[go: up one dir, main page]

WO2018000135A1 - Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps - Google Patents

Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps Download PDF

Info

Publication number
WO2018000135A1
WO2018000135A1 PCT/CN2016/087340 CN2016087340W WO2018000135A1 WO 2018000135 A1 WO2018000135 A1 WO 2018000135A1 CN 2016087340 W CN2016087340 W CN 2016087340W WO 2018000135 A1 WO2018000135 A1 WO 2018000135A1
Authority
WO
WIPO (PCT)
Prior art keywords
port
switching device
network switching
ports
exchange
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2016/087340
Other languages
English (en)
Chinese (zh)
Inventor
程浩
程宏涛
赵冠凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
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
Priority to PCT/CN2016/087340 priority Critical patent/WO2018000135A1/fr
Priority to CN201680086677.9A priority patent/CN109314658A/zh
Publication of WO2018000135A1 publication Critical patent/WO2018000135A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/14—Two-way operation using the same type of signal, i.e. duplex
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/50—Circuit switching systems, i.e. systems in which the path is physically permanent during the communication
    • H04L12/52—Circuit switching systems, i.e. systems in which the path is physically permanent during the communication using time division techniques

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to a network switching device and a method for time slot exchange.
  • the switch fabric is an important part of the network switching device. It directly determines the application performance of the network switching device.
  • the network switching device can be a router or a switch.
  • Crossbar is called crossbar matrix or crossbar switch matrix. It is the industry's recognized switch fabric for building high-capacity systems. The first three switch fabrics share bandwidth to some extent, and the Crossbar switch fabric adopts The matrix structure implements non-blocking switching, breaking the bandwidth limitation.
  • Each port in the Crossbar architecture has a certain number of cache structures.
  • the slot exchange of the Crossbar architecture refers to the exchange of service data in the cache structure of different ports. Since the exchange relationship between the output port and the input port of the Crossbar switch fabric is random, in order to avoid conflicts in the data exchange process, the prior art is for the N ⁇ N Crossbar architecture, that is, N input ports and N Output port, need to set up N 2 cache structure.
  • the hardware redundancy of the Crossbar switch fabric is very large, wasting hardware resources, and the hardware consumed by the Crossbar switch fabric increases exponentially with the number of ports.
  • the Crossbar architecture is increased from 4 ports to 5 ports.
  • the number of cache structures will increase from 16 to 25. Due to the high hardware consumption, it is difficult to continue to expand when the existing Crossbar switch fabric switch ports reach tens or hundreds of scales.
  • the object of the present invention is to provide a network switching device and a time slot exchange method for solving the present problem.
  • a first aspect provides a network switching device, including: a processor, N ports, at least N memories, each of the memories corresponding to one of the ports; N being a positive integer greater than 1; Storing service data received by the corresponding port in each time slot; wherein the memory includes a plurality of storage areas, each of the storage areas storing a service received by the port in a time slot corresponding to the storage area Data; the processor is configured to exchange, between the ports, the service data separately stored by the at least N memories in a switching period; wherein the switching period includes multiple time slots, each of the In the time slot, the processor controls each of the storage areas of the memory to be written into the service data or read a piece of the service data.
  • the number of memories included in the network switching device is linearly related to the number of ports of the network switching device, that is, the number of memories increases linearly with the number of network device ports, and the amount of memory can be at least The number of ports is the same.
  • the number of memory of the network switching device is the square of the number of ports.
  • the network switching device provided by the present invention reduces hardware consumption, and at the same time, since the storage area of each memory is in the same time slot. Write a unique business data or read a unique business data, thus avoiding conflicts caused by time slot exchange.
  • the first port is in a duplex mode, and the first port is any one of the at least N ports; For: using the first port as an input port; storing, in the switching cycle, the service data stored in each storage area of the memory corresponding to the first port in a memory corresponding to the other ports . That is to say, the network switching device completes a service exchange in one switching cycle, that is, it needs to read all the storage areas in each memory once.
  • the second port is in a duplex mode, and the second port is any one of the at least N ports; And the second port is used as an output port, in which the service data is respectively read from the memory corresponding to the other port and each memory of the memory corresponding to the second port is written.
  • the network switching device needs to write once for all storage areas in each memory in one switching cycle.
  • the network switching device includes a preset routing table, where the routing table includes a time slot exchange relationship between each of the ports;
  • the time slot exchange relationship indicates data exchange between any two of the at least N memories;
  • the processor is further configured to determine an execution time of each time slot exchange in the routing table,
  • the conflict-free routing table is configured to: exchange, according to the conflict-free routing table, the service data separately stored by the at least N memories in the exchange period between the ports.
  • the time slot exchange relationship between each of the ports is related to the service handled by the network switching device.
