WO2005109665A1 - Procede de realisation d'une inversion a plaque simple dans un systeme de communication - Google Patents

Procede de realisation d'une inversion a plaque simple dans un systeme de communication Download PDF

Info

Publication number
WO2005109665A1
WO2005109665A1 PCT/CN2005/000637 CN2005000637W WO2005109665A1 WO 2005109665 A1 WO2005109665 A1 WO 2005109665A1 CN 2005000637 W CN2005000637 W CN 2005000637W WO 2005109665 A1 WO2005109665 A1 WO 2005109665A1
Authority
WO
WIPO (PCT)
Prior art keywords
board
switching
boards
original
control
Prior art date
Application number
PCT/CN2005/000637
Other languages
English (en)
French (fr)
Inventor
Xiaolong Mi
Donglei Luo
Jidong Zhang
Original Assignee
Huawei Technologies Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Publication of WO2005109665A1 publication Critical patent/WO2005109665A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/10Packet switching elements characterised by the switching fabric construction
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/15Interconnection of switching modules
    • H04L49/1515Non-blocking multistage, e.g. Clos
    • H04L49/1523Parallel switch fabric planes

Definitions

  • the present invention relates to a communication method, and in particular, to a method for switching a board in a communication system based on a shared memory communication method.
  • Telecom networks require reliable and uninterrupted services for their users, especially in some important business applications, such as electronic money, order processing, customer service, inventory management, email and Internet access, etc., business viability Has become more important than ever, and its availability requirements can reach 99.999% or higher. Therefore, network survivability has become an important factor affecting network design and construction, and the equipment used in telecommunication networks also needs to have high reliability accordingly.
  • Redundant backup and active / standby switchover are common solutions. Redundant backup and active / standby switching technologies have multiple backup modes, such as 1 + 1, 1: 1, 1: N, etc.
  • 1 + 1 backup is a single protection method.
  • the main and backup units in the system form a logical business function unit, that is, the protected equipment unit has a backup unit.
  • the main module is responsible for real-time processing of the service and the backup service is maintained.
  • Consistent with the data of the active unit a switchover is initiated when a failure of the original active unit is detected to take over the services on the original active unit.
  • the difference from a 1 + 1 backup is that the standby unit in a 1: 1 backup only works after a failure.
  • 1: N backup means that multiple equipment units share a backup unit. When any one of the active units fails, the backup unit takes over its work.
  • a complex telecommunication device includes multiple processing units working together, and each unit needs to communicate with each other.
  • different node devices also need to communicate.
  • the switching In order to ensure the equipment's self-healing ability, ensure data integrity and maintain service quality under network failure conditions, the switching must be completed within a specified time limit, and services should be restored as soon as possible to minimize the impact on the communication system. It can be seen that the performance of the protection switching mechanism has a significant impact on the reliability of the system and even the survivability of the entire network.
  • the switching methods currently used in communication systems are mainly switching methods based on circuit switching and communication based on Internet Protocol (Internet Protocol, "IP" for short) communications.
  • Internet Protocol Internet Protocol
  • FIG. 1 (a) shows a switching device in a conventional circuit switching system.
  • the circuit switch is implemented through a "switching network".
  • the switching network unit includes two units of the active and standby switching planes. The effect of redundant backups.
  • Each business processing unit maintains a physical connection with the active and standby switching units, and can send and receive data from both switching planes. However, during normal operation, only data processing on the active plane is selected.
  • the switching network unit When the switching network unit is switched, it will notify all instructing each service module to switch to the new switching plane. In order to ensure that the switching operation is completed within the prescribed time limit, the switching of the switching network unit must be implemented by hardware.
  • each service processing unit also includes a master and backup service processing unit, which is used to protect respective service communications.
  • the switching network adopts a "single transmit and double receive" operating mode, that is, business data is simultaneously exchanged to the input / output ports of the active and standby business processing units.
  • the active and standby service modules determine the operation mode of port sending and receiving through arbitration or heartbeat handshake. During normal operation, only the active unit processes services and sends data, and the output ports of the standby module are in a disconnected or high-impedance state; when a switchover occurs, the active and standby units decide to disconnect or operate the connected input / output ports based on the arbitration result.
  • FIG. 1 (b) shows a switching device in a conventional IP network.
  • Each device or network node in the IP network is assigned an independent IP address, and the switching operation is realized by switching the IP address.
  • Circuit-based switching technology must be implemented in hardware, with poor versatility and scalability, and high cost; IP-based switching technology cannot be applied to real-time, high-quality telecommunications services.
  • the main reason for this is that the switching device implemented by hardware must be designed for the specific system circuit, which is not easy to upgrade and test, and the cost is high; the switching technology of IP address switching is directly limited by IP technology, and IP technology has transmission time Delays are uncertain, transmission bandwidth is not guaranteed, and transmission quality is not high.
  • the main object of the present invention is to provide a method for switching a single board in a communication system, so that the switching operation of the service channel before and after the switching is simple, reliable, fast and efficient, and meets the switching requirements of carrier-grade equipment.
  • the present invention provides a switching method in a communication system.
  • the system includes multiple boards, at least one of which is a control board, and is used to centrally control all boards through a bus.
  • Each board is also It contains a logical slot and physical slot mapping table, which is used to indicate the mapping relationship between the logical slot and the physical slot of each board, and is characterized in that the board switching includes the following steps:
  • the control board modifies the logical slot and physical slot mapping table of all the boards, and maps the logical slot number of the original active board to the physical slot number of the original standby board;
  • the control board modifies its own logical slot number stored on the original active and standby boards.
  • the control board modifies a logical slot and physical slot mapping table on the board
  • the control board synchronizes the contents of the updated logical slot and physical slot mapping table of this board to the logical slot and physical slot mapping tables of all other boards.
  • the method further includes: mapping the delayed slot number of the original standby board to the physical slot number of the original active board.
  • the communication system is based on a compact peripheral device interconnect bus platform.
  • the present invention also provides another method for switching boards in a communication system.
  • the system includes at least two frames that communicate with each other.
  • Each frame includes multiple boards, and at least one control board is used to centralize the bus.
  • Control all the boards in the box; the control board also includes a channel information table for storing the mapping relationship between the logical slot numbers of all the boards in the box and the base address of the data storage area for sending and receiving, in the control board
  • the board needs to be switched it includes the following steps:
  • the original master control board is prohibited from allocating a receiving channel to the device board.
  • the second method also includes:
  • the control board modifies its own logical slot number stored on the original active board and the original standby board.
  • the communication system is based on a compact peripheral device interconnect bus platform.
  • the frames can communicate through an asynchronous transmission mode.
  • a control board is used to centrally control the switching operation of the board on the bus, and the board in the frame passes the migration slot Bit and physical slot mapping mechanism to achieve transparent switching. Only the logical slot and physical slot mapping tables need to be updated during the switching. In the case of inter-frame communication, the sending and receiving channel mapping information and control command word addresses need to be switched. There are two operation modes: control board switching and device board switching.
  • the shared memory communication method provides convenient bus operations
  • the logical slot and physical slot mapping mechanism provides a simple switching operation method, and the concentration of the bus
  • the control guarantees the reliability and real-time of the switching.
  • the channel mapping mechanism realizes the switching in the case of inter-frame communication.
  • the switching method provides a seamless protection mechanism to ensure the real-time and reliability of the switching and reduce the network cost. Improved system scalability and network survivability.
  • FIG. 1 is a schematic diagram of an existing switching technology
  • FIG. 2 is a schematic diagram of a structure of a PSM frame, a shared memory communication mechanism, and an inter-frame communication mechanism;
  • FIG. 3 is a diagram of modifying a logical slot and a physical slot according to an embodiment of the present invention A flowchart of a switching method implemented by a bit map;
