是否被使用无关。 As a result, users often have to pay a fixed rate of cost as part of the total cost of this connection, which costs associated with the connection is being used irrelevant.

现在提出了另一类系统,这类系统可以在拨号的基础上建立连接,要求终端设备配有建立连接的附属装置,这些装置能发送信令请求,而且能响应网络发出的类似信令。 Another system now proposed, such systems can be established on the basis of dial-up connection, the request terminal device is equipped with an accessory device to establish a connection, these devices can send a signaling request, and can respond to similar signaling network issued.

专用通信网占用永久电路被广泛应用。 Private communication network occupies a permanent circuit is widely used. 对这种永久电路的需求可望增加到包括大于2Mbit/s的宽带速率,这些电路传送的是多个信源复用而成的业务量,这些信源可能具有突发性,也可能以恒定速率发送但对时延敏感,如话音传输和恒速图象。 The demand for such circuits is expected to include a permanent than 2Mbit / s broadband speeds, the amount of these circuits is transmitted traffic multiplexed from a plurality of sources, these sources may have a sudden, it may be a constant rate sent but delay-sensitive, such as voice transmission and constant speed images.

异步传递模式(ATM)的信元都有48字节的固定长度的信息段,该信息段能传送用户业务流或起源于用户的控制信息(信令),这两种类型的数据传输由信头中的虚路径(VP)和虚通道(VC)区分。 Asynchronous transfer mode (ATM) cell has a 48-byte fixed-length pieces of information, that piece of information can be conveyed to the user traffic or the origin of the user control information (signaling), these two types of data transmission by the letter head of the virtual path (VP) and Virtual Channel (VC) distinction. ATM信头中的另一个段是信元丢失优先级,该段区分高优先级的信元和低优先级的信元。 Another section of the ATM header is a cell loss priority, the paragraph to distinguish high-priority and low-priority cell cell. 当出现拥塞时,低优先级的信元先被丢弃。 When congestion occurs, the low priority cells to be discarded.

在基于ATM的网络中,专用电路的路由、要求的带宽和业务质量(QOS)由网络管理进程设定。 In the ATM-based network, dedicated circuit routing, bandwidth and quality of service (QOS) requirements set by the network management process. 专用电路称为永久虚电路(PVCs),因为和存贮在交换机中的决定路由、带宽和QOS的信息相关的不是一条实际的物理电路,而是VP/VC值或者称为“标号”。 Dedicated circuit called PVC (PVCs), and stored in the switch because the routing decisions, the bandwidth and QOS information is not related to an actual physical circuit, but VP / VC value or called "label."

所有已知的永久电路的缺点是即使用户不传任何东西时带宽也保持分配给电路。 All disadvantage of the known permanent circuits is even when the user does not pass anything bandwidth remains assigned to the circuit. 这就意味着用户所付的费用比在需要时带宽才可用所付的费用要高。 This means that the user fee paid than the bandwidth available when needed to pay higher costs. 这里假设收费与预定带宽相关,而事实不一定如此,因为公用网操作员可以根据虚电路收费。 It is assumed that charges associated with the predetermined bandwidth, but the fact is not necessarily the case, because the public network operators can charge based on VC. 然而,根据预定带宽收费在将来会成为一个重要因素。 However, according to a predetermined bandwidth charges in the future will be an important factor.

普遍作法是建立一条永久虚电路,以便它只在一天的某几个小时可用,或者是一周的某几天可用。 Common practice is to create a permanent virtual circuit, so that it only at certain hours of the day are available, or are available on certain days of the week. 这种作法的困难在于它不允许用户快速改变使用模式,只能粗略地反映用户的使用要求。 The difficulty with this approach is that it does not allow the user to quickly change the usage mode, only roughly reflect the requirements of users.

第二种建议是向用户提供一条和网管平面相通的独立的信道,从而允许一条永久虚电路可被重新配置。 The second recommendation is to provide an independent and network management plane communication channel to the user, allowing a PVC can be reconfigured. 这种作法的困难在于用户得等一段时延后才能开始使用该虚电路。 Difficulty with this approach is that users have to wait for some time delay to start using the virtual circuit.

第三种建议是在网络的每个交换节点引入一种设备,该设备识别快速资源管理信元,指示某带宽值应分配给某电路。 The third proposal is to introduce a device at each switching node in the network, the device quickly identify resource management cell, indicating a bandwidth value should be assigned to a circuit. 这种作法的困难在于没有一个国际认可的关于不同厂家生产的交换设备都可识别的带宽请求信元的标准。 The difficulty with this approach is that there is no internationally recognized on the exchange of equipment from different manufacturers can identify the bandwidth request standard cell.

根据本发明的一方面所说,一个宽带交换系统至少有一个入口连到相应的信源,至少有一个出口连到选定的信号接收系统,该交换系统对从入口传送到出口的携带信息的异步传递数据信元至少经过一个交换机的交换,有一个通过上述交换机接受并建立从入口到出口的连接的系统控制装置,有一个检测在入口收到的信元的带宽控制装置,而且带宽控制装置能够自动响应信元检测,以使系统控制装置分配足够的带宽,以便能在预定的时间内传输预定数目的信元。 According to an aspect of the present invention said, a broadband switching system having at least one inlet connected to the corresponding source, at least one outlet connected to a selected signal receiving system, the switching system transferred from the inlet to the outlet of information-carrying asynchronous transfer through the exchange of data cells at least one switch, there is an acceptance by the above switch and establish a connection from the inlet to the outlet of the system control device, there is a bandwidth control device detects cell entry received, and bandwidth control device to automatically respond to cell detection, to cause the system control means to allocate sufficient bandwidth in order to transmit a predetermined number of cells within a predetermined time.

该系统优选地包括:表存贮装置,用于存放预定时间周期与信源的关系的表,时间周期是预定的传送一组预定数目的信元的最长时间;带宽控制装置,用于识别被检测到信元的源,从而确定传送该被检测到信元的最长时间。 The system preferably includes: table storage means for storing a predetermined time period table relationships with the source, the time period is the longest scheduled delivery of a set of a predetermined number of cells; bandwidth control means for identifying source cells are detected in order to determine which is the longest transmission detected cell.

这样,就有可能动态地调整某一通道上传送的信元在系统中的可用带宽,而不必事先发送一个请求带宽的信号才能在系统中发送信元,以提供一个有保证的最大传输时间。 Thus, it is possible to dynamically adjust the cell in the system on the available bandwidth of a transmission channel, without having to send a request in advance of bandwidth in the system to send a signal to the cell, to provide a guaranteed maximum transmission time.

更进一步,存贮表包括消息的长度值,该值与相应的信源相关并且是构成消息的预定最大信元数。 Furthermore, including the length value of the message store table, the value associated with the respective sources and is the predetermined maximum number of cells composing messages. 这样,根据存贮的最大消息长度值和最大传输时间,系统控制装置能计算出信元速率(等价于带宽),在该速率下,消息能及时传输完毕。 Thus, according to the stored maximum message length value and maximum transmission time, the system control means can calculate the cell rate (equivalent to the bandwidth), rate at which messages can be transmitted promptly completed. 信息的这两项性质通常是由用户预先登记的。 Both the nature of the information is usually pre-registered by the user.

存贮表还可以包括一个带宽值,该带宽值是相应信源的预定最大带宽分配值。 Storage table may also include a bandwidth value The bandwidth value is a predetermined value corresponding to a maximum bandwidth allocation information source.

同样,该值通常也是由用户预先登记的,而且常常与信源的最大传输带宽一致。 Again, this value is usually also pre-registered by the user, and often the largest source of transmission and consistent bandwidth. 这就避免了信源被分配了一个它不能充分利用的带宽,当系统希望充分利用带宽时,如果不能按需分配带宽,将导致预定数目的信元不能在规定的时间周期内传完,或者导致可用带宽没有被充分利用。 This avoids the source is allocated bandwidth that it can not take full advantage of, when the system hope to take full advantage of the bandwidth, bandwidth on demand, if not, will lead to a predetermined number of cells can not pass within a specified time period finished, or lead to the available bandwidth is not fully utilized.

最大传输时间可以选择性地根据将一条消息传送到出口的预定结束时间计算出来。 The maximum transmission time can be calculated based on the selective transmission of a message to a predetermined end time exports. 这样,用户可以发送一个消息来指定该消息必须到达的时间。 Thus, the user can send a message to specify that the message must arrive. 系统控制装置和/或带宽控制装置可以根据当前时间和预定结束时间计算出最大传输时间。 The system control unit and / or bandwidth control means may calculate the maximum transmission time based on the current time and the scheduled ending time.

系统控制装置最好能设法计算出一个要求的带宽值,在该带宽下,一条消息能够在相应信源的允许时间周期内传输完毕。 The system control means preferably managed to calculate a desired bandwidth value in the bandwidth of a corresponding message can be the source of transfer is completed within the allowable time period.