  • the device includes N memories, and the memory corresponds to the port one-to-one . That is to say, in the specific implementation of the present invention, the number of memories and the number of ports can be the same, and the present invention reduces hardware consumption compared to the number of ports in the existing network switching device is the square of the number of ports.
  • a method for time slot switching is provided, the method being used for a network switching device, where the network switching device includes N ports, at least N memories, each of the memories corresponding to one of the ports, The memory includes a plurality of storage areas, each of the storage areas storing service data received by the port in a time slot corresponding to the storage area, where N is a positive integer greater than one; the method includes: the network switching The device reads the service data stored in each storage area of the memory corresponding to the input port during the exchange period; the network switching device writes a service to each storage area of the memory corresponding to the output port in the exchange period Data; wherein the switching period includes a plurality of the time slots, and in each of the time slots, the network switching device controls each of the storage areas in the memory to be written into the service data Or read a copy of the business data.
  • the first port is In the duplex mode, the first port is any one of the at least N ports; the network switching device reads service data stored in each storage area of the memory corresponding to the input port during the switching period, including The network switching device uses the first port as an input port; the network switching device separately writes the service data stored in each storage area of the memory corresponding to the first port in the switching cycle Corresponding to the memory of the other port.
  • the second port is in a duplex mode, and the second port is any one of the at least N ports;
  • the network switching device Writing a service data to each storage area of the memory corresponding to the output port during the exchange period specifically includes: the network switching device uses the second port as an output port; and the network switching device is in the exchange In the cycle, the service data is respectively read from the memory corresponding to the other ports and written into each storage area of the memory corresponding to the second port.
  • the network switching device includes a preset routing table, where the routing table includes a time slot exchange relationship between each of the ports; The time slot exchange relationship indicates data exchange between any two of the at least N memories; the method further comprising: the network switching device determining execution of each time slot exchange in the routing table At the moment, obtaining a collision-free routing table; the network switching device reads the service data stored in each storage area of the memory corresponding to the input port in the exchange period, and stores the memory corresponding to the output port in the exchange period Writing, by the storage area, a service data, comprising: the network switching device reading the service data stored in each storage area of the memory corresponding to the input port in the exchange period according to the conflict-free routing table, and in the A service data is written to each storage area of the memory corresponding to the output port during the exchange period.
  • each of the memory stores M service data, where the switching period includes at least M Time slots.
  • a computer readable medium for storing a computer program comprising instructions for performing the method of any of the second aspect or any of the possible implementations of the second aspect.
  • FIG. 1 is a schematic diagram of a Crossbar switching architecture in the prior art
  • FIG. 2 is a schematic diagram of an output port in the Crossbar switch fabric shown in FIG. 1 reading data from an input port;
  • FIG. 3 is a schematic diagram of inputting data from an output port of an input port in the Crossbar switching architecture shown in FIG. 1;
  • FIG. 4 is a schematic structural diagram of a network switching device according to an embodiment of the present disclosure.
  • FIG. 5 is a schematic diagram of an output port in the Crossbar switch fabric shown in FIG. 4 reading data from an input port;
  • FIG. 6 is a schematic diagram of an output port in the Crossbar switching architecture shown in FIG. 4 reading data from an input port;
  • FIG. 7 is a schematic flowchart diagram of a method for time slot exchange according to an embodiment of the present invention.
  • the technology described in this paper can be used in various communication systems, such as 2G systems such as Global System for Mobile Communications (GSM), Wideband Code Division Multiple Access (English name: Wideband Code Division Multiple Access Wireless) 3G system such as WCDMA), 4G system such as Long Term Evolution (LTE) system, 5G communication system of LTE subsequent evolution, and wireless local area network (English name: wireless local area network, referred to as: WLAN)
  • GSM Global System for Mobile Communications
  • WCDMA Wideband Code Division Multiple Access
  • 4G system such as Long Term Evolution (LTE) system
  • 5G communication system of LTE subsequent evolution and wireless local area network
  • WLAN wireless local area network
  • the Crossbar architecture consists of an input port, an output port, and a crossbar.
  • the port size is N
  • the input port is denoted by i, 1 ⁇ i ⁇ N
  • the output port is denoted by j, 1 ⁇ j ⁇ N
  • the crosspoint switch is denoted as XP ij.
  • FIG. 1 is a schematic structural diagram of a network switching device adopting a Crossbar switching architecture.
  • the network switching device has a port size of 4, including 16 crosspoint switches.
  • each input port has an independent read cache structure for all output ports, such as Figure 1 shows the read cache structure R shown near each crosspoint switch.
  • Another design is to avoid conflicts when multiple input ports simultaneously write data to the same output port.
  • Each output port has an independent write cache structure for all input ports, as shown in Figure 1.
  • a write cache structure W is shown near each crosspoint switch.
  • Port duplex mode means that the port can transmit in both directions, and it can receive data as well as output data.
  • the duplex mode includes a full duplex mode and a half duplex mode.
  • the port works in full-duplex mode, the port can transmit in both directions at the same time, that is, the port can be used as an output port or an input port at the same time; the port works in half-duplex mode.