  • FIG. 4 is a flowchart of a switching method implemented by modifying a channel information table under inter-frame communication according to an embodiment of the present invention.
  • the invention is based on the shared memory communication technology, and completes the switching operation of the equipment unit through the modification of the address mapping, which can meet the requirements of real-time and transparency.
  • mapping relationship between the physical position and the logical position of each board on the bus. The modification of the mapping relationship is used to realize the transparent switching of the device unit, which has good scalability and real-time performance.
  • a PSM frame based on a Compact Peripheral Component Interconnect (CP-CI) bus platform is used to implement the switching of each board on the bus, that is, the switching of the device unit.
  • CP-CI Compact Peripheral Component Interconnect
  • CPCI is a high-performance industrial bus based on the Peripheral Component Interconnect (PCI) standard. At the same time, it inherits the mechanical characteristics of European-style boards to achieve a reliable structure.
  • the PSM chassis has 16 slots, which can be inserted into the front and rear. There are four types of boards processed, which are called URCU, MCU, UGPU, UFIU, and universal PCI daughter board ("PCI Mezzanine Card", referred to as "PMC").
  • PCI Mezzanine Card PCI Mezzanine Card
  • the URCU board is the control board in the PSM box, which is responsible for the management function of the PSM box, and 1 + 1 backup is used;
  • the UACU board is the auxiliary control board in the PSM box, which assists the URCU board to complete the control function, mainly used for Realize the interconnection between different segments of the bus in the PSM box and use it in pairs with the URCU board.
  • the UGPU board is a business processing unit in the PSM box and supports various business types. A maximum of 12 UGPU boards can be configured in a PSM box.
  • the system can be configured as 1 + 1 backup or 1 + N backup according to the type of business.
  • PMC is a small daughter board in the CPCI system.
  • the UFIU board is used to implement inter-frame connections, and is specifically used in asynchronous transmission mode (Asynchronous Transfer Mode (referred to as "ATM") is an optical interface board for an optical network. It is a PMC daughter board that uses a PMC interface and is buckled on a URCU board.
  • ATM Asynchronous Transfer Mode
  • Figure 2 (a) shows the structure, components, and communication connection modes of the PSM frames.
  • the communication between the boards is mainly based on the PCI bus technology.
  • the communication between the URCU board and the UGPU boards is based on the CPCI backplane bus.
  • the communication between the URCU board or the UGPU board and the PMC daughter board is based on the inside of the carrier board.
  • PCI bus Each single board uses an IC (Special Integrated Circuits) (ASIC) chip such as a PCI bridge to implement electrical connection and address conversion functions. That is, as long as the bridges are properly configured, the CPCI platform All of the single boards can access each other's memory space.
  • ASIC Specific Integrated Circuits
  • the PCI bus can be cascaded through a PCI bridge.
  • the backplane of the 16-slot PSM frame is actually composed of It consists of two CPCI buses.
  • the two segments of the CPCI bus are bridged through the UACU board to achieve cross-bus access and communication between the buses.
  • only one PCI master device can complete the bus arbitration at the same time. Therefore, in the active and standby working mode, the URCU board has only one active single board for controlling two sections of the CPCI bus. When the bus is disconnected and the URCU board is switched, the standby URCU board preempts the bus control right, the master URCU board releases the bus control right, and the corresponding UACU board also switches with the URCU.
  • the shared memory communication of the PSM box is mainly completed by the UFIU bottom driver, the CPCI bottom driver, and the PMC daughter board bottom driver module.
  • the PSM boxes are implemented through the ATM server + client mode, and are carried by the ATM optical transmission network.
  • the optical interface of UFIU is connected to the ATM optical transmission network, and the UFIU bottom-level driver module completes the communication between the PSM frames.
  • the CPCI bottom driver module completes the communication between the URCU board and each UGPU board in the PSM frame.
  • the communication module at the bottom of each PMC daughter board completes the communication between its carrier board and the PMC daughter board. It can be seen that when the switching of the URCU or UGPU board occurs, the communication modules related to the switching only involve the UFIU low-level driver module responsible for inter-frame communication or the CPCI low-level driver module responsible for intra-frame communication.
  • Figure 2 (b) shows the memory sharing mechanism between the individual boards on the CPCI bus.
  • the working principle of the CPCI bottom communication module is the memory sharing between the boards.
  • Each board reserves a shared physical memory for other boards, which is mapped onto the CPCI bus through the PCI bridge chip, and then through PCI The bridge is mapped to the PCI storage space of other single boards. When other boards access this PCI storage space, they actually access the physical memory of the source board. It can be seen that the mapping of the PCI bridge chip realizes the memory sharing between different boards.
  • FIG. 2 (c) shows the mechanism of implementing UFIU inter-frame communication using UFIU.
  • the design idea of the UFIU low-level communication module is similar to that of CPCI. It is also based on shared memory and CPCI address space mapping. The difference is that the UFIU communication module is implemented based on the ATM server + client model.
  • UFIU On the connection mechanism, UFIU is connected to the ATM optical network through the optical interface, and realizes communication with other PSM boxes or nodes under the mode exchange of ATM server + client. Inside the PSM box, the UFIU communicates with each UGPU board to realize the function of sending and receiving data of each UGPU board by sharing the memory.
  • Each UGPU board reserves a certain amount of shared memory for the UFIU.
  • the UFIU can be accessed by other CPCI boards, including the UFIU daughter board on the URCU board.
  • Dedicated processing chips such as CN8236 ASIC chips, directly access and store in the corresponding UFIU receiving shared memory area of the receiving UGPU board; when the data of each UGPU board is sent out of the PSM box, you only need to copy it to the UFIU sending share
  • the memory area is automatically taken out by the SAR special processing chip on the UFIU board and forwarded to the ATM network through the optical interface.
  • the ATM switch finds the path through the virtual path (Virtual Path, referred to as "VP") indicator VPI and the virtual channel (Virtual Channel, referred to as "VC") indicator VCI in the letterhead.
  • VP Virtual Path
  • VC Virtual Channel
  • one VP carries multiple VCs, so a VP can be regarded as multiple VC sets with the same VPI value.
  • VPs have different VPI values, VCs belonging to the same VP have different VCIs, and VCs belonging to different VPs can have the same VCI value.
  • VC can be used to directly complete the connection between two users, and VP can be used to connect between different network segments.
  • VP switching only switches the VPI, and the cell's VCI remains unchanged during the entire VP link transmission process.
  • the VC is connected at the switching node, and the mapping table is found according to the VPI and VCI of the cells in the virtual path to obtain the corresponding new VPI and VCI.
  • PVP Perpetual Virtual Path
  • PVC Perpetual Virtual Channel
  • VCI pathfinding function When communicating between PSM boxes, different VPIs are generally used to indicate different PSM boxes, and different boards in the box are indicated by VCI. Therefore, on the UFIU board, the VCI pathfinding function must be completed, that is, the VC channel of each board is mapped. This is achieved by mapping the sending VCC table (SEG VCC) and receiving the VCC table (RSM VCC).
  • SEG VCC sending VCC table
  • RSM VCC receiving the VCC table
  • the send and receive queues, SEG VCC tables and RSM VCC tables, and send buffer descriptions (Send Buffer Descr iptor, tube ("SBD") table and so on.
  • the operation management and maintenance (Operat ion, Ad ministry, Maintenance, or "0AM") of the URCU are mainly allocated to receive and send data buffers and receive and send status queues for cells and data cells.
  • the sending data buffer of each UGPU board in the same frame.
  • local send and receive buffer descriptors, receive and send status queues, local SEG VCC tables and RSM VCC tables, and receive data buffers are allocated.
  • UFIU's SAR dedicated processing chip mainly relies on VCC tables to send and receive ATM cells with the outside world.
  • the sending VCC table is located in the shared memory of the SAR.
  • Each board can establish up to 256 sending PVCs.
  • the index value of the entry is the sending VCC index.
  • each board can establish up to 256 receiving PVCs.
  • the index value of the entry is the receiving VCC index.
  • One VCC index corresponds to one VPI and VCI link.
  • the PVC configuration of UFIU is to establish a mapping relationship between the VCC index and PVC for the inter-frame transmission and reception of boards, that is, the mapping relationship between VPI and VCI links and each board.
  • the PVC of UFIU boards uses unidirectional PVC. Therefore, the establishment and removal of the sending PVC and the receiving PVC are independent, and the index numbers of the two are also assigned separately.
  • the PSM frame needs to be addressed by the correspondence between the VCC index in the VCC table and the PVC, that is, the sending PVC indexes the VCC table sent by the host according to the physical slot number; the receiving PVC is The logical slot number is used to index the VCC table received by the host.
  • the external ATM exchange will not change, that is, the corresponding VCC index of the VPI and VCI links cannot be switched. Therefore, the actual switching is the corresponding relationship of the VCC index PVC channel, that is, the VCC item and the receiving physical slot. The corresponding relationship of the numbers to change the flow direction of the received data unit.
