WO2006128369A1 - Ethernet access device and method thereof - Google Patents

Ethernet access device and method thereof Download PDF

Info

Publication number
WO2006128369A1
WO2006128369A1 PCT/CN2006/001138 CN2006001138W WO2006128369A1 WO 2006128369 A1 WO2006128369 A1 WO 2006128369A1 CN 2006001138 W CN2006001138 W CN 2006001138W WO 2006128369 A1 WO2006128369 A1 WO 2006128369A1
Authority
WO
WIPO (PCT)
Prior art keywords
physical layer
multiplexing
unit
module
downlink
Prior art date
Application number
PCT/CN2006/001138
Other languages
English (en)
French (fr)
Inventor
Yang Yu
Xiang Lin
Original Assignee
Hangzhou H3C 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 Hangzhou H3C Technologies Co., Ltd. filed Critical Hangzhou H3C Technologies Co., Ltd.
Priority to US11/921,234 priority Critical patent/US7940812B2/en
Priority to JP2008513897A priority patent/JP4819883B2/ja
Priority to EP20060742026 priority patent/EP1887722A4/en
Publication of WO2006128369A1 publication Critical patent/WO2006128369A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04J—MULTIPLEX COMMUNICATION
    • H04J3/00—Time-division multiplex systems
    • H04J3/02—Details
    • H04J3/04—Distributors combined with modulators or demodulators
    • H04J3/047—Distributors with transistors or integrated circuits
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04J—MULTIPLEX COMMUNICATION
    • H04J3/00—Time-division multiplex systems
    • H04J3/24—Time-division multiplex systems in which the allocation is indicated by an address the different channels being transmitted sequentially
    • H04J3/247—ATM or packet multiplexing

Definitions

  • the present invention relates to Ethernet access technologies, and in particular, to an Ethernet access device and an access method. Background technique
  • ADSL Asymmetric Digital Subscriber Loop
  • VDSL Very High Bit-Rate Digital Subscriber Loop
  • ADSL2+ ITU standard G.992.5
  • FIG. 1 a typical structure of an Ethernet access system for accessing a broadband terminal user is as shown in FIG. 1.
  • the Layer 2 switch 810 is used as an access device, and the Layer 3 switch 820 is used as a gateway.
  • the Layer 2 switch 810 accesses the gateway 820.
  • the Layer 2 switch 810 performs Layer 2 forwarding between the user and the gateway 820 to complete Layer 2 isolation between the access terminals.
  • Other operations such as QoS (Quality of Service) can be performed at the gateway 820.
  • QoS Quality of Service
  • the average bandwidth per user is relatively low, for example, 10M (megabytes) of bandwidth can satisfy more than 80% of broadband access applications.
  • the bandwidth of Ethernet products such as MAC (Media Access Control) chips and long-distance transmission media has been developed to meet the 1G (Gigabit) and 10G of enterprise network applications, and its price has considerable advantages over bandwidth.
  • the direct application of these products to broadband access will cause huge waste of bandwidth.
  • the cost of the equipment on the user side is still relatively high. It is difficult for the service provider to deploy the network according to this scheme, and it is difficult to reduce the average to each user. Into the device Equal cost.
  • the service provider needs to arrange a large number of Layer 2 switches on the closest user side, and in order to ensure that every possible user needs to access the Internet service, the access service can be opened immediately.
  • Layer switches typically have access capacity of dozens of ports, which may be comparable to the number of nearby homes, or even larger than the number of homes.
  • the actual access rate is very low.
  • the cost of a Layer 2 Ethernet switch is shared by several users. Obviously, the cost is too high. This is the mature Ethernet.
  • IBM Corporation has innovated to utilize the multiplexing technology to achieve the purpose of the tubular Ethernet switch to a certain extent.
  • the solution cyclically collects input signals of multiple low-speed physical ports of the physical layer chip in a certain period of time, multiplexes them into one high-speed digital signal, and transmits them to the MAC chip for Layer 2 processing, and outputs digital signals to the MAC chip.
  • the process is reversed. This reduces the number of MAC chips in the Ethernet switch.
  • a 100M/GE MAC chip is used instead of multiple 10M/100M MAC chips, and the overall cost is reduced accordingly.
  • Ethernet multiplexing technology disclosed in the prior art and the traditional Ethernet-related multiplexing technology is that it utilizes multiplexing technology to solve the internalization problem of the Ethernet switch, instead of like other ethers.
  • the problem of the physical link cost of the Ethernet transmission is solved by using the multiplexing, or the problem that the 10GE physical channel cannot be directly discharged due to the device not supporting the 10GE data processing capability, for example, multiplexing 10 GE channels through the multiplexing device A 10GE channel is transmitted, and the corresponding GE channel is restored after demultiplexing at the other end.
  • the invention provides an Ethernet access device and an access method, which have lower cost and a simplified device architecture.