  • the port can only transmit in one direction at a time, that is, the port can only be used as an output port or an output port at the same time.
  • the input port 1 and the output port 1 may be a port in a working mode, so that the network switching device shown in FIG. 1 may include four working in duplex mode. Port.
  • Time slot switching refers to the exchange of signals on different ports and different time slots. It should be noted that the time slot exchange used by the network switching device in the embodiment of the present invention may be based on time division multiplexing (English full name: Time Division Multiplex, TDM for short), wherein the TDM exchange is based on the input time slot and the output time. The result of the configuration of the slot can exchange data of the input slot into the same or different output slots.
  • TDM Time Division Multiplex
  • FIG. 1 Still taking the network switching device shown in FIG. 1 as an example, where input port 1 and output port 1 are port 1 on the network switching device, and input port 2 and output port 2 are port 2 on the network switching device, input Port 3 and output port 3 are port 3 on the network switching device, and input port 4 and output port 4 are port 4 on the network switching device.
  • Table 1 below is a schematic diagram of time slot exchange between port 1 and port 4.
  • Each execution time in Table 1 is the start time of performing a data exchange operation, and the duration between two adjacent execution times is one time slot, and the execution time 1 to the execution time 4 in Table 1 are networks.
  • the switching period of the network switching device is specified by the switching protocol adopted by the network switching device, for example, the E1 protocol or the Synchronous Digital Hierarchy (SDH) protocol.
  • p1 refers to port 1
  • p2 refers to port 2
  • p3 refers to port 3
  • p4 refers to port 4.
  • AC1 refers to the first storage area of the service data received by the storage port at the execution time 1 in the cache structure
  • AC2 refers to the second storage area of the service data received by the storage port at the execution time 2 in the cache structure
  • AC3 refers to the cache.
  • the storage area stores the third storage area of the service data received at time 3
  • AC4 refers to the fourth storage area of the service data received by the storage port at the execution time 4 in the cache structure.
  • p1.AC3->p1.AC1 means that the service data in the third storage area in the cache structure corresponding to port 1 is written into the first storage area in the cache structure corresponding to port 1.
  • time slot exchange relationship is determined according to the service processed by the network switching device, that is, if the service handled by the network switching device does not change, the data exchange relationship between the storage areas of all the cache structures is fixed.
  • each time slot exchange is performed and the time slot number is bound. For example, the time slot exchange performed at the execution time 1 is to write data into the first storage area of the cache structure.
  • each column is viewed separately, and each port obtains data from other ports as an output port.
  • each port has an independent cache structure for other ports. For example, at the execution time 2 shown in Table 1, port 3 serves as an input port to the port at the same time. The data written by 1 and port 2 is the data in different cache structures, avoiding conflicts.
  • time slot exchange of the network switching device is illustrated from the perspective of the input port and the output port, respectively.
  • FIG. 2 is a schematic diagram of an output port reading data from an input port.
  • an input port establishes a corresponding independent read buffer structure for all output ports, as shown in FIG. a cache structure in which each output port reads data from a plurality of input ports, for example, the output port 1 is from the cache structure 1 of the input port 1, the cache structure 5 of the input port 2, and the cache structure 9 of the input port 3,
  • the cache structure 13 of the input port 4 reads data.
  • FIG. 3 is a schematic diagram of an input port writing data to an output port.
  • an output port establishes a corresponding independent write cache structure for all input ports, as shown in FIG. a cache structure, wherein each input port writes data to a plurality of output ports, for example, the input port 1 is respectively directed to the cache structure 1 of the output port 1, the cache structure 2 of the output port 2, and the cache structure 4 of the output port 3, The cache structure 4 of the output port 4 writes data.
  • the existing Crossbar switching architecture has the technical problem of excessive hardware consumption.
  • the embodiment of the present invention provides a network switching device and a time slot exchange method, which reduces hardware consumption under the premise of avoiding collision of time slot exchange.
  • the embodiment of the present invention provides a network switching device 40, where the network switching device can be The router can also be a switch. As shown in FIG. 4, the network switching device 40 includes:
  • a processor such as the processor 41 shown in FIG. 4;
  • N ports 42 as shown in Figure 4, port 1, port 2, port 3 and port 4;
  • At least N memories 43 such as memory 1, memory 2, memory 3 and memory 4 as shown in FIG. 4; wherein each of said memories corresponds to one of said ports; N is a positive integer greater than one;
  • the memory may be used to cache business data.
  • the memory 43 is configured to store service data received by a corresponding port in each time slot; wherein the memory 43 includes a plurality of storage areas, each of the storage areas storing the port in a corresponding storage area Service data received in one time slot;
  • the processor 41 is configured to exchange service data stored by the at least N memories in each of the ports during a switching period, where the switching period includes multiple time slots, each of the In the time slot, the processor controls each of the storage areas of the memory to be written into the service data or read a piece of the service data.
  • the network switching device is used, and the number of the memory included in the network switching device is linearly related to the number of ports of the network switching device, that is, the number of the memory increases linearly with the number of the network device ports, compared with the prior art.