  • a mapping conversion relationship between a logical slot and a physical slot is used.
  • a switching method based on CPCI shared memory communication was implemented.
  • the system assigns a logical slot to each board on the bus. High-level applications access the corresponding board through the logical slot.
  • a logical slot and a physical slot are configured in the CPCI bottom driver module on each board.
  • Bit mapping table which reflects the corresponding logical slot allocated to each board at the physical location.
  • the CPCI underlying driver module indexes the board at the physical location according to the table.
  • the logical slot is the same as the actual physical slot of the board.
  • the high-level application module transmits the data unit and the logical slot of the target board to the CPCI bottom driver module.
  • the CPCI bottom driver module is You can search the logical slot and physical slot mapping table, convert it to the actual physical slot of the destination board, and then use the shared address space mapping of the PCI bridge to convert it to the corresponding CPCI space address of the shared memory for access.
  • the logical slot and physical slot mapping table mechanism is directly related to the bottom layer of the PSM box, and can be done through a direct index lookup, so it has no effect on the efficiency of CPCI communication.
  • mapping table of logical slots and physical slots on all boards you only need to modify the mapping table of logical slots and physical slots on all boards to achieve the equivalent board switching. For example, if the logical slot number of the original main board to be switched is 1 and its physical slot number is 1 and the logical slot number of its standby board is 2 and its physical slot number is mapped to 2, then there are two single boards.
  • a board switch occurs, change the mapping table to logical slot number 1 to physical slot number 2, and logical slot number 2 to physical slot number 1, which means that the switchover operation of the active and standby boards is equivalently completed.
  • the switching operation of other standby modes can also be conveniently implemented. It can be seen that the switching method is not only convenient to operate, but also can be completed at a high speed. Before and after the switching operation, the service communication is hardly affected, and the normal communication of the service is guaranteed.
  • the centralized control method is used to implement the switching operation, so that the switching of the equipment unit is more simple and reliable.
  • the CPCI control board URCU board centrally controls the logical and physical slot number mapping tables of all the boards. Each board also places the logical slot and physical slot mapping tables in the shared memory of the board. The centralized modification of the URCU board ensures the normal operation of the entire communication system.
  • each board also maintains its own logical slot and physical slot information.
  • modifying the logical and physical slot mapping relationship table it is necessary to modify the related switching card.
  • Information about its own logical slots stored in the board For example, when the board in physical slot 1 and the board in physical slot 1 are switched as described above, except for all mapping tables, To be modified, the logical slot number stored on the physical slot board 1 needs to be changed from the original number 1 to 2, and the logical slot number stored on the physical slot board 2 needs to be changed from the original Number 2 was changed to number 1.
  • FIG. 3 illustrates a specific flow of a switching method implemented by modifying a logical slot and a physical slot mapping table according to an embodiment of the present invention. This method is applicable to the switching of various boards on the bus, including control boards and equipment boards.
  • step 301 the control board updates the logical slot and physical slot mapping table on the board according to the board switching mode. For example, in the PSM box, URCU modifies the mapping table stored by itself.
  • step 302 the control board synchronizes the updated logical slot and physical slot mapping table contents to the logical slot and physical slot mapping tables of all the boards.
  • URCU modifies the mapping table on each UGPU board.
  • Other boards send data through logical slot numbers.
  • the control board can send the data to the switched active board according to the correspondence in the mapping table.
  • step 303 the control board updates its own logical slot information stored on each board involved in the switching.
  • the "self” mentioned here refers to the "each board", not the control board.
  • an external interface board of each frame stores a channel information mapping table used for data unit forwarding of a board.
  • the channel information mapping table here is used by the external interface board to index the information of the physical slot number of the board in the box according to the channel number for external communication.
  • the aforementioned VCC table and VCC table are received. When switching occurs, this information needs to be modified to ensure the normal operation of inter-chassis communication.
  • the UFIU interface board based on the ATM optical interface stores information such as a reception VCC table and a transmission VCC table, and according to these information, the data units of each board are indexed and forwarded to complete inter-frame communication.
  • the information update of the main and backup boards when the control board is switched is mainly the information related to the control board itself, such as the modification of the transmission and reception channel mapping tables of the main and backup control boards and the channel mapping table of communication between the main and backup boards.
  • the channel mapping table related to the UGPU board is synchronized with the active and standby control boards during normal (configuration / deletion / UGPU board switching) operations, so these information do not need to be updated when the control board is switched.
  • the underlying drive module communication is also implemented by means of shared memory. Specifically, it is This is achieved by writing a control command word to the corresponding control command word address, so the corresponding control command word address also needs to be modified when switching occurs. It can be seen that the other switching operations required under the framework communication only involve the underlying driver module of the interface board, which is reliable and has good real-time performance.
  • FIG. 4 (a) shows a method for controlling a single board switching implemented by modifying a channel information table under inter-frame communication according to an embodiment of the present invention.
  • step 411 the bus control switching is performed, the original master control board releases the bus control right, and the original standby control board takes over the bus control right.
  • the original standby URCU is in the bus disconnected state.
  • the original master URCU transfers bus control to the original standby URCU through a handshake.
  • step 412 to update the channel information table corresponding to the control board on the active and standby control boards.
  • the sending VCC table and the receiving VCC table contain the base address of the data storage area of each board.
  • the so-called base address refers to the basic address. You can find the starting position of the data storage area based on this base value, and then according to the data sent and received
  • the existing data in the storage area is calculated to obtain the location where the data that needs to be sent and received should be stored.
  • step 41 3 to prohibit the original active control board from allocating the receiving channel of each device board.
  • the original master uses the URCU board to control the UGPU board, so it cannot be assigned to the receiving channel of each device board.
  • FIG. 4 (b) shows a method for switching a device board (a board other than a control board) implemented by modifying the receiving and sending channel information under inter-frame communication according to an embodiment of the present invention.
  • step 421 the corresponding channel information table of the device board involved in the switchover stored on the main and standby control boards is updated.
  • the information such as the sending VCC table and receiving VCC table of the UGPU involved in the switching stored on the active and standby URCUs.
  • all UGPU sending and receiving VCC tables and other information are stored on the URCU, while each UGPU also stores its own sending and receiving VCC tables and other information.
  • step 422 the control command word address information of the device board involved in the switching is updated, and the control command word address is used to implement shared memory communication between the underlying driver modules.
  • the control command word address information of the device board involved in the switching is updated, and the control command word address is used to implement shared memory communication between the underlying driver modules.
  • the underlying ATM driver modules of the UGPU board and the URCU board need to write control command words to the control command word address to achieve communication. Therefore, the corresponding control command word address needs to be updated during switching.
  • step 423 the control board modifies the sending channel information and receiving channel information of the active and standby equipment boards involved in the switching.
  • the sending channel information and receiving channel information here are similar to the role of the logical slot number of each board.
  • the switching method provided by the present invention can be used to implement device switching without affecting the essence of the present invention. And scope.
  • the cascading scheme between PSM frames is mainly implemented by using a UFIU daughter board and an ATM optical interface.
  • UFIU daughter board and an ATM optical interface.
  • other physical interfaces are used to implement different PSM frames or different CPCI buses.
  • the switching method provided by the present invention can still complete the device switching without affecting the essence and scope of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Description