带宽控制装置可以周期性地监视一条已开始发送的特定消息的信元数,并将信元计数通知系统控制装置。 Number of cell bandwidth control means may periodically monitor a specific message that has begun to be transmitted, and the cell count notifies the system control unit. 然后,系统控制装置根据能否保证在最大传输时间内完成传输决定给该信源分配的传输该消息的带宽是否需要调整。 Then, the system control means can ensure the completion of transmission within the maximum transmission time decided to allocate the bandwidth of the source of transmission of the message whether adjustments are needed.

信源很可能不充分利用分配的带宽,这时,系统控制装置不仅应做到分配足够的带宽以使消息及时传完,而且系统控制装置有必要监视消息的传送进程。 The source most likely not take full advantage of the bandwidth allocated, then the system control device should be done not only allocate enough bandwidth to make the message transmission finished in time, and the system control unit monitors the transfer process is necessary message. 用带宽控制器统计信元就是一种方法。 Bandwidth controller statistics cell is a way.

带宽控制器可以检测信源的非活动周期,测量非活动周期的长度,当输入端口收到的信元在非活动周期之后,且超过了预定的时长,就认为该信元是一个新消息的开始。 Bandwidth controller can detect the source of non-activity cycle, measure the length of the inactive period, when the input port receives the cell after a period of inactivity, and exceeds a predetermined length of time, it is believed that the cell is a new message begin.

带宽控制装置最好设法检测到信元送到系统输入端口的速率,而系统控制装置根据正在传输的信源应该在指定路由上分到最大的系统可用带宽的原则提供带宽值。 Bandwidth control means is preferably supplied to the system trying to cell rate of the input port is detected, and the system control apparatus according to the source being transmitted bandwidth values should be assigned to the principle of providing maximum system bandwidth available on the specified routes. 这样的带宽分配依赖于系统控制装置分配带宽时系统的流量情况。 Such a bandwidth allocation depends on the system control means to allocate bandwidth when the traffic conditions of the system. 然而,在该发明优选的实施例中,从一个给定信源送到系统输入端口的信元速率受到监测,如果分配的带宽超过了以监视速度传输信元所需的带宽,系统控制装置在带宽控制装置的启动下分配一个较小的带宽给这些信元,过了预定的时间间隔后,其他用户可用这部分带宽。 However, in the preferred embodiment of the invention, from a given source to the input port of the system being monitored cell rate, if the allocated bandwidth exceeds the speed of transmission cells to monitor the bandwidth required, the system control device Under the bandwidth allocation control means start a smaller bandwidth to these cells, and after a predetermined time interval after this part of the bandwidth available to other users.

带宽控制装置可能包括:反馈装置,用来将最大输出信元速率信号传回给信源,反馈信号指出的信元速率依赖于分配的带宽或者与分配的带宽相匹配。 The bandwidth control means may include: feedback means for the maximum output cell rate signal back to the source, the feedback signal indicates that the cell rate depends on the bandwidth or bandwidth allocation and assignment of matches.

带宽控制装置还可以读取输入信元的带宽识别部分,并且使系统控制装置根据识别的带宽给信元分配带宽。 Bandwidth control device may also read bandwidth input cell identification section, and the system control apparatus according to the cell identification bandwidth allocation bandwidth. 带宽控制装置也可以根据缺省带宽限制系统接受输入信元的速率,直到系统控制装置开始分配带宽。 Bandwidth control device can also accept input cell rate based on the default bandwidth limitations of the system until the system control means allocate the bandwidth. 在这种情况下,带宽控制装置中有缓冲器,分配带宽时,信元存在缓冲器中。 In this case, the bandwidth control apparatus in a buffer, when allocating bandwidth, the presence of the cell buffer. 这样的缓冲也可用于其它场合,当信源的发送速率超过了给定时间系统可接受的速率时,就可以用缓冲。 Such a buffer can also be used for other occasions, when the source of the transmission rate exceeds a given time of the system at an acceptable rate, you can use a buffer. 事实上,缓冲器最好能检测出何时缓冲器充满到了门限值,并且有一个反馈装置,能够响应缓冲器的状态,向与输入端口相连的信源发一个重发的信元速率反馈指示信号。 In fact, the best buffer can detect when the buffer is full to the threshold, and there is a feedback device, capable of responding to a buffer state, and input to the source connected to the port to send a retransmission of a cell rate feedback indication signal. 最好,检测出缓冲器充满到门限值时,能向信源发送一个减少的信元速率反馈指示信号,信源减少发送信元速率,以避免缓冲器溢出,丢失数据。 Preferably, the detected value of the threshold when the buffer is full, can send a reduced cell rate indicator signal feedback to the source, the source to reduce the transmission cell rate to avoid buffer overflow, loss of data.

优选实施例的带宽控制装置能够自动为来自若干个信源的信元分配带宽,包括一个活动检测器和一个控制器,活动检测器能检测到输入信元,并识别发送该检测到信元的信源,控制器能够响应上述检测与识别,并向系统控制装置发送一个带宽请求消息,以便分配带宽给被识别信源发送信元。 Bandwidth control apparatus of the preferred embodiment can automatically from a number of sources of the cell to allocate bandwidth, including an activity detector and a controller, motion detector can detect incoming cells, and to identify the sender of the detection cell sources, the controller can respond to the detection and identification, to the system control device sends a bandwidth request message to assign bandwidth to transmit the identified source cell.

活动检测器包含一个信元计数器,用于统计与带宽控制器相连的相应信源发出的信元数。 Activity detector comprises a cell counter for the appropriate number of cells a source connected to the controller bandwidth statistics issued. 这样得到的信元计数可用作产生用户帐单或其它目的的收费信号。 Thus obtained can be used to produce cell count signal from the user fee bill or other purposes.

根据本发明的另外一方面,本发明还包括从属权利要求10所声明的方法。 According to a further aspect of the present invention, the present invention further comprises a method of statement 10 dependent claims.

现在,举例详细说明本发明,参见附图。 Now, detailed examples illustrate the present invention, see FIG. 图中:图1是根据本发明的宽带交换系统图;图2是根据本发明的另一种宽带交换系统图;图3是宽带交换系统的部分图,说明几个终端系统如何共享一个带宽控制器;图4是图1和图2中使用的带宽控制器的方框图;图5是图4中所示活动检测器模块的功能规格和描述语言(SDL)图;图6-1和6-2是图4中所示控制器的SDL图;图7是资源管理(RM)数据信元格式图;图8是图4中所示反馈模块的SDL图; In which: Figure 1 is a broadband switching system diagram of the present invention; Figure 2 is a diagram of another broadband switching system of the present invention; Figure 3 is a partial diagram of broadband switching system, explained how the system several terminals to share a bandwidth control device; Figure 4 is a block 2 in Figs. 1 and used bandwidth controller; Figure 5 is a functional specification and description language activities shown in Figure 4 detector module (SDL); Figures 6-1 and 6-2 the controller shown in FIG. 4 is an SDL diagram; FIG. 7 is a resource management (RM) data cell format; Figure 8 is a SDL diagram shown in Figure 4 feedback module;

图9是图4中所示的带宽控制器中使用的缓存器图;图10是缓存器的SDL图;图11是图6中所示的整形器/复用器模块及其与缓存器的连接图;图12-1,12-2,12-3是整形器/复用器模块的SDL图。 FIG. 9 is a buffer bandwidth controller shown in Fig. 4 used; Figure 10 is a buffer SDL; Figure 11 is a diagram of shaper / multiplexer module and the buffer in Fig. 6 connection; Figure 12-1,12-2,12-3 is shaper / multiplexer module of SDL diagram.

在优选形式中,本发明涉及一个宽带交换网络,该宽带交换网络构成了使各终端系统之间能传送异步传递信元的公用交换网络的一部分。 In a preferred form, the present invention relates to a broadband switching network, a broadband switching network which constitutes a part of the system can be transmitted between the respective terminal cells asynchronous transfer public switched network.

公用网10包含若干在异步传递模式(ATM)下运行的交换机,见图1。 Public network 10 includes a number of asynchronous transfer mode (ATM) switches running, see Figure 1. 在这个简单的例子中,交换机包括两个本地交换机12和一个转接交换机16,每个本地交换机12都有一个端口与相应的终端系统14相连,转接交换机16与本地交换机12互连。 In this simple example, the switches include two local switches 12 and a transit exchange 16, 12 has a port connected with the respective terminals of each local exchange system 14, 16 and 12 interconnected local exchange transit exchange. 与交换机相关联的有一个连接允许控制功能18(CAC)和一个动态带宽控制器(DBC)20,DBC20用于控制交换机12进入网络的流量。 There switch associated with a connection admission control function 18 (CAC) and a dynamic bandwidth controller (DBC) 20, DBC20 for controlling the switch 12 into network traffic. 这个交换机12还包括一个用法参数控制装置22,该装置用于动态调整从终端系统14送入输入端口24的数据信元的优先级。 The switch 12 also includes a usage parameter control means 22, the means for dynamically adjusting the system 14 is fed from the terminal data input port 24 of the cell priority.