  • the number of memories of the network switching device is the square of the number of ports, and the embodiment of the present invention reduces hardware consumption, and at the same time, since the storage area of each memory is written in the same time slot, only one service data is written or is read uniquely. A piece of business data that avoids conflicts that occur when time slots are exchanged.
  • the network switching device provided by the embodiment of the present invention is described in detail below in order to enable a person skilled in the art to better understand the technical solutions provided by the embodiments of the present invention.
  • the processor 41 is specifically configured to use the first port as an input port, and store service data stored in each storage area of the memory corresponding to the first port in a memory corresponding to the other ports in a switching cycle.
  • the first port is any one of the at least N ports, that is, the network switching device completes a service exchange in one switching cycle, that is, needs to perform all storage areas in each memory. Read once.
  • the processor 41 is further configured to use the second port as an output port, in the exchange period
  • Each of the memories corresponding to the other ports is read into each storage area of the memory corresponding to the second port.
  • the second port is any one of the at least N ports, that is, the network switching device needs to write once for all storage areas in each memory in one switching cycle.
  • first port and the second port may be the same port, that is, a port in the duplex mode of the network switching device, which can serve as an input port or an output port at different times.
  • the execution time 1 to the execution time 4 are the times of performing the data exchange operation four times included in one exchange cycle, and each port reads data from other ports as the output port in the exchange cycle. It also writes data to other ports as an output port.
  • each port writes the service data stored in each storage area of its corresponding memory to the memory of the other port under the control of the processor as an input port in one switching cycle, and also serves as an output port.
  • the service data is read from other ports and written into each storage area of the corresponding memory.
  • each memory is written into a single service data or a unique service data is read in one time slot, so that multiple ports are not simultaneously generated to a certain storage area of one port. Data conflict caused by writing or reading data.
  • the memory and the port in the network switching device may have a one-to-one correspondence. That is to say, only one memory is set up for each port of the network switching device.
  • the network switching device exchanges the service data in the storage area of each memory according to the routing table as shown in Table 2, which avoids conflicts between data read and write.
  • the number of memories is the square of the number of ports to avoid data collision. The embodiment of the invention reduces the hardware consumption of the network switching device.
  • each of the memories stores M shares of the service data, that is, the memory includes M storage areas, and the switching period includes at least M time slots.
  • the network switching device can also perform time slot switching in more than M time slots to avoid data collisions.
  • the processor can always calculate a collision-free routing table to ensure that the time slot exchange between the ports does not cause a collision.
  • the network switching device includes a preset routing table, where the routing table includes a slot exchange relationship between each of the ports; the slot exchange relationship indicates any of the at least N memories Data exchange between two storage areas; the processor is further configured to: determine an execution time of each time slot exchange in the routing table, to obtain a collision-free routing table; the processor is specifically configured to: according to the The collision-free way exchanges the service data respectively stored by the at least N memories in the exchange period between the ports.
  • the attribute of the service sent by the upper-layer network of the network switching device to the network switching device determines the data exchange relationship between the storage areas. Therefore, when the network switching device is initially started, the routing table can be preset according to the service. .
  • n color patches each of which is m, and it is randomly placed into a grid matrix of m rows and n columns, and each grid is placed with a color patch, and it is required to take all the color patches in m times. Every Take n times, and ask to take each color column and take only one color block at a time, and each of these color blocks should be different colors.
  • the n color color patches represent n ports, and each color block m represents each port buffer m service data.
  • the network switching device can always calculate the conflict-free routing table according to the preset routing table. So that the network switching device can exchange information without conflict during one switching cycle.
  • the network switching device can calculate the collision-free routing table by using the branching method, the dichotomy method or the Euler path search. Specifically, based on the above mathematical model, the branching method, the dichotomy method or the Euler path search can be used to calculate the collision-free routing. table. It can be understood that the conflict-free routing table can be obtained by other methods, which is not specifically limited in this embodiment of the present invention.
  • time slot exchange of the network switching device provided by the embodiment of the present invention is described below from the perspective of the input port and the output port.
  • FIG. 5 is a schematic diagram of an output port reading data from an input port.
  • the processor controls each output port to read data from a plurality of input ports.
  • the output port 1 is respectively input from the memory 1 of the input port 1.
  • the memory 2 of the input port 2, the memory 3 of the input port 3, and the memory 4 of the input port 4 read the data. It is worth noting that the same output port reads the service data from the memory of the different input ports at different time slots.