通信系统中单板的倒换方法
技术领域 本发明涉及一种通信方法, 特别涉及基于共享内存通信方式的通信系 统中单板的倒换方法。
背景技术
电信网络要求能为其用户提供可靠的不间断的服务, 尤其是在一些重 要业务的应用中, 如电子货币、 订单处理、 客户服务、 库存管理、 电子邮 件和国际互联网接入等, 业务生存性变得比以往更加重要, 其可用性要求 能达到 99. 999%甚至更高。 因此, 网络生存能力成为影响网络设计与构建 的重要因素, 而在电信网络中使用的设备也相应的需要有很高的可靠性。
冗余备份和主备倒换技术是一种常用的解决方法。 冗余备份和主备倒 换技术有多种备份模式, 比如 1+1 , 1: 1 , 1: N等。 1 + 1备份是一种筒单 的保护方式, 系统中的主备用单元组成一个逻辑上的业务功能单元, 即被 保护设备单元拥有一个备用单元, 主用模块负责业务的实时处理, 备用业 务保持与主用单元的数据一致, 在检测到原主用单元故障时发起倒换, 接 管原主用单元上的业务。 与 1 + 1备份不同的是 1: 1备份中备用单元只在 出现故障后开始工作。 而 1: N备份就是多个设备单元共享一个备用单元, 当任意一个主用单元出现故障时, 备用单元即接替其工作。
通常, 一个复杂的电信设备包含多个处理单元协同工作, 各单元之间 需要进行通信; 而在整个通信网络中, 不同节点设备之间也需要通信。 为 了保证在网络故障条件下设备自愈能力、 确保数据的完整性和维持服务质 量, 倒换必须在指定的时限内完成, 尽快恢复业务, 使对通信系统影响最 小。 可见, 保护倒换机制的性能对于系统的可靠性乃至整个网络的生存能 力都有着重大的影响。
目前通信系统中采用的倒换方法主要有基于电路交换和基于网际协 议 ( Internet Protocol , 简称 "IP" )通信等倒换方法。
图 1 ( a )示出了传统的电路交换系统中的倒换装置。 电路交换机是通 过 "交换网络" 实现的, 交换网络单元包含主备交换平面两个单元, 达到 冗余备份的效果。 各业务处理单元都与主备交换单元都保持物理连接, 并 且从两个交换平面上都可以收发数据, 但正常工作时只选用主用平面上的 数据处理。 当交换网络单元发生倒换时, 会通知所有指示各业务模块切换 到新的交换平面上。 为了保证倒换操作在规定的时限内完成, 交换网络单 元的倒换必须通过硬件方式实现。 另外, 每个业务处理单元也包含主备业 务处理单元, 用于保护各自的业务通信。 对于主备业务处理单元, 交换网 络采用 "一发双收" 的工作模式, 即将业务数据同时交换到主备业务处理 单元的输入 /输出端口上。 主备业务模块之间通过仲裁或心跳握手方式确 定端口收发的操作方式。 正常工作时, 只有主用单元会处理业务并发送数 据, 备用模块的输出端口处于断开连接或高阻状态; 发生倒换时, 主备单 元根据仲裁结果决定断开或操作连接输入 /输出端口。
图 1 ( b )示出了传统的 IP网络中的倒换装置。 IP网络中每个设备或 网络节点都分配有一个独立的 IP地址, 通过切换 IP地址实现倒换操作。 比如网络上有一个双机备份的网络服务器, 当正在提供服务的一台服务器 突然掉电或者发生异常宕机了,另一台备用的服务器会立刻切换为它的 IP 地址接管服务。
可见, 对于传统的基于电路交换实现的倒换技术, 由于主备仲裁、 倒 换通知等都需要直接由硬件实现, 因此必须背板连接或配线, 其通用性和 可扩展性差, 成本高。 对于基于 IP的双机备份实现的倒换技术, 其倒换 性能直接受 IP技术局限性的限制, 如传输时延不确定、 传输带宽不保证 等, 导致在电信网络中远远不能满足实时业务、 高服务质量的需求。
在实际应用中, 上述方案存在以下问题: 基于电路交换的倒换技术必 须釆用硬件实现, 通用性和可扩展性差, 成本高; 基于 IP的倒换技术不 能应用于实时、 高质量的电信业务中。
造成这种情况的主要原因在于, 硬件实现的倒换装置必须针对具体系 统电路设计, 不易升级和测试, 而且成本较高; IP地址切换的倒换技术直 接受 IP技术的限制, 而 IP技术具有传输时延不确定、 传输带宽不保证和 传输质量不高等缺点。
发明内容 有鉴于此, 本发明的主要目的在于提供一种通信系统中单板的倒换方 法, 使得倒换前后业务通道的切换操作简单可靠、 快速高效, 满足电信级 设备的倒换要求。
为实现上述目的, 本发明提供了一种通信系统中的倒换方法, 所述系 统包括多块单板, 其中至少一块控制单板, 用于通过总线集中控制所有单 板; 每一块单板上还包含逻辑槽位和物理槽位映射表, 用于指示每一块单 板的逻辑槽位和物理槽位之间的映射关系, 其特征在于, 单板倒换时包含 以下步骤:
A、 所述控制单板修改所有单板上的逻辑槽位和物理槽位映射表, 将 原主用单板的逻辑槽位号映射到原备用单板的物理槽位号;
B、 所述控制单板修改所述原主用和备用单板上所存储的自身逻辑槽 位号。
其中, 所述控制单板倒换时, 通过下述步骤修改所有单板上的逻辑槽 位和物理槽位映射表:
Al、 所述控制单板修改本板上的逻辑槽位和物理槽位映射表;
A2、 所述控制单板将本板更新的逻辑槽位和物理槽位映射表内容同步 到所有其他单板上的逻辑槽位和物理槽位映射表中。
所述方法还包括: 将原备用单板的遲辑槽位号映射到原主用单板的物 理槽位号。
所述通信系统基于紧凑外围设备互连总线平台。
本发明还提供了另一种通信系统中单板的倒换方法, 所述系统包括至 少两个相互通信的框, 每一个框包含多块单板, 其中至少一块控制单板, 用于通过总线集中控制该框中所有所述单板; 所述控制单板还包括通道信 息表, 用于保存框内所有单板的逻辑槽位号与收发数据存储区基址的映射 关系, 在所述控制单板需要倒换时包括以下步驟:
D、 进行总线控制切换, 原主用控制单板释放总线控制权, 原备用控 制单板接管所述总线控制权;
E、 更新存储在所述原主用控制单板和所述原备用控制单板上的通道 信息表, 将所述原主用控制单板的逻辑槽位号映射到所述原备用控制单板 的收发数据存储区基址;
F、 禁止所述原主用控制单板向所述设备单板分配接收通道。
上述第二种方法还包括:
G、 更新存储在所述控制单板上的通道信息表, 将原主用单板的逻辑 槽位号映射到原备用单板的收发数据存储区基址;
H、 更新所述原主用单板和原备用单板上用于实现底层驱动模块之间 共享内存通信的控制命令字地址信息;
I、 所述控制单板修改所述原主用单板和原备用单板上存储的自身逻 辑槽位号。
在倒换以前, 检查倒换涉及的主用和备用单板是否都在使用, 如果是 则开始倒换。
所述通信系统基于紧凑外围设备互连总线平台。
所述框之间可以通过异步传输模式进行通信。