  • the Ethernet access device is located between a plurality of access nodes and a gateway device, and includes at least two downlink physical port units, a multiplexing demultiplexing unit, and an uplink physical port unit, where:
  • the downlink physical port unit is configured to convert between a low-speed physical layer signal of the input/output access device and a low-speed physical layer load carried therein;
  • the uplink physical port unit is configured to convert between a high speed physical layer signal of the input and output access device and a high speed physical layer load carried therein;
  • the multiplexing demultiplexing unit is configured to multiplex and demultiplex between the low-speed physical layer payload of the downlink physical port unit and the high-speed physical layer payload of the uplink physical port unit.
  • the multiplexing and demultiplexing of the multiplexing demultiplexing unit is performed according to a label corresponding to a downlink physical port unit in the physical layer payload.
  • the identifier is a field in a data frame in a physical layer payload, and the field has a corresponding relationship with a downlink physical port unit;
  • the multiplexing demultiplexing unit includes a tag storage module and a frame multiplexing module, wherein: the tag storage module is configured to store the tag and its corresponding downlink physical port unit; and the frame multiplexing module is configured to input the downlink physical port unit into the low speed
  • the physical layer load is converted into a high-speed physical layer load output to the uplink physical port unit in units of data frames, and the high-speed physical layer load input from the uplink physical port unit is converted into a low-speed physical layer load in units of data frames and outputted thereto.
  • the corresponding downlink physical port unit is configured to store the tag and its corresponding downlink physical port unit; and the frame multiplexing module is configured to input the downlink physical port unit into the low speed
  • the physical layer load is converted into a high-speed physical layer load output to the uplink physical port unit in units of data frames, and the high-speed physical layer load input from the uplink physical port unit is converted into a low-speed physical layer load in units of data frames and outputted thereto.
  • the multiplexing demultiplexing unit includes a tag storage module, a tag module, and a code stream multiplexing module, where:
  • the tag storage module is configured to store the tag and its corresponding downlink physical port unit;
  • the tag module is configured to add a tag corresponding to the downlink physical port unit to the low-speed physical layer load input from the downlink physical port unit, and output the code to the stream multiplexing.
  • the code stream multiplexing module is used to convert between the low speed physical layer load of the tag module and the high speed physical layer load of the upstream physical port unit.
  • the tagging module adds and removes tags in the physical layer payload in units of data frames or in units of fixed length code streams.
  • the multiplexing and demultiplexing of the multiplexing demultiplexing unit is performed in a timing cycle, and the time period includes a time slot corresponding to each downlink physical port unit.
  • the time slot length corresponding to each downlink physical port unit in the timing cycle is the same;
  • the multiplex demultiplexing unit includes a timing storage module and a fixed length multiplexing module, where: the sequential storage module is configured to store a downlink physical port unit corresponding to the time slot in the sequence period;
  • the fixed length multiplexing module converts the low-speed physical layer load received from the corresponding downlink physical port unit into a high-speed physical layer load output to the uplink physical port unit and the high-speed physical layer load received from the uplink physical port unit in each time slot.
  • the conversion to the low-speed physical layer load is output to the corresponding downstream physical port unit in each time slot.
  • the multiplexing demultiplexing unit includes a timing and duration storage module and a variable length multiplexing module, where:
  • the timing and duration storage module is configured to store the length of each time slot in the time series, and the downlink physical port unit corresponding to the time slot;
  • the variable length multiplexing module is configured to convert the low-speed physical layer load that matches the length of the code stream received from the corresponding downlink physical port unit to the length of the time slot into a high-speed physical layer load and output to the uplink physical port unit in each time slot. And converting the high-speed physical layer load received from the uplink physical port unit to the low-speed physical layer load, and outputting the low-speed physical layer load whose code stream length matches the length of the time slot to the corresponding downlink physical port unit in each time slot.
  • an Ethernet access device is located between a plurality of access nodes and a gateway device, and includes a plurality of lower-level downlink physical port units, a top-level uplink physical port unit, and at least two-stage multiplexing units.
  • the uplink interface of the lower multiplexing unit is connected to the downlink interface of the upper multiplexing unit, where:
  • the lower-level downlink physical port unit is connected to the downlink interface of the lower-level multiplex unit, and is used for converting between the low-speed physical layer signal of the input/output access device and the low-speed physical layer load carried therein;
  • the upper-level uplink physical port unit is connected to the uplink interface of the upper-level multiplex unit for converting between the high-speed physical layer signal of the input/output access device and the high-speed physical layer load carried therein;
  • the multiplexing units at each level perform step-by-step multiplexing and demultiplexing between the low-speed physical layer payload of the downlink interface of the lowest-level multiplexing unit and the high-speed physical layer payload of the uplink interface of the upper-level multiplexing unit.
  • multiplexing and demultiplexing of the multiplexing units of the levels are performed according to a label corresponding to a downlink interface of the lowest level multiplexing unit in the physical layer payload.