实际上,网络10会有多个本地交换机和转接交换机12,16和若干DBC20,这些交换机和DBC互连形成一个网络,这个网络有若干个端口和若干个终端系统相连,比如端口24连在终端系统14上。 In fact, the network will have more than 10 local exchanges and transit exchanges 12, 16 and several DBC20, these switches and DBC interconnected to form a network that has several ports and several terminals connected to the system, such as port 24 connected terminal system 14. 通过DBC20;公用网10能够提供可用比特率(ABR)业务,DBC的作用是:检测送到输入端口24的信元,并且对检测作出自动响应,使CAC18分配足够的带宽,保证要传送到终端系统的信元在预定时间内送到。 By DBC20; public network 10 to provide available bit rate (ABR) service, DBC role is: to detect the cell to the input port 24, and an automatic response to the detection of so CAC18 allocate sufficient bandwidth to ensure to be transmitted to the terminal system cell to within a predetermined time. 通常,带有ABR业务的终端系统14都有一个固定的DBC20。 Typically, a terminal systems with 14 ABR service has a fixed DBC20. 每个本地ATM交换机12可能有不止一个DBC20。 Each local ATM switch 12 may have more than one DBC20. 终端系统还可能连接一个备份DBC(图中未画),以防出错。 End systems may also be connected to a backup DBC (not shown in figure), in case of error.

数据以异步传递模式(ATM)信元传送的,每个ATM信元有一个48字节的信息段及一个5字节的信头,信头中包含便于信息在网络中传输的信息。 Data in asynchronous transfer mode (ATM) cell transfer, each ATM cell has a 48-byte information field and a 5-byte header, the header is included to facilitate the transmission of information in the network information. 这样,路由是以信元为基础,一个信元一个信元控制的,数据可以在一条链路的若干通路和若干时隙上传输,ATM信元是根据信头中定义的虚路径和虚通路传送的。 This route is a cell-based, a cell of a cell control, the data available on the number of paths of a link and a number of time slots transmission, ATM cell header based virtual path and virtual channel defined transmission.

虚路径和虚通路由虚路径标志(VPI)和虚信道标志(VCI)确定的,5字节信头中的VPI和VCI有效地确定了终端系统之间的连接,使构成一条普通消息的所有信元通过同一条连接传送。 Virtual path and virtual channel mark by the virtual path (VPI) and virtual channel identifier (VCI) determined 5-byte header of VPI and VCI effectively determine the connection between the end-systems, constituting a common message to all Cell transmitted through the same connection. 如图1所示,进入公用网10的ABR业务流量根据它们的VPI和VCI选择路由,通过DBC20送到外部路由。 As shown in Figure 1, into the public network 10 ABR traffic routing according to their VPI and VCI, to an external routed through DBC20. 从DBC20处开始,每个虚路径和虚通路的流量都限制CAC20确定的信元速率(此后称为“CR”)以内。 Starting DBC20 at each virtual path and virtual channel flow are determined to limit cell rate CAC20 (hereinafter referred to as "CR") or less.

另外一个说明的例子见图2。 Another example of a illustrated in Figure 2. 在这种情况下,终端系统14A受到多个DBC的控制。 In this case, end-system 14A by a plurality of DBC control. 实际上,两个终端系统14A,14B是通过两个公用网10-1,10-2进行编路的,每个网络10-1,10-2都有各自的DBC 20-1,20-2,分别根据各自网络中的连接允许控制功能(CAC)18-1,18-2分配的带宽限制进入网络的流量。 In fact, the two end systems 14A, 14B is to encode the road through two public networks 10-1, 10-2, 10-1 and 10-2 each network has its own DBC 20-1,20-2 , respectively, according to their network connection admission control function (CAC) 18-1,18-2 allocated bandwidth limit traffic entering the network. 每个DBC20-1,20-2还负责向终端系统14A建议当前可申请的CR。 Each DBC20-1,20-2 is also responsible for the current proposal to end the system 14A can apply for CR.

在图1和图2的各系统中,用户预先登记终端系统的最大传送速率,传输一个文件的最长时间,还可以选择登记希望发送的文件的最大长度,以及预定的消息到达时间。 In the system of Figure 1 and Figure 2, the maximum transfer rate of pre-registered user terminal system, the maximum time to transfer a file, you can also select the maximum length of the registration document you want to send, and the scheduled arrival time of the message.

用户并不一定传送登记的文件长度,也可以不以最大速率传送,而有可能传送一个比最大长度小的文件,而且/或者以低于最大值的速率传送。 The user does not have file transfer length register to be not at the maximum transfer rate, and it is possible to transmit smaller than the maximum length of a document, and / or less than the maximum rate transmission. CAC 18,18-1,18-2给信元分配一个足够在最长传输时间内传送完文件的带宽。 CAC 18,18-1,18-2 assigned to the cell transmitted within a sufficient time to complete the longest transmission bandwidth of the file. 此分配值送到DBC20,20-1,20-2,由DBC将CR值知传输终端系统14。 This value is assigned to the DBC20,20-1,20-2, the DBC will be the CR-known transmission terminal system 14.

几个终端系统或信源可能共享一个动态带宽控制器(DBC)。 Some end systems or sources may share a dynamic bandwidth controller (DBC). 例如,见图3,DBC 20-3与构成网络10的一部分的带宽ATM交换机12-3相连,三个信源14C的业务流量用输出缓存器28处理。 For example, in Figure 3, DBC 20-3 constituting part of the network bandwidth ATM switch 10 12-3 connected to traffic three sources 14C by processing the output buffer 28. DBC20-3能处理的信源数由链路速率L决定(信源数一定不能太多,否则链路速率总保持为决定可用速率的限制因子L)。 Source DBC20-3 number that can be processed by the link L rate decision (number of sources must not be too much, otherwise the link rate always remains available to determine the rate limiting factor L). 信源14C发出的ABR业务流量的合计速率一定不能超过L。 Total rate ABR traffic information source 14C issued must not exceed L. 这意味着如果每个信源的流量都有突发,那么输出缓存器28可能会有拥塞。 This means that if the flow rate of each source has burst, the output buffer 28 may be congested. 这种情况可以通过向终端系统14C发送一个信元速率(CR)的反馈信号,该反馈信号通过一般流量控制(GFC)信号发送,使每个信源都立即停止发数据。 This situation can be by sending a cell rate to the terminal system 14C (CR) of the feedback signal, the feedback signal through the general flow control (GFC) signaling, so that each source will immediately stop sending data.

当动态带宽控制器(DBC)与图1,2,3中的各系统合作运行时,它的主要功能如下:首先,它为输入的数据信元提供缓存,任一时刻缓存的程度根据信元传输,传输由上面所说的VPI和VCI信息确定。 When the dynamic bandwidth controller (DBC) and Fig. 1, 2, co-operation of each system, its main function is as follows: First, it provides caching for data input cells at any one time according to the cell level cache transmission, the transmission is determined by the above mentioned VPI and VCI information. DBC还控制或者说“整形”送入网络10的流量,以使信元速率等于当前的特定传输可使用的速率,信元速率与分配的带宽有关。 DBC further controls or "plastic" into the traffic network 10 to allow cell rate equal to the current rate of transmission can be used in a particular bandwidth related to cell rate and distribution.

CAC18(见图1)根据路由上正在进行的传输数量及在预定最长传输时间内完成传输的必要性确定路由的可用容量,合理地划分路由的可用容量并确定传输路由后,CAC18可以给特定传输分配带宽,即确定CR。 CAC18 (Figure 1) according to the number of transport routes and the necessity of being completed within a predetermined maximum transmission time to determine the route of transmission of the available capacity, a reasonable division of routes available capacity and determine the transmission route after, CAC18 can give a specific allocate bandwidth transmission, i.e., to determine CR.

当传输开始时,DBC开始检测,并立即向相应的终端系统14(见图1)发送一个低的缺省CR。 When transmission begins, DBC began testing, and immediately to the appropriate terminal system 14 (see FIG. 1) sends a low default CR. 这样总是允许建立连接,但是注意,低的缺省CR可能是零。 Always allow the establishment of such a connection, but note that the low default CR may be zero. 减少到低缺省CR确保了一个新的活的传输源在得到CAC18分配的带宽和该传输的CR之前,不会使系统10过载。 Reduced to the low default CR ensures that a new live transmission source before getting CAC18 allocated bandwidth and the transmission of the CR, the system 10 does not cause overload. 这种过载一般会使该传输发生信源丢失与离开DBC20的传输相关的业务流量经过整形后与缺省CR相匹配。 This overload may typically occur in the transmission loss and leaving the transmission source DBC20 traffic associated with the default CR after shaping match. 这就是DBC的第二个功能,即为了控制传输信元速率,向终端系统发送一个反馈信号。 This is the second function of the DBC, that in order to control the transmission cell rate to the terminal system sends a feedback signal. 实际上,每次CAC18获得一个新的传输CR时,一个CR建议信号都反馈给终端系统。 In fact, every time I get a new transmission CAC18 CR, the proposal signals a CR feedback to end systems.