通过比较可以发现, 本发明的技术方案与现有技术的区别在于, 在基 于共享内存通信方式的通信系统中用控制单板集中控制总线上单板的倒 换操作, 框内单板通过遷辑槽位和物理槽位映射机制来实现透明倒换, 倒 换时只需更新逻辑槽位和物理槽位映射表, 在框间通信下还需要更新发送 和接收通道映射信息以及控制命令字地址等进行倒换, 包括控制单板倒换 和设备单板倒换两种操作方式。
这种技术方案上的区别, 带来了较为明显的有益效果, 即由于共享内 存通信方式提供了方便的总线操作, 逻辑槽位和物理槽位映射机制提供了 简单的倒换操作方式, 总线的集中控制保证了倒换的可靠性和实时性, 通 道映射机制实现了框间通信情况下的倒换, 该倒换方法提供了无缝的保护 机制, 保证倒换的实时性和可靠性, 降低了网络成本, 大大提高了系统可 扩展性和网络生存能力。
附图说明
图 1为现有的倒换技术的原理示意图;
图 2为 PSM框的结构、共享内存通信机制和框间通信机制示意图; 图 3 为根据本发明的一个实施例的通过修改逻辑槽位和物理槽 位映射表实现的倒换方法流程图;
图 4 为根据本发明的一个实施例的框间通信下通过修改通道信 息表实现的倒换方法流程图。
具体实施方式 为使本发明的目的、 技术方案和优点更加清楚, 下面将结合附图对本 发明作进一步地详细描述。
本发明基于共享内存通信技术, 通过地址映射的修改完成设备单元的 倒换操作, 能满足实时性、 透明性的要求。
在共享内存通信中, 多个设备单元通过总线访问实现内存共享, 同一 总线上的各个单板通过共享内存区的读写实现数据通信, 而不同总线之间 可以通过其他接口技术实现通信。 在本发明中, 总线上各个单板的物理位 置和逻辑位置之间存在映射关系, 通过映射关系的修改实现设备单元的透 明倒换, 其可扩展性和实时性较好。
在本发明的一个实施例中, 采用基于紧凑外围设备互连 (Compact Peripheral Component Interconnect , 筒称 "CP— CI" ) 总线平台的 PSM框 实现总线上各个单板的倒换, 即设备单元的倒换。
CPCI 是一种基于外 围设备互连 ( Peripheral Component Interconnect , 简称 " PCI" )标准的高性能工业总线, 同时继承了欧式板 卡的机械特性以实现可靠结构。 PSM机框有 16个槽位, 前后都可以插板, 处理的单板主要有 4类, 分别称为 URCU、 MCU、 UGPU、 UFIU以及通用的 PCI扣板 ( PCI Mezzanine Card, 简称 "PMC" )0 其中, URCU板为 PSM框 中的控制单板, 负责 PSM框的管理功能, 釆用 1 + 1备份; UACU板为 PSM 框中的辅助控制单板, 辅助 URCU板完成控制功能, 主要用于实现 PSM框 中物理上不同段总线之间的互连, 与 URCU板成对使用; UGPU板为 PSM框 中的业务处理单元,支持各种业务类型, 1个 PSM框中最多配置 12块 UGPU 板, 系统可根据业务类型配置成 1 + 1备份或 1 + N备份等。 PMC即为 CPCI 系统中的小扣板, 可根据配置需要扣在 URCU单板或 UGPU单板上, URCU 和 UGPU都提供通用的 PMC接口, PMC—般用于完成特定的功能或增强处理 功能; UFIU板用于实现框间连接, 专门用于异步传输模式(Asynchronous Transfer Mode, 筒称 "ATM" )光网络的光接口板, 是一种 PMC扣板, 采 用 PMC接口, 扣在 URCU板上使用。
图 2 ( a ) 示出了 PSM框的结构、 组成部分以及框间通信连接方式。 在 PSM框中,各单板之间的通信主要基于 PCI总线技术, URCU板与各 UGPU板之间的通信基于 CPCI背板总线, URCU板或 UGPU板与 PMC扣板的 通信则基于承载板内部的 PCI总线。 各单板间通过 PCI桥等专用集成电路 ( Appl icat ion Specif ic Integrated Circui t s , 筒称 "ASIC" ) 芯片来 实现电气连接和地址转换功能, 即只要对各桥片进行正确配置, CPCI平台 中的所有单板都可以相互访问彼此的内存空间, 这种通过访问共享内存实 现的通信方式即为共享内存通信。 根据 PCI总线规范, PCI总线可以通过 PCI桥进行级联, 但由于受到电气等特性的限制, 每一段 PCI总线上最多 只能挂 8个 PCI设备, 所以 16槽位的 PSM框的背板实际由两段 CPCI总线 组成。 两段 CPCI总线之间通过 UACU板完成桥接, 实现跨总线的访问和总 线之间的通信。 但对于每一段 CPCI总线同时只能有一个 PCI主设备完成 总线仲裁, 因此 URCU板在主备工作模式下只有 1块主用单板用于控制两 段 CPCI总线, 备用 URCU板则与两段 CPCI总线都断开 URCU板倒换时, 备 用 URCU板抢占总线控制权, 主用 URCU板释放总线控制权, 相应的 UACU 板也与 URCU—起倒换。
PSM框的共享内存通信主要由 UFIU底层驱动、 CPCI底层驱动和 PMC 扣板底层驱动模块完成。 PSM框之间通过 ATM的服务器 +客户端模式实现, 采用 ATM光传输网络承载。 UFIU所带有的光接口连到 ATM光传输网络, UFIU 底层驱动模块完成 PSM框间的通信。 CPCI底层驱动模块完成 PSM框内 URCU 板与各 UGPU板之间的通信。 每块 PMC扣板底层通信模块完成其承载板与 PMC扣板间的通信。 可见, URCU或 UGPU板的倒换发生时, 与倒换相关的 通信模块只涉及负责框间通信的 UFIU底层驱动模块或负责框内通信的 CPCI底层驱动模块。
图 2 ( b )示出了 CPCI总线上各个单板之间内存共享机制。 CPCI底层 通信模块的工作原理就是各单板之间的内存共享。 每块单板都为其它板预 留出一块共享物理内存, 通过 PCI桥片映射到 CPCI总线上, 再通过 PCI 桥片映射到其他单板的 PCI存储空间 ,当其他板访问这段 PCI存储空间时, 实际上就访问到了源板的物理内存。 可见, PCI桥片的映射实现了不同单 板之间的内存共享。