  • FIG. 6 is a schematic diagram of an input port writing data to an output port, wherein the processor controls each input port to write data to a plurality of output ports, for example, the input port 1 is respectively directed to the memory 1 of the output port 1, and the output port 2 is The memory 2, the memory 3 of the output port 3, and the memory 4 of the output port 4 write data. It is worth noting that the same input port writes service data to the memory of the different output ports at different time slots.
  • the processor 41 in the embodiment of the present invention may be a central processing unit (English name: Center Processing Unit, abbreviated as CPU).
  • the processor 41 may also be a field programmable gate array (English name: Field Programmable Gate Array, FPGA for short) to implement all operations of the time slot exchange in the embodiment of the present invention, or
  • the processor 41 can also In the case of the CPU and the FPGA, the FPGA and the CPU respectively perform part of the operation of the time slot exchange in the embodiment of the present invention.
  • the memory 43 in the embodiment of the present invention may be a random access memory (English name: Random-Access Memory, RAM for short) for buffering service data, or may be a register file.
  • the storage area is a storage address in the memory, and the storage area may be a physical storage address in the memory or a logical storage address, which is not limited in the present invention.
  • the network switching device shown in FIG. 4 is only an example.
  • the number of ports included in the network switching device is not limited by the present invention.
  • the network switching device may further include other components. Not shown in the figure.
  • the embodiment of the present invention further provides a method for time slot exchange, where the method is applied to a network switching device, where the network switching device includes N ports, at least N memories, and each of the memories corresponds to one of the ports.
  • the memory includes a plurality of storage areas, each of the storage areas storing service data received by the port in a time slot corresponding to the storage area, where N is a positive integer greater than one.
  • the method includes:
  • the network switching device reads, in a switching period, service data stored in each storage area of the memory corresponding to the input port.
  • the network switching device writes a service data to each storage area of the memory corresponding to the output port in the exchange period.
  • the switching period includes a plurality of the time slots, and in each of the time slots, the network switching device controls each of the storage areas in the memory to be written into the service data or is Read a copy of the business data.
  • the storage area of each memory is written into a single service data in the same time slot or a single service data is read, setting a memory for each port does not occur.
  • the number of the memory of the network switching device can be linearly compared with the number of the ports. Compared with the number of the existing network switching devices, the number of the memory increases exponentially with the number of ports, and the embodiment of the present invention reduces hardware consumption and is beneficial to increase. Large exchange scale.
  • the first port is in a duplex mode, and the first port is any one of the at least N ports; then the step S701 includes: the network switching device uses the first port as And inputting, in the exchange period, the service data stored in each storage area of the memory corresponding to the first port is respectively written into a memory corresponding to the other ports. That is to say, the network switching device completes a service exchange in one switching cycle, and needs to read all the storage areas in each memory once.
  • the second port is in a duplex mode, and the second port is any one of the at least N ports; then the step S702 includes: the network switching device uses the second port as an output port The service data is respectively read from the memory corresponding to the other port in the exchange cycle and written into each storage area of the memory corresponding to the second port. That is to say, the network switching device needs to write once for all storage areas in each memory in one switching cycle.
  • the network switching device includes a preset routing table, where the routing table includes a time slot exchange relationship between each of the ports; the time slot exchange relationship Determining data exchange between any two storage areas of the at least N memories; the method specifically includes: determining, by the network switching device, an execution time of each time slot exchange in the routing table, obtaining no a conflict routing table, and reading, according to the conflict-free routing table, service data stored in each storage area of the memory corresponding to the input port in the exchange period, and each of the memories corresponding to the output port in the exchange period The storage area writes a business data.
  • the time slot exchange relationship between each port is related to the service handled by the network switching device.
  • each of the memories stores M pieces of service data
  • the switching period includes at least M time slots.
  • all the memories include four storage areas, and the period lengths of Table 1 and Table 2 are the same, each including four time slots, that is, the embodiment of the present invention and the existing Compared with technology, hardware consumption is reduced without affecting the switching delay.
  • the network switching device may also perform time slot exchange in more than M time slots to avoid data collision.
  • the method for the network switching device to exchange time slots may be implemented in the form of a hardware plus software functional unit.
  • the above-described integrated unit implemented in the form of a software functional unit can be stored in a computer readable storage medium.
  • the above software functional unit is stored in a storage medium and includes a number of instructions for causing a router or switch to perform the above-described method of exchanging time slots.
  • the foregoing storage medium includes various non-volatile storage media that can store data, such as a USB flash drive, a mobile hard disk, a random access memory, a magnetic disk, or an optical disk.