使用DBC的预传输缓冲使相关的终端系统能够有充分的时间将输出调整到最新的CR反馈建议值。 Use DBC's pre-transfer buffer so associated terminal system can have sufficient time to adjust the output to the latest CR feedback recommendation. 这意味着DBC的缓存足可以容纳至少是DBC和终端系统之间往返时延这么长的时间内传输的额外信元数。 This means DBC cache enough to accommodate at least an extra number of cells within the round trip delay between the DBC and the end systems such a long time transmission. 但是,如果信元不断地以高于建议反馈CR值从终端系统14送入(例如,CR在传输途中丢失,或者终端系统出错),那么过量的信元将因为缓存器溢出而在DBC处丢失。 However, if the cell constant feedback CR value higher than the recommendation system 14 is fed from the terminal (for example, CR is lost in transit, or terminal system error), then the excess cells will be lost due to buffer overflow at the DBC .

在建议的DBC中也可以利用缓存器门限确定一个错误容限。 In the proposed DBC can also use the buffer threshold determining an error tolerance. 当某一给定传输的存贮信元数达到门限时,向终端系统重发一个反馈建议CR。 When a metadata stored letter given transmission reach the threshold, the end-system retransmission of a feedback suggestion CR. 这个功能也可用于终端系统的管制机制,以防止无论是终端出错还是故意违反约定的传输规则而造成的带宽的无效利用。 This function can also be used to end the system of control mechanisms to prevent invalid use either terminal error or deliberate breach of the contract of transport rules caused bandwidth. 这就避免了其它符合规则的终端系统的业务质量受到干扰。 This avoids the quality of service in line with the rules of other terminal system is disturbed.

现在详细说明模块DBC20,参见图4。 Now a detailed description module DBC20, see Figure 4.

图4中的DBC20是一个分立的设备,它有一个接收异步传递数据信元的输入端口30和一个向交换网络10中的交换机12或16(见图1和图2)发送信元的输出端口32。 Figure 4 DBC20 is a separate device, which has a receiving asynchronous transfer data cell input port 30 and a switch to the network 10 to exchange 12 or 16 (see Figures 1 and 2) transmission cell output port 32. 该设备另外还有一个接收从交换机12或16发回来的消息的输入端口34和一个向终端系统14(见图1)发送反馈消息的反馈输出端口35。 The receiving apparatus in addition to a hair from the back of the switch 12 or 16 messages to the input port 34 and an end-system 14 (see FIG. 1) sends feedback message feedback output port 35. 虽然DBC20是一个独立设备,图4也可以看作是表示一个大的数据处理设备的子系统的功能图,这个数据处理设备的大部分体现为软件功能块。 Although DBC20 is a standalone device, can also be seen in Figure 4 is a functional diagram of a large data processing device subsystems, most reflect the data processing device software function blocks.

作为用户信元流送到输入端30的输入信元首先被送到活动检测器36。 As a user cell stream to the input terminal of the first cell 30 is sent to the activity detector 36. 活动检测器的作用是向控制器38提供每个收到的传输的状态信息,每次传输由信头中的VPI和VCI标识。 Action activity detector is provided to the controller 38 for each transmission received status information, each transmission by the letter head of the VPI and VCI identification. 当一个传输先前是静的,而现在观察到一个与其有相同VIP和VCI的信元从终端系统送到输入端口30时,该传输被标识为活动的。 When a transmission previously quiet, but now we observe a VIP and VCI its same cell to the input port 30, which is identified as the active transmission from the terminal system. 通过信头中的差错校验段使活动检测器36与信头的起始同步。 By letter header error checking section 36 is synchronized with the start so that the header of the activity detector. 差错校验段有一定的冗余度,因此可用信头中的信息进行差错校验。 Error Checking section have some redundancy, so the header information available for error checking. 提供信头差错校验段的主要原因是确保信头的信息是正确的,也就确保信元不会被送到错误的地址去。 The main reason for providing the header error check section is to ensure that the header information is correct, also ensure that the cell will not be sent to the wrong address to go.

当一个传输先前是活动的,而经过一段时间t后,仍没有收到与其有相同VPI和VCI的信元时,这个传输就被认为是静止的。 When a transmission was previously active, but after a period of time t, the same does not receive its VPI and VCI of the cell, the transmission is considered to be stationary.

活动检测器为每个VPI/VCI对值保留一个定时器和状态表,最好把t设为几秒钟,这样当VPI/VCI对在几毫秒之内出现活动-静止-活动的变化时不会被检测到,从而传输仍被认为处于活动状态。 Activity detector for each VPI / VCI values are reserved for a timer and state table, the best t set for a few seconds, so that when the VPI / VCI to appear activities within a few milliseconds - still - change activities without It will be detected, so that the transmission is still considered active. 这样会降低从DBC20同CAC18发送消息的频率,以降低网络利用率为代价。 This will reduce the transmission of messages from DBC20 CAC18 same frequency, in order to reduce the cost of network utilization.

活动检测器36的另外一个功能是,从控制器38收到“开始信元计数”的请求之后,统计传输信元数。 Another function of the activity detector 36 is received from the controller 38 after the request "start cell count" the number of cells transport statistics. 这个信息可用来收费,也可使控制器38估计收到传输的实际信元速率。 This information can be used to charge, but also allows the controller 38 receives the estimated actual cell rate transmission.

活动检测器的伪码如下所示,相应的SDL示于图5。 Activity detector pseudo code shown below, the corresponding SDL is shown in Figure 5.

<pre listing-type="program-listing"><![CDATA[BEGIN {cell arrival} cell arrival from end-system read VC reset VC inactivity timer IF VC is newly acrive THEN update state table advise CONTROLLER of newly active VC ELSE IF counting.cells(VPI/VCI) THEN increment cell.count(VPI/VCI) ELSE do nothingENDBEGIN {VC timer exptres} VC inacuvitv umer expires indicauing quiet VC update state table advise CONTROLLER of quier VC counuring.cells:=FALSEENDBEGIN {receve a start.cell.count signal} receive a start.cell.count(VPI/VCI)signal from CONTROLLER cell.count(VPI/VCI):=0 counung.cells:=TRUE END BEGIN{cell.count timer expires} cell.count umer expires send cell.count(VPI/VCI)to CONTROLLER restart cell.count timer END]]></pre>就输入用户信元流来说,活动检测器36从收到的信元流中的每个信元的信头中读取VPI/VCI值,用此信息更新为每个VPI/VCI对保存的状态表。 <Pre listing-type = "program-listing"> <! [CDATA [BEGIN {cell arrival} cell arrival from end-system read VC reset VC inactivity timer IF VC is newly acrive THEN update state table advise CONTROLLER of newly active VC ELSE IF counting.cells (VPI / VCI) THEN increment cell.count (VPI / VCI) ELSE do nothingENDBEGIN {VC timer exptres} VC inacuvitv umer expires indicauing quiet VC update state table advise CONTROLLER of quier VC counuring.cells: = FALSEENDBEGIN {receve a start.cell.count signal} receive a start.cell.count (VPI / VCI) signal from CONTROLLER cell.count (VPI / VCI): = 0 counung.cells: = TRUE END BEGIN {cell.count timer expires} cell .count umer expires send cell.count (VPI / VCI) to CONTROLLER restart cell.count timer END]]> </ pre> to enter the user cell stream, the activity detector 36 from the received cell stream Each cell header reads VPI / VCI value, with this information update for each VPI / VCI to save the state table. 就与控制器38的通信来说,检测器36通知控制器每个VPI/VCI对的状态变化。 To communicate with the controller 38, the detector 36 informs the controller of each VPI / VCI to the state changes. 控制器通知活动检测器要用的定时器的时长t。 Controller notifies activity detector use the timer value t.

最好,所有的VPI/VCI对都使用相同的t值。 Preferably, all of the VPI / VCI to use the same t values. 每次定时器超时后,活动检测器36通知控制器38信元计数信息。 Every time the timer expires, the activity detector 36 informs the controller 38 cell count information. 为了确定到目前为止已经完成了多少传输,控制器38把信元计数传给CAC18。 In order to determine how far the transmission has been completed, the controller 38 cell count passed CAC18.

送到输入端口30的用户信元流的信元立即传送到缓存模块40,在此处信元按先进先出(FIFO)缓存队列的方式存贮,构成每个队列的信元都有一个指定的VPI/VCI对。 To the input port of the cell user cell stream 30 immediately transferred to the cache module 40, where the cell according to FIFO (FIFO) memory buffer queue manner constituting each cell has a specific queue The VPI / VCI pair. 检测器36不区分信元类型。 Detector 36 does not distinguish between cell types. 因此,任何数据信元的到达都被检测,而且无论控制或管理信元存在与否,都会潜在影响与其相关的VPI/VCI对的活动状态。 Therefore, to reach any data cells are detected, and whether the control or management cells exist or not, the potential impacts are associated with VPI / VCI active right. 缓存器40中的信元先送入整形器/复用器模块40,再经过输出端口32送入ATM交换机。 Cache 40 cell first into the shaper / multiplexer module 40, and then through the output port 32 into the ATM switch. 以后再详细说明缓存器和整形器/复用器40,42的运行。 Described in detail later re-buffer and shaper / multiplexer 40, 42 runs. 暂且知道缓存器模块在缓存队列达到预定的充溢门限时能够向控制器38发送信号即可。 Let us know in the cache buffer module overflowing queue reaches a predetermined threshold to be able to send a signal to the controller 38. 整形器/复用器42的功能是将信元从缓存器模块40取出并将其送到目的地。 Function shaper / multiplexer 42 is to the cell module 40 removed from the buffer and sent to the destination. 它包括复用器功能,且整形器保存了每个VPI/VCI对的CR值,以便从输出端口送出的信元流经过整形后,能确保由分配带宽所决定的通过网络的每个传输的相应路径的容量不会过载。 It includes a multiplexer function and the shaper save the CR value for each VPI / VCI pair to the output port from sending cell stream after surgery, to ensure that the allocated bandwidth is determined by each of the transmission network of capacity corresponding path is not overloaded. 控制器38还控制反馈模块44接收控制器及网络经过输入端口34送来的反馈信号,并经过输出端口35向终端系统14传输信号。 The controller 38 also controls a feedback module 44 receives the network controller and sent via the input port 34 of the feedback signal and the output port 35 via a signal 14 to the transmission terminal system. 缓存器、整形器/复用器和反馈模块40、42、44的功能在以后详细描述,现在首先说明控制器38。 Buffer, shaper / multiplexer, and feedback modules 40,42,44 function described in detail later, the controller 38 is now explained first.