  • the mark is a field in a data frame in a physical layer payload of a downlink interface of a lower-level multiplexing unit that is different from other downlink interfaces;
  • Each of the multiplexing units includes a tag storage module and a frame multiplexing module, where: the tag storage module is configured to store the tag and its corresponding downlink interface of the multiplexing unit; and the frame multiplexing module is configured to input the downlink interface
  • the physical layer load is converted into a physical layer load in units of data frames and output by the uplink interface, and a physical layer load input from the uplink interface is converted into a multi-path physical layer load in units of data frames and marked by the corresponding one.
  • Downstream interface output is configured to input the physical layer load.
  • the lower-level multiplexing unit includes a tag storage module, a tag module, and a code stream multiplexing module, where:
  • the tag storage module is configured to store the tag and its corresponding downlink interface of the multiplexing unit; the tag module is configured to add a tag corresponding to the downlink interface to the physical layer payload input from the downlink interface, and output the tag to the code stream multiplexing module, and Removing the mark from the physical layer load input from the code stream multiplexing module and outputting from the downlink interface corresponding to the mark;
  • the code stream multiplexing module is used for marking multiple physical layer loads of the module and the uplink interface. Converting between physical layer loads;
  • the other level multiplexing unit includes a label storage module and a code stream upper multiplexing module, wherein the code stream upper multiplexing module is configured to convert the multiple physical layer loads input from the downlink interface into one physical layer load and output from the uplink interface. And converting one physical layer load input from the uplink interface into multiple physical layer loads and outputting from the downlink interface corresponding to the label.
  • the marking module of the lowermost multiplexing unit adds and removes tags in the physical layer payload in units of data frames or in units of fixed length code streams.
  • the marking includes an offset flag of each multiplexing unit
  • the multiplexing units of the levels include an offset tag storage module, an offset tag module, and a code stream multiplexing module, where:
  • the offset tag storage module is configured to store the offset tag of the multiplexing unit and its corresponding downlink interface of the multiplexing unit;
  • the offset marking module is configured to add an offset flag corresponding to the downlink interface to the physical layer payload input from the downlink interface, output to the code stream multiplexing module, and remove the physical layer load input from the code stream multiplexing module.
  • An offset flag of the multiplexing unit is output from a downlink interface corresponding to the offset flag;
  • the code stream multiplexing module is used to convert between the multiple physical layer loads of the offset tag module and one physical layer load of the uplink interface.
  • the offset marking module of the multiplexing unit of each level adds and removes the offset flag in the physical layer payload in units of data frames or in units of fixed length code streams.
  • multiplexing and demultiplexing of the multiplexing units of the levels are performed according to a timing cycle of the multiplexing unit, and the time period includes a time slot corresponding to a downlink interface of the multiplexing unit.
  • the time slots in the timing cycle of each multiplexing unit have the same length
  • Each of the multiplexing units includes a timing storage module and a fixed length multiplexing module, wherein: the timing storage module is configured to store a downlink interface corresponding to the time slot in the timing period of the multiplexing unit;
  • the multiplexed multiplex module receives the multiplexed physics from the corresponding downlink interface in each time slot.
  • the layer load is converted into a physical layer load and then output from the uplink interface, and a physical layer load received from the uplink interface is converted into multiple physical layer loads and then outputted by the corresponding downlink interface in each time slot.
  • the multiplexing units of the levels include a timing and duration storage module and a variable length multiplexing module, where:
  • the timing and duration storage module is configured to store the length of the time slot in the timing period of the multiplexing unit and the downlink interface corresponding to the time slot;
  • the variable length multiplexing module is configured to convert the physical layer load that matches the length of the code stream received from the corresponding downlink interface to the length of the time slot into a physical layer load in each time slot, and output from the uplink interface, and receive from the uplink interface. After the rate conversion of the physical layer load of one path, the physical layer load matching the length of the code stream to the length of the time slot in each time slot is output from the corresponding downlink interface.
  • the access device further includes a downlink physical port unit of each level except the lowest level, and each downlink interface of the multiplex unit is respectively connected to perform conversion between the physical layer signal and the physical layer load;
  • Each multiplex unit and its connected downlink physical port unit are encapsulated in a physical layer PHY-MAC interface multiplexing chip;
  • the access device further includes an uplink physical port unit of each level except the uppermost level, and is connected between the lower-level PHY-MAC interface multiplexing chip and the upper-level PHY-MAC interface multiplexing chip, and is used to perform physical layer signal and physical layer. Conversion between loads.
  • the access device further includes downlink physical port units of all levels except the lowest level, and each downlink interface of the multiplexing unit is respectively connected to perform conversion between physical layer signals and physical layer loads. ;
  • the access device further includes an uplink physical port unit of each level except the uppermost level, and each uplink interface of the multiplex unit is connected to perform conversion between the physical layer signal and the physical layer load;
  • Each of the multiplexing unit and the uplink physical port unit and the uplink physical port unit connected thereto are encapsulated in one PHY interface multiplexing chip, and the uplink physical port unit of the lower PHY interface multiplexing chip is connected to the downstream physical port of the upper PHY interface multiplexing chip. unit.