Landscapes

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

Abstract

L'invention concerne un dispositif de commutation de réseau et un procédé d'échange d'intervalles de temps servant à résoudre le problème technique comme quoi la consommation de matériel de l'architecture de commutation à barres croisées existante est trop élevée. Le dispositif de commutation de réseau comprend un processeur, N ports, au moins N mémoires, chacune des mémoires correspondant à l'un des ports. La mémoire est utilisée pour stocker des données de trafic reçues par le port correspondant à chaque intervalle de temps. La mémoire comporte une pluralité de zones de stockage, dont chacune stocke les données de trafic reçues par le port à un intervalle de temps correspondant à la zone de stockage. Le processeur est utilisé pour échanger des données de trafic stockées dans lesdites au moins N mémoires respectivement parmi chacun des ports au cours du cycle d'échange. Le cycle d'échange comporte une pluralité d'intervalles de temps. Dans chacun des intervalles de temps, le processeur commande chacune des zones de stockage de la mémoire à des fins d'écriture d'un élément de données de trafic ou à des fins de lecture d'un élément de données de trafic.
PCT/CN2016/087340 2016-06-27 2016-06-27 Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps Ceased WO2018000135A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2016/087340 WO2018000135A1 (fr) 2016-06-27 2016-06-27 Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps
CN201680086677.9A CN109314658A (zh) 2016-06-27 2016-06-27 网络交换设备及时隙交换的方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2016/087340 WO2018000135A1 (fr) 2016-06-27 2016-06-27 Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps

Publications (1)

Publication Number Publication Date
WO2018000135A1 true WO2018000135A1 (fr) 2018-01-04

Family

ID=60785011

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/087340 Ceased WO2018000135A1 (fr) 2016-06-27 2016-06-27 Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps

Country Status (2)

Country Link
CN (1) CN109314658A (fr)
WO (1) WO2018000135A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11307913B2 (en) 2020-08-26 2022-04-19 Coupang Corp. Systems and methods for automated application programming interface evaluation and migration
CN114697275A (zh) * 2020-12-30 2022-07-01 深圳云天励飞技术股份有限公司 数据处理方法和装置
WO2024187353A1 (fr) * 2023-03-13 2024-09-19 华为技术有限公司 Procédé de routage, procédé de configuration d'informations de routage et dispositif électronique