控制器38的作用是通知CAC18任何VPI/VCI对的一个ABR类型的传输都应在系统中分配到或重新协商的一个带宽值。 The role of controller 38 is to inform the VPI / VCI for an ABR transmission type should be allocated to the system or re-negotiation of a bandwidth value any CAC18. 本实施例中,控制器向CAC18发送一个带宽重新协商信号,CAC18认为此信号的意义是,就传输而言,必须在预先登记的最长时间内完成。 In this embodiment, the controller sends a signal to renegotiate the bandwidth to CAC18, CAC18 meaning that this signal is to transmit, it must be completed within a pre-registered maximum time. 因此,CAC向接收带宽值的控制器发出一个带宽建议值。 Therefore, CAC issued a recommended value of the bandwidth controller bandwidth value to the receiver. (以后说明CAC的带宽分配策略)。 (Described later CAC bandwidth allocation strategy).

新的CR经过反馈送到终端系统,以及整形器模块42。 The new CR through the feedback system to the terminal, and a shaper module 42. 控制器38指挥活动检测器36对接收的信元计数,并且根据此计数值向CAC18周期性地发送信元计数信号。 The controller 38 directing the activity detector 36 received cell count, and periodically transmits a signal to the CAC18 cell count based on this count.

控制器38还负责监视传输获得大的有效带宽时用户的实际发送速率。 The controller 38 is also responsible for monitoring the transmission of large effective bandwidth obtained actual transmission rate of the user. 方法是:初始化活动检测器36中的定时器,该定时器的超时时间为几秒钟,超时到后,活动检测器向控制器38发送一个信元计数值。 Methods are: initialization activity detector 36 in a timer timeout a few seconds, after a timeout, the activity detector sends a cell count to the controller 38. 该信息经过处理后可以估计等价于用户实际发送速率的系统容量。 This information can be estimated after treatment system capacity equivalent to the actual sending rate of users. 如果该容量实际上小于当前分配容量,说明用户试图通过向输入端口30发送较小的信元流,以保证活动检测器将与其相关的传输视为活动的,从而保持一个大的系统容量。 If the capacity is actually less than the current allocation of capacity, indicating that the user tries to enter port 30 to send a small cell stream, in order to ensure the activity detector will be regarded as the activities associated with the transmission, thereby maintaining a large system capacity. 显然,这将导致网络没有被充分利用。 Obviously, this will cause the network has not been fully utilized. 当控制器38发现分配的容量与当前使用的容量不匹配时,它将采取措施。 When the controller 38 found allocated capacity does not match the current use of capacity, it will take measures. 首先,整形器42立即复位CR到该VPI/VCI对的实际CR值。 First, the shaper 42 is immediately reset to the actual CR CR values of the VPI / VCI pair. 然后,命令反馈模块44经输出端口35向用户发送此实际CR值。 Then, the command feedback module 44 via the output port 35 to send this practical CR value to the user. 最后,控制器38向CAC发一个带宽重新协商信号,该信号包括等价于用户当前使用容量的带宽值。 Finally, the controller 38 sends a bandwidth of the CAC renegotiation signal that comprises a user is currently using equivalent bandwidth capacity value. CAC可以在收到此信号后,将容量分给其他用户。 CAC may, after receiving this signal, the capacity allocated to other users. 另外一种措施是,当发现不匹配时,给整形器和反馈模块42,44分配缺省CR值,而不是实际CR值。 Another measure is when a mismatch is found, the shaper and the feedback module to assign default CR values of 42, 44, rather than the actual CR.

如上所述,控制器38也接收从缓存模块40发来的信号,因为某VPI/VCI对的缓存队列的长度达到了门限,而使缓存器充满时,缓存器向控制器发送一个信号。 As described above, the controller 38 also receives a signal from the cache module 40 sent, because the length of the buffer queue of a VPI / VCI pair reached the threshold, leaving when the buffer is full, the cache sends a signal to the controller. 控制器38收到此信号后,命令反馈模块44发送一个资源管理(RM)信元,以后详细说明RM信元。 The controller 38 receives this signal, the feedback module 44 sends a command resource management (RM) cells, described in detail later RM cell. 控制器38也可以接收每一个新建传输(由新的VPI/VCI对标志)的DBC标志值,该DBC标志值是从CAC发出的。 The controller 38 may also receive each new transmission (by a new VPI / VCI to the mark) of DBC flag value, the DBC flag value is sent from the CAC. 当然,如果没有提供标志值,DBC可用一个缺省标志值。 Of course, if no flag value, DBC available for a default flag value.

控制器的伪码如下所示: The controller of the pseudo-code as follows:

<pre listing-type="program-listing"><![CDATA[BEGIN{receive an active VPI/VCI from ACTIVITY DETECTOR} receive active (VPI/VCI) from ACTIVITY DETECTOR send minimum CR to feedback send minimum CR to shaper send bandwidth request (VPI/VCI) to CAC send start. cell.count to ACTIVITY DETECTORENDBEGIN {receive a cell.count from AD} receive a cell.count from ACTIVITY DETECTOR send cell.counr to CACENDBEGIN {receive a VPI/VCI buffer threshold signal} receive a VPI/VCI buffer threshold signal signal feedback to rerransmit CR to end-systemEND BEGIN {receive a CR from CAC} receive a CR from CAC send CR to shaper send CR to feedbackENDBEGIN {receive an inactive signal} receive an inacrive signal advise CAC of inactive VPI/VCIEND]]></pre>控制器的SDL图见图6-1和图6-2。 <Pre listing-type = "program-listing"> <! [CDATA [BEGIN {receive an active VPI / VCI from ACTIVITY DETECTOR} receive active (VPI / VCI) from ACTIVITY DETECTOR send minimum CR to feedback send minimum CR to shaper send bandwidth request (VPI / VCI) to CAC send start. cell.count to ACTIVITY DETECTORENDBEGIN {receive a cell.count from AD} receive a cell.count from ACTIVITY DETECTOR send cell.counr to CACENDBEGIN {receive a VPI / VCI buffer threshold signal } receive a VPI / VCI buffer threshold signal signal feedback to rerransmit CR to end-systemEND BEGIN {receive a CR from CAC} receive a CR from CAC send CR to shaper send CR to feedbackENDBEGIN {receive an inactive signal} receive an inacrive signal advise CAC of inactive VPI / VCIEND]]> </ pre> SDL controller is shown in Figure 6-1 and 6-2.

控制器38向反馈模块44写入DBC,VPI/VCI的标志值对。 The controller 38 is written DBC, VPI / VCI value of the flag to the feedback module 44. 它也命令反馈模块44针对特定的VPI/VCI对发送资源管理命令。 It also commands the feedback module 44 for a specific VPI / VCI transmission resource management commands. 这个命令可能也包括适当的CR值对T,τ,(T是信元平均到达间隔时间,τ是突发容限)。 This command may also include appropriate CR values T, τ, (T is the average cell arrival time, τ is the burst tolerance). 值得注意的是,无论什么时候CAC要更新CR值时,针对每个新VPI/VCI对的RM信元中定义的值中只能有一个值被修改。 It is worth noting that, whenever you want to update CAC CR value, the value of RM cell for each new VPI / VCI pairs defined in only one value is modified. 在公用网中,这种情况通常是每30秒钟或更长时间发生一次,而且决定于DBC20的活动检测器的灵敏度的设定。 In the public network, which is often the case every 30 seconds or longer occurs once, and decided in DBC20 activity detector sensitivity settings. 因而需要的反馈控制带宽能够相对小一些。 Thus the required feedback control bandwidth can be relatively small.

正如将要看到的伪码所示,每当具有指定VPI/VCI的信元的缓存达到充满门限时,控制器就接收到来自缓存模块40的信号。 As shown in pseudo-code will be seen, the cache whenever the cell with the specified VPI / VCI reaches full threshold, the controller receives a signal from the buffer module 40.

与活动检测器36的接口已经说明过了。 And the activity detector 36 interface has already been described.

现在说明反馈模块44的作用。 The feedback module 44 will now be described effects.