  • an Ethernet access method includes the following steps: multiplexing physical layer loads carried in physical layer signals of at least two access nodes into One physical layer load;
  • the demultiplexed physical layer load is converted to a low-speed physical layer signal and then downlinked.
  • the multiplexing the physical layer load of the access node into one physical layer load is specifically: multiplexing the multiple physical layer loads of the same group of access nodes into one physical layer load of each group; The physical layer payloads are again grouped and then multiplexed or directly multiplexed until they are multiplexed into one physical layer load;
  • Demultiplexing the one-way multiplexed physical layer load into a physical layer load corresponding to the access node is specifically: demultiplexing the multiplexed physical layer load into multiple physical layer loads; and demultiplexing the multiplexed physical physics The layer payloads are again demultiplexed separately until demultiplexed into physical layer loads corresponding to the access nodes.
  • the method before the multiplexing, further includes: adding, in the physical layer payload before multiplexing, each physical layer load that is multiplexed at the current time, a corresponding offset flag;
  • Each of the demultiplexing is performed according to an offset flag having a corresponding relationship between the demultiplexed physical layer load and the demultiplexed physical layer load;
  • the method further includes, at each demultiplexing, removing an offset flag according to the physical layer load for performing the current demultiplexing.
  • the method before performing physical layer load multiplexing, the method further includes: adding a label corresponding to the access node in a physical layer load of the access node;
  • the method further includes: removing the flag in a physical layer payload corresponding to the access node before converting the physical layer payload corresponding to the access node to the low-speed physical layer signal.
  • the multiplexing the multiple physical layer loads into one physical layer load is cyclically performed according to a timing cycle; the timing cycle includes a time slot corresponding to each physical layer load before multiplexing, and each time slot will be The corresponding pre-multiplexed physical layer load output is a multiplexed physical layer load, and the length of the pre-multiplexed physical layer load matches the length of the time slot; Demultiplexing a physical layer load into multiple physical layer loads according to a timing cycle; the timing cycle includes a time slot corresponding to each physical layer load after demultiplexing, and is to be solved in each time slot.
  • the physical layer load matching the length of the preamble with the length of the preamble is the corresponding demultiplexed physical layer payload.
  • the demultiplexing a physical layer load into multiple physical layer loads is performed according to a field in the data frame of the physical layer load corresponding to the access node.
  • an Ethernet access switching device is disposed between an access device and a gateway device; and includes an access side physical port unit, a multiplexing conversion unit, and a gateway side physical port unit;
  • the access side physical port unit is configured to convert between a physical layer signal transmitted between the access switching device and the access device and a multiplexed physical layer load carried therein;
  • the gateway side physical port unit is used to Converting between a physical layer signal transmitted between the access switching device and the gateway device and a physical layer load carried in units of data frames carried therein;
  • the multiplexing conversion unit is configured to convert between a multiplexed physical layer load of the access side physical port unit and a physical layer load of the gateway side physical interface unit in units of data frames.
  • the multiplexing conversion unit includes a multiplexing labeling module and a multiplexing label processing module;
  • the multiplexing labeling module is configured to store a correspondence between the label and the identifier of the access node; the multiplexing label processing module is configured to extract a label in a physical layer load input by the physical port unit of the access side, and generate a data frame by using the label For the physical layer load of the unit, the physical layer load in the data frame unit after the clear flag is cleared is output to the physical port unit on the gateway side; the physical layer load in the data frame unit received from the physical port unit on the gateway side is complex The tag corresponding to the access node identifier in the data frame is searched by the tag module, and the tag is added to the physical layer payload and output to the access side physical port unit.
  • the multiplexing label processing module obtains a physical layer load in units of data frames after clearing the label;
  • the multiplexing tag processing module regroups physical layer loads having the same tag Then, clear the mark in it to get the physical layer load in data frame.
  • the multiplexing conversion unit includes a multiplexing timing module, a timing conversion module, and a data frame identification module, where
  • the multiplexing timing module is configured to store a correspondence between a time slot and an access node identifier in a timing cycle
  • the timing conversion module is configured to write the multiplexed physical layer load segment received from the access side physical port unit into the uplink buffer of the time slot according to the time slot arrangement order in the multiplexed sequence period. And outputting a physical layer load of a certain length from the downlink buffer area of each time slot to the access side physical port unit according to the time slot arrangement order in units of time slots in the multiplexing time period;
  • the data frame identification module is configured to extract physical layer payloads in units of data frames from uplink buffer areas of each time slot, and output physical layer loads in units of data frames to physical port units on the gateway side; physical port units on the side of the gateway side The received physical layer payload in units of the data frame is written to the downlink buffer of the time slot.