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5291482A (en) * 1992-07-24 1994-03-01 At&T Bell Laboratories High bandwidth packet switch
CN1274224A (zh) * 1999-05-12 2000-11-22 深圳市华为技术有限公司 流水线式共享存储器读写方法
CN1283379A (zh) * 1997-11-06 2001-02-07 耐特因塞特公司 在时分多路复用网络的位流之间交换数据的方法和设备
EP1233580A1 (fr) * 2001-02-17 2002-08-21 Alcatel Système de planification multi-serveurs et procédé pour un élément à commutation rapide
CN101212424A (zh) * 2006-12-28 2008-07-02 杭州华三通信技术有限公司 融合了电路交换和分组交换的以太网交换方法与设备

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101212822B (zh) * 2006-12-28 2010-12-01 杭州华三通信技术有限公司 在以太网上进行同步时分交换的以太网交换方法与设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5291482A (en) * 1992-07-24 1994-03-01 At&T Bell Laboratories High bandwidth packet switch
CN1283379A (zh) * 1997-11-06 2001-02-07 耐特因塞特公司 在时分多路复用网络的位流之间交换数据的方法和设备
CN1274224A (zh) * 1999-05-12 2000-11-22 深圳市华为技术有限公司 流水线式共享存储器读写方法
EP1233580A1 (fr) * 2001-02-17 2002-08-21 Alcatel Système de planification multi-serveurs et procédé pour un élément à commutation rapide
CN101212424A (zh) * 2006-12-28 2008-07-02 杭州华三通信技术有限公司 融合了电路交换和分组交换的以太网交换方法与设备

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11307913B2 (en) 2020-08-26 2022-04-19 Coupang Corp. Systems and methods for automated application programming interface evaluation and migration
CN114697275A (zh) * 2020-12-30 2022-07-01 深圳云天励飞技术股份有限公司 数据处理方法和装置
CN114697275B (zh) * 2020-12-30 2023-05-12 深圳云天励飞技术股份有限公司 数据处理方法和装置
WO2024187353A1 (fr) * 2023-03-13 2024-09-19 华为技术有限公司 Procédé de routage, procédé de configuration d'informations de routage et dispositif électronique
WO2024187817A1 (fr) * 2023-03-13 2024-09-19 华为技术有限公司 Procédé et appareil de transmission, et dispositif et réseau

Also Published As

Publication number Publication date
CN109314658A (zh) 2019-02-05

Similar Documents

Publication Publication Date Title
US10476697B2 (en) Network-on-chip, data transmission method, and first switching node
US9477280B1 (en) Specification for automatic power management of network-on-chip and system-on-chip
US10818325B2 (en) Data processing method and data processing system for scalable multi-port memory
US9571420B2 (en) Integrated NoC for performing data communication and NoC functions
US10554496B2 (en) Heterogeneous SoC IP core placement in an interconnect to optimize latency and interconnect performance
US7802025B2 (en) DMA engine for repeating communication patterns
CN102685017A (zh) 一种基于fpga的片上网络路由器
JP6383793B2 (ja) 可変な数のコア群、入出力（ｉ／ｏ）装置、ディレクトリ構造、及びコヒーレンシポイントを有する、キャッシュコヒーレントｎｏｃ（ネットワークオンチップ）
US20200393991A1 (en) Efficient data movement method for in storage computation
WO2018018874A1 (fr) Procédé de traitement de mémoire cache de données et système de traitement de données destinés à un paquet entièrement partagé 4r4w
WO2018000135A1 (fr) Dispositif de commutation de réseau et procédé d'échange d'intervalles de temps
US20220015588A1 (en) Dual mode interconnect
EP3631639B1 (fr) Communications pour dispositif de réseau prédiffusé programmable par l'utilisateur
US9942146B2 (en) Router path selection and creation in a single clock cycle
CN105630727A (zh) 多SoC节点之间的访问方法、装置和系统
CN105446935B (zh) 共享存储并发访问处理方法及装置
CN117742996A (zh) 用于计算的通信优化方法、装置、电子设备和存储介质
CN102013984B (zh) 二维网状片上网络系统
CN104991745A (zh) 一种存储系统数据写入方法和系统
CN108541365B (zh) 用于交换机中拥塞信息的分发的设备和方法
US12056072B1 (en) Low latency memory notification
US12242753B2 (en) Reduced network load with combined put or get and receiver-managed offset
CN111357016A (zh) 用于神经网络处理器的片上通信系统
US20150101035A1 (en) Duplicating packets efficiently within a network security appliance
CN103313439A (zh) 单卡多模终端及其通信处理方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16906539

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16906539

Country of ref document: EP

Kind code of ref document: A1