正如上面提到的,反馈模块44经输出端口36向终端系统发送当前CR值(由控制器38送来的信号)。 As mentioned above, the feedback module 44 via the output port 36 to send the current CR value (the signal sent by the controller 38) to the terminal system. 通过资源管理信元送出CR,如图7所示。 Sent through the resource management cell CR, as shown in Figure 7. 该信元的一个段是DBC的标志值,以便终端系统14(见图1)能够区别出是哪个DBC(见图2中的DBC20-1和20-2)发来的CR建议。 A segment of the cell is a sign of the value of DBC to the terminal system 14 (see Figure 1) can distinguish which DBC (see Figure 2 DBC20-1 and 20-2) made recommendations to the CR. 在图7中,DBC标志段位于段50,CR位于段52。 In Figure 7, DBC flag segment in the sections 50, CR 52 in the sections. RM信元和其它信元一样,也有5个字节的信头,信头中包含一个PT段54,用来指示该信元是资源管理(RM)信元。 RM cells and other cells, there are also five-byte header, the header contains a PT segment 54, to indicate that the cell is a resource management (RM) cells.

如果使用DBC标志,建议该值不要固定不变,而是在网络为VPI/VCI对建立传输通道时设定的。 If you are using DBC flag, it recommended that the values are not fixed, but VPI / VCI to the establishment of the transmission channel set in the network. 这意味着,CAC18为每个VPI/VCI对指定一个DBC标志,而反馈模块44保存(DBC,VPI/VCI)标志对表。 This means that, CAC18 for each VPI / VCI specified DBC a flag, and the feedback module 44 to save (DBC, VPI / VCI) table flag. 例如,在图2中,公用网10-1为一个VPI/VCI对选定了一个DBC标志并将该信息通知公用网10-2,那么公用网10-2就不会选同样的值作DBC标志(例如,公用网10-1选定了标志1,公用网10-2则选标志2,等等)。 For example, in Figure 2, the public network 10-1 as a VPI / VCI selected a DBC flag and the information notifying the public network 10-2, then 10-2 on the public network will not choose the same value as the DBC flag (e.g., public network 10-1 is selected flag 1, public network 10-2 is selected from the flag 2, and so on). DBC标志存贮在一个由反馈模块44保存的表中。 Flag stored in the DBC by the feedback module 44 to save a table.

RM信元(见图7)的CR段52包括CR建议值和突发容限τ,其中CR建议值是由CAC提供的,作为信元平均到达间隔时间。 RM cell (see Fig. 7) of section 52 include CR CR recommended values and burst tolerance τ, where CR recommended value is provided by the CAC, as the average cell arrival time.

满足下列条件时,控制器38触发反馈模块44。 When the following conditions are met, the controller 38 triggers a feedback module 44. (a)当CAC18提出新的CR建议值时,和(b)与任何一个VPI/VCI对相应的缓存模块40中的缓冲程度超过充满容限时。 (A) When the CAC18 propose new CR recommended values, and (b) with any VPI / VCI to the corresponding level buffer cache module 40 over full of tolerance. 这时向终端系统发送一个资源管理信元。 Then the system sends to the terminal a resource management cell.

反馈模块44的伪码如下,相应的SDL图见图8。 Feedback module 44 of the pseudo-code below, the corresponding SDL is shown in Figure 8.

<pre listing-type="program-listing"><![CDATA[ BEGIN{Receive a CR} receive a CR for a VPI/VCI from Controller default_CR:=CR END{Receive an CR} BEGIN{RM.cell timer expires} RM.cell timer expires create RM.cell write default.CR into RM.cell send RM.cell to end-system restart RM.cell timer END{cell arrival from network}]]></pre>下面说明缓存模块40。 ! <Pre listing-type = "program-listing"> <[CDATA [BEGIN {Receive a CR} receive a CR for a VPI / VCI from Controller default_CR: = CR END {Receive an CR} BEGIN {RM.cell timer expires } RM.cell timer expires create RM.cell write default.CR into RM.cell send RM.cell to end-system restart RM.cell timer END {cell arrival from network}]]> </ pre> The following describes the cache module 40 .

图9是缓存模块的详细图。 Figure 9 is a detailed diagram of the cache module. 它的作用是根据信元中的VPI/VCI对存贮输入的数据信元。 Its role is based on the cell VPI / VCI to store incoming data cells. 缓存信元使得终端系统14(图1)有时间对模块44发出的反馈信号做出反映。 Cache cell so that the terminal system 14 (FIG. 1) time module feedback signal 44 issued to make reflect. 缓存模块40的另一个作用是当缓存达到充满门限时,向控制器38发送一个信号,说明终端系统没有对反馈信号作出反映(这反过来使控制器38向终端系统重发一个信元速率的信号,正如上面提到的)当某个VPI/VCI的最大分配缓存容量溢出时,缓存模块40丢弃收到的信元。 Another effect of the cache module 40 is when the cache reaches full threshold, a signal is sent to the controller 38, indicating that the terminal system does not react to the feedback signal (which in turn causes the controller 38 retransmits a letter to the end-system cell rate signal, as mentioned above) when the maximum cache size allocated to a VPI / VCI overflow, buffer module 40 discards cell. 这是由缓存溢出造成的。 This is caused by a buffer overflow.

对交换系统10进行接入控制的DBC20的缓存容量可以相对小一些。 10 switching system access control DBC20 cache capacity can be relatively small. 例如,如果与DBC20相关的所有信源的和速率是150Mbit/s,那么,当到终端系统的往返时延是100μs时,每当CR值发生变化时,最多有35个信元在传送。 For example, if associated with DBC20 all sources and rate of 150Mbit / s, then, when the round-trip delay of the terminal system is 100μs, whenever the CR value changes, up to 35 cells in transmission.

因为突发时,速率的变化只能导致收到较少数目的多余信元(例如,大约35个信元),所以共享存贮区56的容量主要用于提供突发容限的变化。 Because the burst, the rate of change can result in a fewer number of received redundant cells (e.g., about 35 cells), the capacity of the shared memory area 56 is mainly used to provide a burst tolerance variations. 分配给相应VPI/VCI对的固定信元的位置在图9中的参考序号是58。 Assigned to the position of the fixed cell corresponding VPI / VCI pair in Figure 9. The reference number is 58. 处于这些位置上的信元是若干队列的头信元,每个队列都有各自的VPI/VCI。 In the cell at these locations are the first cell of several queues, each queue has its own VPI / VCI. 换句话说,队列可看作头信元在图9中的右侧在同一起跑线上赛跑。 In other words, the queue can be seen as the right of first cell in Figure 9 race on the same starting line. 到达缓存模块40的信元按照先入先出规则(FIFO)存入队列。 Arrive cache module cell 40 according to the first-in, first-out rule (FIFO) into the queue.

当缓存器收到了从整形器/复用器模块42的整形器部分发来的相应信号时,将信元移出缓存器。 When the buffer is received from the shaper / multiplexer module shaper section 42 of the corresponding signal sent to the cell out of the buffer. 该过程由下列伪码描述。 This process is described by the following pseudo-code.

<pre listing-type="program-listing"><![CDATA[ BEGIN{Receive a cell} receive a cell put cell in buffer IF buffer-fill level=Threshold THEN transmit butter-full signal to CONTROLLER ELSE do nothing END{Receive a cell} BEGIN{Receive a fetch} receive a VPI/VCI ferch signai from the SHAPER/MUX pass cell from butter to the SHAPER/MUX decrement buffer-fill level END{Receive a fetch}]]></pre>相应的SDL示于图10。 <Pre listing-type = "program-listing"> <! [CDATA [BEGIN {Receive a cell} receive a cell put cell in buffer IF buffer-fill level = Threshold THEN transmit butter-full signal to CONTROLLER ELSE do nothing END { Receive a cell} BEGIN {Receive a fetch} receive a VPI / VCI ferch signai from the SHAPER / MUX pass cell from butter to the SHAPER / MUX decrement buffer-fill level END {Receive a fetch}]]> </ pre> corresponding The SDL is shown in Figure 10.

现在同时看图4和图11,整形器/复用器模块42把信元从缓存器40移出,经过网络交换传向目的地。 Now while Figure 4 and Figure 11, the shaper / multiplexer module 42 Cell 40 is removed from the buffer, spread through the network to the destination exchange. 模块42有两部分,它们是复用器60和整形器62。 Module 42 has two parts, which are a multiplexer 60 and a shaper 62. 整形器62为每个VPI/VCI对保存一个信元速率(CR)和定时器。 Shaper 62 for each VPI / VCI to save a cell rate (CR) and timer.