  • the multiplexing conversion unit includes a multiplexing timing and duration module, a timing and duration conversion module, and a data frame identification module;
  • the multiplexing timing and duration module is configured to store the length and arrangement order of each time slot in the multiplexing timing cycle, and the correspondence between the time slot and the access node identifier;
  • the timing and duration conversion module is configured to, in units of time slots in the multiplexing timing period, correspond to the time length of the code stream in the multiplexed physical layer payload received from the physical port unit of the access side according to the order of time slots.
  • the code stream segment of the slot length is written into the uplink buffer area of the time slot; in units of time slots in the multiplexing timing cycle, from the downlink buffer area of each time slot to the access side physical port unit in the order of time slot arrangement
  • the output code stream length corresponds to the physical layer load of the length of the time slot;
  • the data frame identification module is configured to extract a physical layer load in units of data frames from the uplink buffer area of each time slot, and output the physical layer load in units of data frames to the physical port unit on the gateway side; The physical layer payload received by the port unit and sent to the access node in units of data frames is written into the downlink buffer of the slot.
  • a multiplexing gateway device includes a physical port unit, a multiplexing interface unit, and a forwarding unit, where: The physical port unit is configured to convert between a physical layer signal of the input/output multiplexing gateway device and a multiplexed MAC layer code stream carried therein;
  • the multiplexing interface unit is configured to convert the multiplexed MAC layer code stream received from the physical port unit into a data frame and output to the forwarding unit, and inversely convert the data frame received from the forwarding unit into a multiplexed MAC layer code stream and output to the Physical port unit;
  • the forwarding unit is used for data frame forwarding.
  • the present invention multiplexes the load of the multi-path physical layer signal, that is, the physical layer load carried therein, and transmits the multiplexed physical layer load through the physical layer, so that the present invention can be compared with the prior art.
  • the cost of the Ethernet access device is reduced to a greater extent, the unnecessary MAC layer and the upper layer function chip are removed, the architecture of the access device is simplified, and the functions necessary for the access device between the gateway and the user are provided; At the same time, the reliability of the access device is correspondingly improved as the access device is simplified.
  • FIG. 1 is a diagram showing an example of the structure of an Ethernet access system in the prior art
  • FIG. 2 is a block diagram showing the structure of a first type of Ethernet access system according to an embodiment of the present invention
  • Figure 3 is a block diagram showing the structure of a second type of Ethernet access system according to an embodiment of the present invention.
  • FIG. 4 is a schematic diagram showing the structure of a third type of Ethernet access system according to an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of Embodiment 1 of an access device according to the present invention.
  • Embodiment 2 is a schematic structural diagram of Embodiment 2 of an access device according to the present invention.
  • Embodiment 3 of an access device according to the present invention is a schematic structural diagram of Embodiment 3 of an access device according to the present invention.
  • Embodiment 4 of an access device according to the present invention is a schematic structural diagram of Embodiment 4 of an access device according to the present invention.
  • FIG. 9 is a diagram showing an example of a logical structure of an access device employing cascade multiplexing according to an embodiment of the present invention.
  • FIG. 10 is a schematic structural diagram of a fifth multiplexing unit of an access device according to an embodiment of the present invention
  • FIG. 11 is a schematic structural diagram of a lower-level multiplexing unit of an access device according to Embodiment 6 of the present invention
  • FIG. 12 is a schematic structural diagram of an upper multiplexing unit of Embodiment 6 of an access device according to the present invention.
  • FIG. 13 is a schematic structural diagram of an multiplex unit of an access device according to an embodiment of the present invention
  • FIG. 14 is a schematic structural diagram of an multiplex unit of an access device according to an embodiment of the present invention
  • FIG. 16 is a schematic diagram of a logical structure and a connection manner of a multiplexing chip according to an embodiment of the present invention
  • FIG. 17 is a schematic structural diagram of Embodiment 1 of an access switching apparatus according to the present invention
  • FIG. 18 is a schematic structural diagram of Embodiment 2 of an access switching apparatus according to the present invention
  • FIG. 20 is a flowchart of an Ethernet uplink access method according to an embodiment of the present invention
  • FIG. 21 is a flowchart of an Ethernet downlink access method according to an embodiment of the present invention.
  • the access nodes in the access network mainly communicate with the external network through the gateway device, and there are few cases where the access nodes directly communicate with each other.
  • the access device between the gateway device and the access node can transmit the signal sent by the access node to the gateway device and downlink the signal sent by the gateway device to the destination access node as long as it can provide the uplink and downlink transmission function.
  • the function of the access network can be satisfied, and the function of forwarding packets between the downlink ports in the Layer 2 switch is basically idle in the access network.
  • the forwarding function can be completed by the gateway device even if there is a case where a small number of access nodes require direct communication.
  • the scheme for implementing the forwarding function only on the gateway device is a scheme suitable for the access system, and has a relatively low Access cost.