图 2 ( c )示出了采用 UFIU实现 PSM框间通信的机制。 UFIU底层通讯 模块的设计思想同 CPCI相似, 同样是基于共享内存以及 CPCI地址空间映 射, 不同之处 UFIU通信模块基于 ATM服务器 +客户端的模式实现。 在连 接机制上, UFIU通过光接口与 ATM光网络相连,在 ATM服务器 +客户端的 模式交换下, 实现与其他 PSM框或节点之间的通信。 在 PSM框内部, UFIU 又与各个 UGPU板通信之间通过共享内存的方式实现各个 UGPU板的收发数 据功能。 每块 UGPU板为 UFIU预留一定的共享内存, 通过前述 CPCI地址 空间映射方式, 能被其它 CPCI单板, 自然也包括 URCU板上的 UFIU扣板 访问。 当 UFIU从光口接收到数据时, 由 UFIU板上的 ATM适应层(ATM adaptat ion layer 5 , 简称 "AAL5" ) 的分段与重组子层 ( Segmentat ion and Reas semble, 筒称 " SAR" )专用处理芯片, 比如型号为 CN8236的 ASIC 芯片等, 直接访问并存储在对应的接收 UGPU板的 UFIU接收共享内存区; 当各 UGPU板的数据发往 PSM框外时, 只需拷贝在 UFIU发送共享内存区, 由 UFIU板上的 SAR专用处理芯片自动取出并通过光接口转发到 ATM网上。
在承载 PSM框间互连的 ATM 网中, ATM交换机通过信头中的虚通道 ( Virtual Path, 简称 "VP" )指示符 VPI和虚通路 ( Virtual Channel , 简称 "VC" )指示符 VCI寻路。 在组成上, 一条 VP承载多条 VC, 因此 VP 可以看成是多个具有相同 VPI值的 VC集合。 VP间具有不同的 VPI值, 同 属一个 VP的各 VC具有不同的 VCI , 而分属不同的 VP的 VC可以具有相同 的 VCI值。 通常, VC可以用于直接完成两个用户间的连接, VP则可用于 不同网段间的连接。 VP 交换只切换 VPI , 信元的 VCI在整个 VP链路的传 输过程中保持不变。 VC连接在交换节点处, 根据虚通路中信元的 VPI 和 VCI查找映射表, 获得相应新的 VPI、 VCI。
由于 PSM 框间的网络结构相对稳定, 因此一般采用永久虚通道 ( Perpetual Virtual Path,简称 "PVP" )和永久虚通路 ( Perpetual Virtual Channel , 筒称 "PVC" ), PVP和 PVC的建立是根据预期的需要通过预先规 定好相应连接的 VPI /VCI 以及在各交换节点处的输入 VPI /VCI 与输出 VPI/VCI间映射关系来完成。
在 PSM框间通信时, 一般采用不同 VPI表示不同 PSM框, 而框内的不 同单板则用 VCI表示。 因此在 UFIU板上, 必须完成 VCI寻路功能, 即映 射各个单板的 VC通道, 这通过发送 VCC表( SEG VCC )和接收 VCC表( RSM VCC ) 的映射实现。
在 UFIU的 SAR专用处理芯片上的共享内存地址区域, 分配了 URCU板 和同框各 UGPU板的发送和接收队列、 SEG VCC表和 RSM VCC表以及发送緩 冲区描述 ( Send Buffer Descr iptor , 筒称 "SBD" )表等。 在 URCU的共 享内存地址区域, 主要分配了 URCU 的操作管理维护 (Operat ion , Admini s trat ion, Maintenance, 筒称 " 0AM" )信元和数据信元的收发数 据緩冲区、 收发的状态队列、 以及同框各 UGPU板的发送数据緩冲区。 在 各 UGPU板上的共享内存地址区域, 分配了本地的收发緩冲区描述符、 收 发状态队列、 本地的 SEG VCC表和 RSM VCC表及接收数据緩冲区。
UFIU的 SAR专用处理芯片主要依靠 VCC表来与外界之间进行收发 ATM 信元。 发送 VCC表位于 SAR的共享内存中, 每个单板最多允许建立 256条 发送 PVC, 表项的索引值就是发送 VCC索引; 同样对于接收 VCC表, 每个 单板最多允许建立 256条接收 PVC, 表项的索引值就是接收 VCC索引。 一一 个 VCC索引与一条 VPI、 VCI链路对应。 UFIU的 PVC配置就是为单板的框 间收发建立一个 VCC索引与 PVC的对应关系, 也即 VPI、 VCI链路与每块 单板之间的映射关系。 UFIU单板的 PVC均采用单向 PVC, 因此发送 PVC和 接收 PVC的建立和拆除是独立的, 两者索引号也是各自分配的。
可见,在框间通信下,在 PSM框内需要通过 VCC表中的 VCC索引与 PVC 的对应关系来寻址, 即发送 PVC按物理槽位号对主机发送 VCC表进行索引 配置; 接收 PVC则按逻辑槽位号对主机接收 VCC表进行索引配置。 当倒换 发生时, 外界 ATM交换不会改变, 即并不能对 VPI、 VCI链路相应的 VCC 索引倒换,因此实际倒换的就是这条 VCC索引 PVC通道的对应关系,即 VCC 项与接收物理槽位号的对应关系, 以改变接收数据单元的流向。
在本发明的一个实施例中, 采用逻辑槽位和物理槽位的映射转换关系 实现了基于 CPCI 共享内存通信的倒换方法。 系统给总线上的每个单板分 配一个逻辑槽位, 高层应用通过该逻辑槽位访问相应单板, 同时在每块单 板上的 CPCI底层驱动模块中都配置有一张逻辑槽位与物理槽位的映射表, 该表反映了物理位置上每个单板分配到的对应逻辑槽位。 通信过程中, CPCI底层驱动模块根据该表来索引物理位置上的单板。系统初始状态下逻 辑槽位和单板的实际物理槽位是一致的, 高层应用模块将数据单元和目的 单板的逻辑槽位传到 CPCI底层驱动模块, 根据该逻辑槽位 CPCI底层驱动 模块即可查找逻辑槽位与物理槽位映射表, 转换为目的单板的实际物理槽 位, 再通过 PCI 桥的共享地址空间映射, 转换成对应的共享内存的 CPCI 空间地址进行访问。 该逻辑槽位与物理槽位映射表机制直接与 PSM框的底 层相关, 可通过直接索引查找的方式完成, 因此对 CPCI通信的效率没有 影响。
在某一块或多块单板发生倒换时, 只需修改所有单板上的逻辑槽位和 物理槽位映射关系表, 即可等效的实现单板的倒换。 比如, 需要倒换的原 主用单板的逻辑槽位号为 1映射到其物理槽位号 1 , 而其备用单板的逻辑 槽位号为 2映射到其物理槽位号 2 , 则两块单板发生倒换时, 将映射表改 为逻辑槽位号 1映射到物理槽位号 2, 逻辑槽位号 2映射到物理槽位号 1 , 即等效完成了主备单板的倒换操作。 同理也可以方便的实现其他备用模式 的倒换操作。 可见, 该倒换方法不但操作方便, 而且可以高速完成, 在该 倒换操作前后, 业务通信几乎没有受到影响, 保证了业务的正常通信。
在本发明的一个较佳实施例中, 采用集中控制的方法实现倒换操作, 使得设备单元的倒换更加筒单可靠。 比如, 前述 PSM框中, CPCI控制单板 URCU板集中控制所有单板的逻辑物理槽位号映射表,每块单板将逻辑槽位 与物理槽位映射表也放在单板的共享内存中, 由 URCU板集中同步修改, 确保整个通信系统的正常运行。