送到输出端口32的信元流由整形器整形,以便那些没有超过突发容限的突发不因整形器62而延时。 Cell to the output port 32 traffic shaping by the shaper so that those who do not exceed the tolerance sudden burst not because shaper 62 and delay. 但是,如果与不同VPI/VCI对应的多个传输同时突发,复用器可能会使信元延迟。 However, if a plurality of different transmission VPI / VCI corresponding to the same time a burst, the multiplexer may cause cell delay. 这种情况下,复用器60为每个活动的VPI/VCI指定一个公平合理的DBC输出带宽。 In this case, the multiplexer 60 to each active VPI / VCI to specify a fair DBC output bandwidth. 这是通过按照不分先后的方式向活动的VPI/VCI探询实现的。 This is in accordance with the alphabetical order of the way through to the activity VPI / VCI polling implementation. 等待时间等于或超过速率间隔T的信元被标志为较高优先级“信元必须送出”,复用器首先发送这些信元(见图11)。 Waiting time is equal to or exceeds the rate interval T cell is flagged as high priority "cell must send" multiplexer first sends these cells (Figure 11). 当突发的出现频率高于突发容限时,整形器将迫使这些信元处于等待状态。 When the frequency is higher than the sudden emergence of unexpected tolerance, the shaper will force these cells in a wait state. 整形器/复用器模块42的详细运行过程可见于下列伪码。 Shaper / multiplexer module 42 during operation can be found in detail in the following pseudo-code.

1. 1.

<pre listing-type="program-listing"><![CDATA[ BEGIN {STATE=ACTIVE} receive a cell.waiting[VPI/VCI]signal from buffer IF burst credit ok THEN cell.can. go:=TRUE STATE:=WAIT for multiplexer ELSE {burst credit not ok} STATE:=WAIT for credir timer to expire END]]></pre>2. <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = ACTIVE} receive a cell.waiting [VPI / VCI] signal from buffer IF burst credit ok THEN cell.can go:!. = TRUE STATE : = WAIT for multiplexer ELSE {burst credit not ok} STATE: = WAIT for credir timer to expire END]]> </ pre> 2.

<pre listing-type="program-listing"><![CDATA[ BEGIN {STATE=WAIT for credit timer to expire} credit timer expires increment burst tolerance credit counter cell.can.go :=TRUE cell.must.go:=TRUE STATE:=WAIT for multrplexer END]]></pre>3. ! <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = WAIT for credit timer to expire} credit timer expires increment burst tolerance credit counter cell.can.go: = TRUE cell.must.go: = TRUE STATE: = WAIT for multrplexer END]]> </ pre> 3.

<pre listing-type="program-listing"><![CDATA[ BEGIN {STATE=WAIT for multiplexer} receive a fetch.cell[VPI/VCI] from multiplexer decrement credit coumter cell.can. go :=FALSE cell.must.go:=FALSE STATE:=ACTIVE END]]></pre>4. ! <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = WAIT for multiplexer} receive a fetch.cell [VPI / VCI] from multiplexer decrement credit coumter cell.can go:. = FALSE cell. must.go: = FALSE STATE: = ACTIVE END]]> </ pre> 4.

<pre listing-type="program-listing"><![CDATA[ BEGIN {STATE=WAIT for multiplexer} credit umer expires IF credit couner<τ THEN increment credit counter ELSE do nothing cell.must.go:=TRUE END]]></pre>5. <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = WAIT for multiplexer} credit umer expires IF credit couner <τ THEN increment credit counter ELSE do nothing cell.must.go:! = TRUE END] ]> </ pre> 5.

<pre listing-type="program-listing"><![CDATA[ BEGIN {STATE=SHAPER ACTIVE} credit timer expires IF credit counter<τTHEN increment credit coumer ELSE do nothing cell.must.go:=TRUE END]]></pre>6. ! <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = SHAPER ACTIVE} credit timer expires IF credit counter <τTHEN increment credit coumer ELSE do nothing cell.must.go: = TRUE END]]> </ pre> 6.

<pre listing-type="program-listing"><![CDATA[ BEGIN {STATE=SHAPER.CR ACTIVE} new CR advised (T,τ) nextT:=T nextcredit:=τ END]]></pre>7. <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = SHAPER.CR ACTIVE} new CR advised (T, τ) nextT: = T nextcredit: = τ END]]!> </ Pre> 7.

<pre listing-type="program-listing"><![CDATA[ BEGIN{STATE=SHAPER TIMER ACTIVE} timer expires reset umer(nexT) END]]></pre>8. <Pre listing-type = "program-listing"> <! [CDATA [BEGIN {STATE = SHAPER TIMER ACTIVE} timer expires reset umer (nexT) END]]> </ pre> 8.

<pre listing-type="program-listing"><![CDATA[ BEGIN{STATE=MULTIPLEXER ACTIVE} output cell timer expires index:=pointer REPEAT{lst loop of searching for cell.must.go} increment index IF cell.must. go[index]THEN pointer:=index fetch cell[index]from buffer send fetch cell signal to SHAPER STATE:=MUX.ACTIVE ELSE IF index=max.buffer.size THEN index:=0 UNTIL index=pointer REPEAT{2nd loop or searching for cell.can.go} increment index IF cell.can. go[index]THEN pointer:=index fetch cell[index]from buffer send fetch cell signal to SHAPER STATE:=MUX.ACTIVE ELSE IF index=max.buffer.size THEN idex:=0 UNTLL index=pointer send no.cell.waitng.signal END]]></pre> ! <Pre listing-type = "program-listing"> <[CDATA [BEGIN {STATE = MULTIPLEXER ACTIVE} output cell timer expires index: = pointer REPEAT {lst loop of searching for cell.must.go} increment index IF cell. must go [index] THEN pointer:. = index fetch cell [index] from buffer send fetch cell signal to SHAPER STATE: = MUX.ACTIVE ELSE IF index = max.buffer.size THEN index: = 0 UNTIL index = pointer REPEAT { 2nd loop or searching for cell.can.go} increment index IF cell.can go [index] THEN pointer:. = index fetch cell [index] from buffer send fetch cell signal to SHAPER STATE: = MUX.ACTIVE ELSE IF index = max.buffer.size THEN idex: = 0 UNTLL index = pointer send no.cell.waitng.signal END]]> </ pre>

当CAC18从控制器38收到带宽请求时,它必须首先确定是否有充足的带宽允许文件在预定最长时间内传完。 When the CAC18 bandwidth requests received from the controller 38, it must first determine whether there is sufficient bandwidth to allow file transfer completed within a predetermined maximum time. 如果没有充足的带宽,将从其他用户处“掠夺”带宽,如下所述。 If you do not have enough bandwidth from other users "predatory" Bandwidth, as described below. 然后,将分配带宽的证实送到控制器38,控制器38再经反馈模块44将带宽分配值送到终端系统。 Then, confirm the allocated bandwidth to the controller 38, the controller 38 and then by the feedback module 44 to the terminal system bandwidth allocation value.

如上所述,CAC18周期性地从控制器38接收信元计数信元,CAC18通过此信号监视有多少信元已经发送(有多少信元要发送)。 As described above, CAC18 periodically received cell count of the cell from the controller 38, CAC18 through the signal track of how many cells have been sent (the number of cells to be sent). 根据信元计数可重新估计保证传输及时完成的带宽值。 Guaranteed bandwidth can be re-estimate the value of the transmission based on the timely completion of the cell count. 即使终端系统得到了足够的带宽,它也可能因为没有完全利用分配带宽而“落后”这时CAC18会分给它更多带宽,使它“赶上”。 Even if the terminal system has been enough bandwidth, it may not fully utilized because the allocation of bandwidth and "backward" Then CAC18'll give it more bandwidth, making it to "catch up."

CAC18也保存与终端系统最大传输速率有关的信息。 CAC18 also save the information and terminal system has a maximum transmission rate dependent. 这样,CAC18就不会给终端系统分配多余的带宽。 Thus, CAC18 not give terminal system allocates extra bandwidth. 一旦CAC18算出了终端系统完成传输必须得到的带宽,这个带宽就被贴上“不许掠夺”的标签,防止带宽被挪用,而使传输不能及时完成。 Once the CAC18 calculated terminal system to complete the transfer must be the bandwidth that was labeled "no predatory" label, to prevent misappropriation of bandwidth, leaving the transmission can not be completed in time.

当DBC20请求修改某传输的带宽时,CAC18必须控制网络中的其它业务流量,以便网络容量能够得到最有效利用。 When DBC20 request to modify a transmission bandwidth, CAC18 must control other network traffic so that network capacity can be most effectively utilized. 下面说明解决流量重新分配问题的连接允许控制方法。 The following describes the flow redistribution to troubleshoot connectivity allows control methods.

现在说明两种连接允许控制方法。 Now show two connection admission control method. 两种方法都解决重新分配流量问题。 Both methods solve redistribute traffic problems. 换句话说,当一个传输静止或者刚活动时,有必要决定会为其它传输产生多少控制消息。 In other words, when a transmission still or newly active, it is necessary to decide how much control messages will be generated as other transport. 目的是使这种控制消息数尽可能少。 The purpose is to make this as little as possible to control the number of messages.

第一种策略是一个简单的连接允许控制方法,实际上并没有重新分配流量。 The first strategy is a simple connection admission control method does not actually reallocate traffic. 在这种方法中,一个新活动的传输(VPI/VCI对)只分配给一个勉强能完成任务的信元速率,直到该传输再次静止为止。 In this method, a new active transmission (VPI / VCI pair) assigned to only one barely able to complete the task of cell rate until the transmission still in operation again. 只有当传输后来再被激活时,才给它一个不同的CR。 Only later on, when the transmission is activated, which gave it a different CR. 这意味着与一个VPI/VCI对相应的静止信号不会生成与其共享网络容量的其它VPI/VCI对的控制信号。 This means that with a VPI / VCI does not generate static signal corresponding to share network capacity with other VPI / VCI control signal pairs.