  • the Ethernet access system may have the structure shown in FIG. 2, FIG. 3 or FIG. 4, wherein the access device 920 converts the uplink physical layer load and converts it.
  • the physical layer signal is transmitted to the gateway side, and the downlink signal of the gateway side is transmitted to the destination access node 910.
  • the access device 920 is directly connected to the gateway device 901. This system structure does not need to be modified for the existing gateway.
  • the second type shown in FIG. In the Ethernet access system the access device 920 is connected to the multiplexing gateway device 902, and the multiplexing gateway device 902 and the access device 920 use the corresponding multiplexing technology to achieve access; the third type shown in FIG. In the Ethernet access system, the access switching device 930 is added between the access device 920 and the gateway device 901 as an interface between the access device 920 and the gateway device 901, so that the multiplexing technology of the access device 920 is applied to the gateway device 901. Transparent.
  • FIG. 8 are schematic structural diagrams of Embodiments 1 to 4 of the access device according to the present invention.
  • the downlink physical port units 111, 112 to Ilk are respectively connected to the multiplexing demultiplexing unit 120, and the uplink physical port unit 130 is also complex. Connected by the demultiplexing unit 120. It should be noted that although k physical port units are illustrated, the above four embodiments support two or more downlink physical port units.
  • the downlink physical port units 111, 112 to Ilk may each be connected to an access node, and respectively parse the low-speed physical layer load carried by the low-speed physical layer signals of the input access device. And outputting to the multiplexing demultiplexing unit 120; the multiplexing demultiplexing unit 120 multiplexes each low-speed physical layer load into one high-speed physical layer load, and outputs it to the uplink physical port unit 130; the uplink physical port unit 130 performs high-speed physical The layer load carries the access device in the high speed physical layer signal.
  • the uplink physical port unit 130 After receiving the high-speed physical layer signal, the uplink physical port unit 130 outputs the high-speed physical layer payload that is parsed out to the multiplexing demultiplexing unit 120; the multiplexing and demultiplexing unit 120 demultiplexes the high-speed physical layer payload to correspond to The low-speed physical layer load of each downlink physical port unit, and the low-speed physical layer load of each downlink is output to the corresponding downlink physical port unit, and is converted into a low-speed physical layer signal by the downlink physical port unit receiving the low-speed physical layer load, and then outputted Into the device.
  • Embodiment 1 to Embodiment 4 The difference from Embodiment 1 to Embodiment 4 is that the internal implementation of the multiplexing demultiplexing unit 120 is different due to the different multiplexing techniques employed.
  • the first embodiment and the second embodiment are multiplexed according to the traffic that arrives at the downlink physical port unit, and are demultiplexed by the label corresponding to the downlink physical port unit in the physical layer load.
  • the third embodiment and the fourth embodiment are cycled according to the timing cycle.
  • the timing period is decomposed into time slots corresponding to the downlink physical port units, and each time slot is dedicated to the corresponding downlink physical port unit, even if the corresponding downlink physical port unit does not have network traffic at the time.
  • the idle signal is not used to multiplex the physical layer payloads of other downstream physical port units.
  • the idle signal can be an idle signal conforming to the Ethernet standard or an idle signal agreed by both parties.
  • the multiplexing demultiplexing unit 120 includes a tag storage module 121 and a frame multiplexing module 122, and the frame multiplexing module 122 and each downlink physical port unit and tag respectively.
  • the storage module 121 and the uplink physical port unit 130 are connected.
  • a field having a corresponding relationship with the access node in the data frame is used as a flag, and each access node is connected to the access device through a downlink physical port unit, and therefore transmitted through each downlink physical port unit.
  • the mark is different in the physical layer load.
  • the data frame in the present invention refers to the MAC layer data frame of the Ethernet standard, which is included in the physical layer payload.
  • the tag may be a MAC address of an access node in the data frame, an IP (Internet Protocol) address, and the like.
  • a downlink physical port unit converts the received physical layer signal into a physical layer payload and sends the signal to the frame multiplexing module 122.
  • the frame multiplexing module 122 checks whether the label in the data frame and its corresponding downlink interface have been saved in the tag storage module. 121, if not saved or changed, the tag storage module 121 is updated, so that the corresponding relationship between the current tag and the downlink physical port unit is maintained in the tag storage module 121, and then converted into a high-speed physical layer load and output to the uplink physical port unit. 130;
  • the uplink physical port unit 130 transmits the high-speed physical layer payload to the access device in a high-speed physical signal.
  • the uplink physical port unit 130 converts the received high-speed physical layer signal into a physical layer payload and sends it to the frame multiplexing module 122.
  • the frame multiplexing module 122 searches the tag storage module 121 for the tag in the data frame in the high-speed physical layer payload.
  • the downlink physical port unit converts the high-speed physical layer load into a low-speed physical layer load and outputs it to the downlink physical end.