在本发明的另一个实施例中 , 每块单板还保存有自身的逻辑槽位和物 理槽位信息, 则在修改逻辑和物理槽位映射关系表的同时, 需要修改相关 的发生倒换的单板中所保存的自身的逻辑槽位信息。 比如, 前面所述的 1 号物理槽位的单板和 1号物理槽位的单板发生倒换时, 除了所有映射表需 要修改以外, 存储在 1号物理槽位单板上的逻辑槽位号需要从原来的 1号 改为 2号, 而存储在 2号物理槽位单板上的逻辑槽位号需要从原来的 2号 改为 1号。
根据上述内容, 图 3示出了根据本发明的一个实施例的通过修改逻辑 槽位和物理槽位映射表实现的倒换方法具体流程。 该方法对于总线上各种 单板的倒换都适用, 包含控制单板和设备单板。
在步驟 301中, 控制单板根据单板倒换方式来更新本板上的逻辑槽位 和物理槽位映射表。 比如 PSM框中, URCU修改自身所存储的映射表。
接着进入步驟 302 , 控制单板同步所更新的逻辑槽位和物理槽位映射 表内容到所有单板上的逻辑槽位和物理槽位映射表中。比如 PSM框中, URCU 接着修改每块 UGPU板上的映射表。 修改时将原先主用的单板的逻辑槽位 号和原先备用的单板的物理槽位号对应起来, 将原先备用的单板的逻辑槽 位号和原先主用的单板的物理槽位号对应起来。 其他单板是通过逻辑槽位 号来发送数据的, 数据到达时, 控制单板可以根据映射表中的对应关系把 该数据送到倒换后的主用单板。
接着进入步骤 303 , 控制单板更新该倒换所涉及的每块单板上所存储 的其自身逻辑槽位信息。 这里所述 "自身" 指所述 "每块单板", 而非控 制单板。
在本发明的一个实施例中, 考虑到框间通信的情况, 每个框的对外接 口板保存有用于单板数据单元转发的通道信息映射表。 这里的通道信息映 射表是对外接口板用于根据对外通信的通道号来索引框内单板物理槽位 号的信息,在 PSM框中即为前述发送 VCC表和接收 VCC表。在倒换发生时, 需要修改这些信息, 才能保证框间通信的正常运行。 比如,前述 PSM框中, 基于 ATM光接口的 UFIU接口板上保存有接收 VCC表和发送 VCC表等信息, 根据这些信息来索引转发各个单板的数据单元, 完成框间通信。 因此, 在 单板倒换发生时, 根据前述基于逻辑物理槽位映射表的倒换方法更新所有 逻辑物理槽位表之后,还需要修改相关的发送 VCC表和接收 VCC表等信息; 在控制单板发生倒换时, 则需要进行总线切换操作, 将总线控制权转交给 备用单板, 将存储在原主用控制单板上的信息更新到备用单板。 需要说明 的是, 控制单板倒换时进行的主备板信息更新主要是与控制板本身相关的 信息, 如主备控制板自身的收发通道映射表和主备板之间通信的通道映射 表修改。 与 UGPU单板相关的通道映射表在平时(配置 /删除 /UGPU板倒换 ) 操作时就对主备控制板进行了同步, 所以控制板倒换时这些信息不需要再 更新。另夕卜, 由于设备单板与控制单板上均带有 ATM驱动模块,而不同 ATM 驱动模块之间需要进行通信, 该底层驱动模块通信也是通过共享内存的方 式实现的, 具体的说, 就是通过往相应的控制命令字地址写控制命令字来 实现, 因此在倒换发生时还需要修改相应的控制命令字地址。 可见, 框架 通信下另外需要的倒换操作只涉及接口板底层驱动模块, 筒单可靠, 实时 性较好。
在框间通信下, 控制单板发生倒换和各设备单板发生倒换时有不同的 处理步骤。 图 4 ( a )示出了根据本发明的一个实施例的框间通信下通过修 改通道信息表实现的控制单板倒换方法。
在步骤 411中,进行总线控制切换,原主用控制单板释放总线控制权, 原备用控制单板接管总线控制权。 比如在 PSM框中, 原备用 URCU处在总 线断开状态, 倒换发生时, 原主用 URCU通过握手转交总线控制权给原备 用 URCU。
接着进入步骤 412, 更新主备控制单板上的控制单板对应的通道信息 表。 比如在 PSM框中, URCU板发生倒换时, 需要更新主备 URCU自身的发 送 VCC表和接收 VCC表等信息, 而这些发送 VCC表和接收 VCC表都存储在 URCU上。发送 VCC表和接收 VCC表中包含了每一块单板收发数据存储区的 基址, 所谓基址是指基本的地址, 可以根据这个基值找到收发数据存储区 的起始位置, 然后根据收发数据存储区内现有数据的情况经过计算得到当 前需要收发的数据应该存放的位置。通过把原主用控制单板的逻辑槽位号 映射到原备用控制单板的收发数据存储区基址, 发向原先主用控制单板的 数据现在就送到原先的备用控制单板中, 从而实现倒换。
接着进入步骤 41 3,禁止原主用控制单板分配给各设备单板接收通道。 比如在 PSM框中, 原主用 URCU板上控制 UGPU板, 因此不能再分配给各设 备单板接收通道。 图 4 ( b )示出了根据本发明的一个实施例的框间通信下通过修改收发 通道信息实现的设备单板(除控制单板以外的其他单板)倒换方法。
在步骤 421中, 更新存储在主备用控制单板上的倒换所涉及的设备单 板相应的通道信息表。 比如在 PSM框中, 需要更新主备 URCU上所存储的 倒换涉及的 UGPU相应的发送 VCC表和接收 VCC表等信息。 如前所述, 所 有 UGPU的发送 VCC表和接收 VCC表等信息都是存储在 URCU上的 , 同时每 个 UGPU还存储有自身的发送 VCC表和接收 VCC表等信息。
接着进入步骤 422 , 更新倒换涉及的设备单板上的控制命令字地址信 息, 这里的控制命令字地址是用于实现底层驱动模块之间的共享内存通 信。 比如 PSM框中, UGPU板与 URCU板的底层 ATM驱动模块需要通过往控 制命令字地址上写控制命令字来实现通信, 因此倒换时需要更新相应的控 制命令字地址。
接着进入步骤 423 , 控制单板修改倒换所涉及的主备设备单板的发送 通道信息和接收通道信息。 这里的发送通道信息和接收通道信息对每一块 单板来说类似于自身逻辑槽位号的作用。
在实际应用中, 为了保证切换的顺利, 一般在倒换以前, 还要检查倒 换涉及的主用和备用单板是否都在使用, 如果是则开始倒换, 否则向网管 系统报警。
熟悉本领域的技术人员可以理解, 除了 CPCI平台或 PSM框以外, 在 其他基于共享内存通信方式的通信系统中, 均可以采用本发明给出的倒换 方法实现设备倒换, 而不影响本发明的实质和范围。
在本发明中实施例中 PSM框间级联方案主要采用 UFIU扣板和 ATM光 接口实现, 熟悉本领域的技术人员可以理解, 采用其他物理接口实现不同 PSM框之间或者不同 CPCI总线之间的通信时,本发明所给出的倒换方法照 样能完成设备倒换, 而不影响本发明的实质和范围。
虽然通过参照本发明的某些优选实施例, 已经对本发明进行了图示和 描述, 但本领域的普通技术人员应该明白, 可以在形式上和细节上对其作 各种各样的改变, 而不偏离所附权利要求书所限定的本发明的精神和范 围。