这种方法与一种填充方法联合应用,填充方法包括(i)给第一个刚活动的连接所有可用容量的一半作为其有效容量;(ii)给下一个刚活动的连接剩余容量的一半作为其有效容量;(iii)给再下一个刚活动的连接剩余容量的一半作为其有效容量;等等。 This method is a combination of filling method, padding method comprises (i) for the first half of a newly active connections all available capacity as its effective capacity; (ii) the remaining capacity connected to half the next newly active as (iii) to a newly active then the next half of the remaining capacity of the connection an effective capacity;; its effective capacity and so on. 这种方法是在VPI/VCI标志的路由上逐条链路应用的,无论哪段产生的最低容量都是反馈给DBC20的CR的决定因素。 This method is applied in the by-link routing VPI / VCI mark, the minimum capacity to produce regardless of which segment is back to DBC20 of CR determinants.

可见,一个新活动的有一个VPI/VCI对的信号不会生成与其共享容量的其它VPI/VCI对的控制信号。 Visible, a new activity has a VPI / VCI pair of signal does not generate control signals to share the capacity of other VPI / VCI pair.

DBC20的设计使得只要活动检测器36(见图4)认为VPI/VCI对处于活动状态,用户就能在网络中拥有较大的有效容量,而且用户产生信元的速率接近有效带宽值(参见上面说过的活动检测器所具有的信元计数功能)。 DBC20 is designed so that as long as the activity detector 36 (see Figure 4) that the VPI / VCI pair is active, the user can have a large effective capacity in the network, and user-generated cells close to the rate effective bandwidth value (see above It said activity detector has cell counting function). 因此,只要用户原意付出与他们申请的带宽相应的较大的负载的费用,就能拥有较大的有效带宽。 Therefore, as long as users pay with their intent to apply for larger bandwidth corresponding load fee, you can have a larger effective bandwidth.

这种方法对用户来说是公平的,因为在足够长的一段时间内,没有一个用户是故意被分到较差的容量的。 This method is fair to users, since a sufficiently long period of time, no one was assigned to the user intentionally poor capacity.

但是,在有些情况下,增加能得到大带宽值的用户数是合乎需要的,下面说明的第二种修改的方法可以提供这种功能。 However, in some cases, increase the number of users able to obtain a large bandwidth value is desirable, a second modification of the method described below can provide this functionality.

这种情况下,基本的原则是:如果一个活动的信号导致其它VPI/VCI值对的控制信号的产生,那么让这种控制信号限制在每条链路上只有一个,即具有最大容量的VPI/VCI对,这可以叫做有限重分配方法或者“劫富济贫”(罗宾汉)方法。 In this case, the basic principle is: if an activity signal results in a control signal other VPI / VCI value pairs, so let this control signal constraint on each link VPI only one that has a maximum capacity of / VCI pair, which can be called a limited re-allocation method or "Robin Hood" (Robin Hood) method.

从填充方法的例子可以得到最好的说明:(i)第一个刚活动的VPI/VCI对被分到总可用容量的一半作为其有效容量。 Examples from filling method can be best illustrated by: (i) the first newly active VPI / VCI and half were assigned to the total available capacity of an effective capacity.

(ii)下一个刚活动的连接被分到剩余容量的一半再加上第一个VPI/VCI对的有效容量的五分之一(当前最富裕的容量的1/5)。 (Ii) the next newly active connection is assigned half of the remainder of the first fifth (the most affluent fifth of capacity) coupled with the effective capacity of the VPI / VCI pair.

(iii)再下一个刚活动的连接被分到剩余容量的一半再加上当前最富裕的容量的五分之一;等等。 (Iii) The next newly active connection is assigned half of the remainder plus the richest fifth of the current capacity; and so on.

为了说明这个过程,可以想象有一条100Mbit/s的链路,根据上面的步骤可得出下面的步骤:(i)第一个刚活动的VPI/VCI对得到50Mbit/s,还剩50Mbit/s,(ii)下一个刚活动的VPI/VCI对得到剩余容量的一半(结果是25Mbit/s)再加上第一个活动VPI/VCI容量的1/5,这说明第一个VPI/VCI对现在有40Mbit/s,第二个有35Mbit/s;(iii)再下一个活动VPI/VCI对得到剩余容量的一半,即12.5Mbit/s再加上第一个VPI/VCI对的1/5,所以第一个VPI/VCI对现在有32Mbit/s,第二个仍然有35Mbit/s,第三个有20.5Mbit/s,等等。 To illustrate this process, imagine there is a 100Mbit / s link, according to the above steps can be drawn from the following steps: (i) the first newly active VPI / VCI to get 50Mbit / s, left 50Mbit / s , (ii) the next newly active VPI / VCI to receive half of the remaining capacity (the result is 25Mbit / s) plus the first one active VPI / VCI capacity of 1/5, indicating that the first VPI / VCI pair There 40Mbit / s, the second has 35Mbit / s; (iii) The next active VPI / VCI pair gets half of the remaining capacity, i.e. 12.5Mbit / s plus the first VPI / VCI of 1/5 , so the first VPI / VCI pair is now 32Mbit / s, the second still 35Mbit / s, the third has 20.5Mbit / s, and the like.

注意到大多数用户现在都得到了大容量,但是链路只发送了一条额外的控制消息。 Noting that most users now get a large capacity, but the link to send only one extra control message. 这就是有限重分配,即“劫富济贫”策略。 This is a limited redistribution, namely "Robin Hood" policy.

把这种方法扩展到具有多条链路的路由上,上面的过程就逐条链路进行,哪条链路得到了最小有效容量,它就是发给DBC20的CR值的决定因素。 To extend this approach to a route with multiple links, the above process is carried out by one link, which links get the minimum effective capacity, it is the determinant of the CR value of DBC20 the issue. CAC逐条链路指定该最小有效容量,即那条链路剩余容量的一半再加上那条链路最富有的VPI/VCI对的容量的1/5,而该最富有VPI/VCI对的容量减少相应值。 CAC by-link specify the minimum effective size for the remaining capacity of the piece of half plus links that link the richest VPI / VCI to the capacity of 1/5, while the richest VPI / VCI to capacity decrease the corresponding values. 结果,每条链路上所有活动的VPI/VCI信号至多产生一条附加CR控制消息发给网络。 As a result, each of the VPI / VCI signal on link up all the activities that an additional CR control message to the network. 静止的信号不生成附加控制消息。 Stationary signals do not generate additional control messages.

这种策略使得一个用户在其它用户是活动的时候,不能保持有很大容量。 This strategy enables a user when the other user is active, a large capacity can not be maintained. 而且,在保持重分流量的复杂度为最小的情况下,尽管使更多的用户得到合理的大容量。 Moreover, while maintaining minimal complexity redistribution of traffic, although so that more users get reasonably large capacity.

总之,提供一个交换异步传递信元的宽带交换系统,动态带宽控制器(DBC)控制信元送入该系统的输入端口,这些信元是由一系列传输终端系统提供的。 In short, there is provided a switching asynchronous transfer cells broadband switching system, dynamic bandwidth controller (DBC) control cell into the input port of the system, these cells are provided by a series of transmission terminal system.

当终端系统开始传输数据信元时,DBC检测输入信元存在与否,并向系统的CAC申请带宽。 When the terminal system begins transmitting data cells when, DBC detects the presence or absence of input cells, and to apply the system CAC bandwidth.

交换系统有一个表,该表保存与输入端口相连的若干信源及其最长传输时间,最大传输速率。 Switching system has a table, the table holds several sources connected with the input port and the maximum transmission time, the maximum transmission rate. 检测到一个信源向输入端口发送信元时,DBC和CAC合作在交换系统中分配一个足够的带宽,使得相应长度的文件可以在相应的最长传输时间内传完。 Detection of a source to the input port sends cells, DBC and CAC collaborate allocate sufficient bandwidth in the exchange system, so that the corresponding length of the file can be transferred completely within the respective maximum transmission time. 信源也确保有一个最小传输速率。 Source also ensure that there is a minimum rate.

Classifications
International ClassificationH04L12/54, H04L12/70, H04Q3/00, H04Q11/04
Cooperative ClassificationH04L2012/5615, H04L2012/5639, H04L2012/5632, H04L2012/5635, H04L12/5602, H04L2012/5636, H04L2012/5649, H04Q11/0478
European ClassificationH04L12/56A1, H04Q11/04S2
Legal Events
DateCodeEventDescription
27 May 1998C06Publication
3 Jun 1998C10Request of examination as to substance
3 Nov 1999C53Change or modification
3 Nov 1999CORBibliographic change or correction in the description
Free format text: CORRECT: APPLICANT; FROM: OXFORD BROOKS UNIVERSITY TO: BRITISH TELECOMMUNICATIONS PUBLIC LIMITED COMPANY
17 Sep 2003C12Rejection of an application for a patent