  • the port unit: the downlink physical port unit carries the low-speed physical layer payload in the low-speed physical layer signal and sends out the access device.
  • the frame multiplexing module 122 has a function of parsing data frames from the physical layer payload and a function of marking learning, and these two functions can be implemented by existing methods in the existing MAC chip.
  • a buffer area may be opened for each downlink physical port unit and the uplink physical port unit 130 in the frame multiplexing module 122, and the high-low speed clock is included to include the complete data.
  • the transmission rate conversion is performed in units of the physical layer load of the frame.
  • the physical layer signal of the input/output access device carries the physical layer load including the complete data frame, conforms to the Ethernet standard, and can be directly connected with the standard gateway device, and is applicable to FIG. 2
  • this multiplexing method requires the access device to parse the data frame, and this function needs to be implemented at the MAC layer, so the access device needs to embed some simple Layer 2 processing capabilities.
  • FIG. 6 is a schematic structural diagram of Embodiment 2 of an access device according to the present invention.
  • the multiplexing demultiplexing unit 120 includes a tag storage module 121, a tag module 123, and a code stream multiplexing module 124.
  • the tag module 123 is respectively connected to the tag storage module 121 and the code.
  • the stream multiplexing module 124 and the respective downlink physical port units, the code stream multiplexing module 124 are connected to the uplink physical port unit 130.
  • the tag storage module 121 stores the tag and the downlink physical port unit corresponding to the tag, and the tag and its correspondence with the downlink physical port unit can be set in advance in the tag storage module 121.
  • a downlink physical port unit converts the received low-speed physical layer signal into a low-speed physical layer payload and sends it to the marking module 123.
  • the marking module 123 finds a label corresponding to the downlink physical port unit in the label storage module 121, at a low-speed physical layer.
  • the tag is added to the payload and output to the code stream multiplexing module 124.
  • the code stream multiplexing module 124 converts the tagged low-speed physical layer payload into a high-speed physical layer payload and outputs it to the uplink physical port unit 130;
  • the unit 130 converts the high speed physical layer payload into a high speed physical layer signal and transmits the access device.
  • the uplink physical port unit 130 converts the received high-speed physical layer signal into a physical layer load output to the code stream multiplexing module 124; the code stream multiplexing module 124 converts the high-speed physical layer load into a low-speed physical layer load and outputs the result to the marking module 123; Marker module 123 is marked
  • the downlink physical port unit corresponding to the mark in the low-speed physical layer load is found in the storage module 121, and is output to the downlink physical port unit after clearing the mark in the low-speed physical layer load; the downlink physical port unit carries the low-speed physical layer load
  • the access device is transmitted in the low speed physical layer signal.
  • the marking module 123 may add or clear the mark in the physical layer load in units of data frames, or may be performed at a fixed length smaller than the length of the data frame.
  • a fixed length it is considered that the data frames from or to the same downlink physical port unit may be separated by a certain time, in order to provide good QoS (Quality of Service) performance, even when the end of the data frame is reached, even if it has not been reached.
  • Fixed length can also be multiplexed.
  • the marker can be added to the set position in the physical layer load, and the marker can also be found and cleared from the set position.
  • each low-speed port is used to multiplex one frame in turn, but in order to achieve differentiated QoS to some extent, the order of multiplexing can be changed. For example, for three ports A, B, and C, A and B are multiplexed two frames in turn, and then one frame of C is multiplexed, and the same method can be used for the fixed length multiplexing, and the improvement effect is more obvious.
  • the high and low speed physical layer load translation function of the code stream multiplexing module 124 can be implemented by referring to the frame multiplexing module 122.
  • the VLAN number that conforms to the Ethernet standard As a flag, so that the physical layer load can be directly processed on the gateway side, and the first type of access shown in FIG. 2 can be used. system.
  • the access device since the VLAN number is configured on the access device and is independent of the access node, the access device repeatedly learns the tag due to the change of the MAC address or IP address of the access node, thereby ensuring the stability of the access device. And with the maturity of multi-layer VLAN technology, the number of VLANs will not become a problem, and the two-layer VLAN can support 4094*4094 nodes.
  • the other scheme in this embodiment is a physical layer load carried in the physical layer signal input and output from the uplink physical port 130 and does not include a complete Ethernet data frame, and thus is applicable to the second type of FIG. 3 or the third type of FIG. Access to the system.
  • the schematic diagram of the third embodiment of the access device of the present invention is shown in FIG. 7.
  • the multiplexing and demultiplexing unit 120 includes a timing storage module 125 and a fixed length multiplexing module 126, and the fixed length multiplexing module 126 and each downlink physical port unit respectively.
  • the timing storage module 125 and the uplink physical port unit 130 are connected.
  • each of the time slots in the timing cycle of this embodiment has the same length.
  • the downlink physical port unit corresponding to each of the time slots sequentially arranged in the timing cycle is stored.