Claims

权 利 要 求
1. 一种通信系统中单板的倒换方法, 所述系统包括多块单板, 其中 至少一块控制单板, 用于通过总线集中控制所有单板; 每一块单板上还包 含逻辑槽位和物理槽位映射表, 用于指示每一块单板的逻辑槽位和物理槽 位之间的映射关系, 其特征在于, 单板倒换时包含以下步骤:
A、 所述控制单板修改所有单板上的逻辑槽位和物理槽位映射表, 将 原主用单板的逻辑槽位号映射到原备用单板的物理槽位号;
B、 所述控制单板修改所述原主用和备用单板上所存储的自身逻辑槽 位号。
2. 根据权利要求 1所述的通信系统中单板的倒换方法, 其特征在于, 通过下述步錄修改所有单板上的逻辑槽位和物理槽位映射表:
A1、 所述控制单板修改本板上的逻辑槽位和物理槽位映射表;
A2、 所述控制单板将本板更新的逻辑槽位和物理槽位映射表内容同步 到所有其他单板上的逻辑槽位和物理槽位映射表中。
3. 根据权利要求 1或 1所述的通信系统中单板的倒换方法, 其特征 在于, 还包括: 将原备用单板的逻辑槽位号映射到原主用单板的物理槽位 号。
4. 根据权利要求 3所述的通信系统中单板的倒换方法, 其特征在于, 所述通信系统基于紧凑外围设备互连总线平台。
5. 一种通信系统中单板的倒换方法, 所述系统包括至少两个相互通 信的框, 每一个框包含多块单板, 其中至少一块控制单板, 用于通过总线 集中控制该框中所有所述单板; 所述控制单板还包括通道信息表, 用于保 存框内所有单板的逻辑槽位号与收发数据存储区基址的映射关系, 其特征 在于:
D、 进行总线控制切换, 原主用控制单板释放总线控制权, 原备用控 制单板接管所述总线控制权;
E、 更新存储在所述原主用控制单板和所述原备用控制单板上的通道 信息表 , 将所述原主用控制单板的逻辑槽位号映射到所述原备用控制单板 的收发数据存储区基址; F、 禁止所述原主用控制单板向所述设备单板分配接收通道。
6. 根据权利要求 5所述的通信系统中单板的倒换方法, 其特征在于, 还包括:
G、 更新存储在所述控制单板上的通道信息表, 将原主用单板的逻辑 槽位号映射到原备用单板的收发数据存储区基址;
H、 更新所述原主用单板和原备用单板上用于实现底层驱动模块之间 共享内存通信的控制命令字地址信息;
I、 所述控制单板修改所述原主用单板和原备用单板上存储的自身逻 辑槽位号。
7. 根据权利要求 5或 6所述的通信系统中单板的倒换方法, 其特征 在于, 在倒换以前, 检查倒换涉及的主用和备用单板是否都在使用, 如果 是则开始倒换。
8. 根据权利要求 5或 6所述的通信系统中单板的倒换方法, 其特征 在于, 所述通信系统基于紧凑外围设备互连总线平台。
9. 根据权利要求 5或 6所述的通信系统中单板的倒换方法, 其特征 在于, 所述框之间通过异步传输模式进行通信。
PCT/CN2005/000637 2004-05-10 2005-05-09 Procede de realisation d'une inversion a plaque simple dans un systeme de communication WO2005109665A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200410044466.8 2004-05-10
CNB2004100444668A CN100342666C (zh) 2004-05-10 2004-05-10 通信系统中的倒换方法

Publications (1)

Publication Number Publication Date
WO2005109665A1 true WO2005109665A1 (fr) 2005-11-17

Family

ID=35320528

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2005/000637 WO2005109665A1 (fr) 2004-05-10 2005-05-09 Procede de realisation d'une inversion a plaque simple dans un systeme de communication

Country Status (2)

Country Link
CN (1) CN100342666C (zh)
WO (1) WO2005109665A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101207409B (zh) * 2006-12-21 2012-01-04 华为技术有限公司 一种业务倒换的方法及一种单板
CN111478947A (zh) * 2020-03-20 2020-07-31 珠海高凌信息科技股份有限公司 主备控制板实时同步方法及系统
TWI723217B (zh) * 2016-10-17 2021-04-01 南韓商愛思開海力士有限公司 記憶體系統及其操作方法

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100459503C (zh) * 2006-08-25 2009-02-04 华为技术有限公司 一种获取单板的物理槽位号和槽位类型的方法及系统
CN101425929B (zh) * 2008-12-01 2011-02-09 成都市华为赛门铁克科技有限公司 一种板号表分配方法、装置和系统
CN103220373B (zh) * 2012-01-19 2018-05-04 中兴通讯股份有限公司 一种分布式逻辑地址与物理地址互换方法、装置及设备
CN106603303A (zh) * 2016-12-30 2017-04-26 杭州迪普科技股份有限公司 云聚合板卡的切换方法和装置
CN107124316B (zh) * 2017-05-23 2019-10-11 烽火通信科技股份有限公司 一种数据通信设备中基于硬件的快速倒换动作实现方法
CN111049566B (zh) * 2019-11-20 2022-03-08 中国航空工业集团公司西安航空计算技术研究所 信息传递方法和机载lrm模块

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1321004A (zh) * 2000-04-25 2001-11-07 华为技术有限公司 交换机的主备倒换方法及其实现装置
JP2002247017A (ja) * 2001-02-20 2002-08-30 Yamatake Building Systems Co Ltd 通信回線切替装置
TW542961B (en) * 2000-12-29 2003-07-21 Chunghwa Telecom Co Ltd Fault tolerance system for memory sharing by multiple buses
US20030212927A1 (en) * 2002-05-08 2003-11-13 Navar Venu Gopal Fault-protection mechanism for protecting multi-protocol-label switching (MPLS) capability within a distributed processor router operating in an MPLS network

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1321004A (zh) * 2000-04-25 2001-11-07 华为技术有限公司 交换机的主备倒换方法及其实现装置
TW542961B (en) * 2000-12-29 2003-07-21 Chunghwa Telecom Co Ltd Fault tolerance system for memory sharing by multiple buses
JP2002247017A (ja) * 2001-02-20 2002-08-30 Yamatake Building Systems Co Ltd 通信回線切替装置
US20030212927A1 (en) * 2002-05-08 2003-11-13 Navar Venu Gopal Fault-protection mechanism for protecting multi-protocol-label switching (MPLS) capability within a distributed processor router operating in an MPLS network

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101207409B (zh) * 2006-12-21 2012-01-04 华为技术有限公司 一种业务倒换的方法及一种单板
TWI723217B (zh) * 2016-10-17 2021-04-01 南韓商愛思開海力士有限公司 記憶體系統及其操作方法
CN111478947A (zh) * 2020-03-20 2020-07-31 珠海高凌信息科技股份有限公司 主备控制板实时同步方法及系统
CN111478947B (zh) * 2020-03-20 2023-02-17 珠海高凌信息科技股份有限公司 主备控制板实时同步方法及系统

Also Published As

Publication number Publication date
CN100342666C (zh) 2007-10-10
CN1697337A (zh) 2005-11-16

Similar Documents

Publication Publication Date Title
WO2005109665A1 (fr) Procede de realisation d'une inversion a plaque simple dans un systeme de communication
JP3109591B2 (ja) Atm交換機
TWI233018B (en) Multi-port high-speed serial fabric interconnect chip in a meshed configuration
US7233570B2 (en) Long distance repeater for digital information
US5956337A (en) ATM switch interface
Malavalli High-speed fiber channel switching fabric services
US20030152182A1 (en) Optical exchange method, apparatus and system for facilitating data transport between WAN, SAN and LAN and for enabling enterprise computing into networks
US6285674B1 (en) Hybrid distributed broadcast and unknown server for emulated local area networks
EP0961443B1 (en) Switching system comprising a mask mechanism for altering the internal routing process
WO2022017099A1 (zh) 通信方法、cp设备及nat设备
GB2401518A (en) Efficient arbitration using credit based flow control
KR100429899B1 (ko) Fudcom에 의한 결함허용 제어장치 및 방법
Cisco ATM-focused supplement to Packet Magazine
Cisco ATM-focused supplement to Packet Magazine
Cisco ATM-focused supplement to Packet Magazine
Cisco ATM-focused supplement to Packet Magazine
Cisco D
Cisco D
Cisco D
Cisco D
Cisco D
US11909708B2 (en) Protection switching device with link selective switching between host devices for maintenance of communication connection
US7145909B2 (en) Packet switching access platform
KR19980061785A (ko) Atm 교환기에서 망관리보드의 이중화회로
Garcia et al. Future I/O

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Country of ref document: DE

122 Ep: pct application non-entry in european phase