  • each physical port unit may be corresponding to the same number of time slots; and for the downlink physical port unit with different bandwidth, each downlink physical port unit may be corresponding to its bandwidth. Match the number of time slots.
  • a downlink physical port unit converts the received low speed physical layer signal into a low speed physical layer payload for transmission to the fixed length multiplexing module 126.
  • the fixed length multiplexing module 126 writes the low-speed physical layer load input by each channel into the buffer area corresponding to the downlink physical port unit, and sequentially selects the downlink physical port unit corresponding to the time slot in units of time slots in the sequence period.
  • the buffer area outputs a certain length of physical layer load with a high-speed clock; for a downlink physical port unit that does not currently have network traffic, the corresponding time slot is filled with an idle standard signal of the Ethernet standard or an idle signal agreed by the communication party.
  • the fixed length multiplexing module 126 multiplexes the low speed physical layer payload into a high speed physical layer payload, including the code stream segments corresponding to the respective downlink physical port units.
  • the multiplexed high-speed physical layer payload is output to the upstream physical port unit 130, which converts the high-speed physical layer payload into a high-speed physical layer signal and transmits the access device.
  • the uplink physical port unit 130 converts the received high speed physical layer signal into a physical layer payload output to the fixed length multiplexing module 126.
  • the fixed length multiplexing module 126 writes the high speed physical layer load into the cache area, and sequentially outputs a certain length from the cache area to the downlink physical port unit corresponding to the time slot in a sequence of time slots in the time series. Physical layer load.
  • the fixed length multiplexing module 126 demultiplexes one high speed physical layer load into multiple low speed physical layer loads.
  • Downstream physical port unit will The received low-speed physical layer load bearer is transmitted out of the access device in the low-speed physical layer signal.
  • FIG. 8 The schematic diagram of the fourth embodiment of the access device of the present invention is shown in FIG. 8.
  • the multiplexing and demultiplexing unit 120 includes a timing and duration storage module 127 and a variable length multiplexing module 128, and the variable length multiplexing module 128 and each downlink physical medium respectively.
  • the port unit, the timing and duration storage module 127, and the uplink physical port unit 130 are connected.
  • the node users may adopt different access rates.
  • This embodiment is more suitable for the case where the various downlink physical port units may have different bandwidths.
  • the difference between this embodiment and the third embodiment is that the length of each slot in the timing period is introduced.
  • the sequence and duration module 127 in addition to storing the downlink physical port units corresponding to the respective time slots arranged in the sequence, the length of the time slot is also stored, and the time slot length matches the bandwidth of the downlink physical port unit.
  • a downlink physical port unit converts the received low-speed physical layer signal into a low-speed physical layer payload and sends it to the variable length multiplexing module 128.
  • the variable length multiplexing module 128 writes the low-speed physical layer load of each input into the buffer area corresponding to the downlink physical port unit, and simultaneously buffers the downlink physical port unit corresponding to each time slot according to the time slot sequence in the time series period.
  • the area outputs a physical layer load of a certain length with a high-speed clock, and the length of the physical layer payload corresponds to the length of the time slot; for the downlink physical port unit that does not currently have network traffic, the idle standard signal of the Ethernet standard or the idled by the communication party The signal fills its corresponding time slot.
  • variable length multiplexing module 128 multiplexes the low-speed physical layer payload into a high-speed physical layer payload, including codestream segments that may have different lengths corresponding to the respective downlink physical port units.
  • the multiplexed high-speed physical layer payload is output to the upstream physical port unit 130, which converts the high-speed physical layer payload into a high-speed physical layer signal and transmits the access device.
  • the uplink physical port unit 130 converts the received high speed physical layer signal into a physical layer payload output to the variable length multiplexing module 128.
  • the variable length multiplexing module 128 writes the high-speed physical layer load into the cache area, and outputs a certain length from the cache area to the downlink physical port unit corresponding to the time slot according to the time slot sequence in the time series period. Physical layer load, the length of the output physical layer payload corresponds to the length of the time slot.
  • the variable length multiplexing module 128 demultiplexes one high speed physical layer load into multiple low speed physical layer loads.
  • the downlink physical port unit will receive the low-speed physical layer load bearer
  • the access device is transmitted in the low speed physical layer signal.
  • the time slots in the third embodiment and the fourth embodiment can also be regarded as a measure of the physical layer payload length, for example, the number of bytes of the physical layer payload.
  • the multiplexing demultiplexing unit 120 does not need to know the content of the physical layer payload, and does not need to identify or parse the data frame therein.
  • the uplink bandwidth is shared by all the downlink connections, and the bandwidth allocated to each downlink connection in the uplink bandwidth in the two embodiments is strictly reserved, and is guaranteed from the physical layer. QoS performance.
  • these two embodiments have higher requirements for physical layer transmission from the upstream physical port 130 to the gateway device side to avoid data retransmission that may occur in the event of a code stream misalignment.
  • the physical layer load carried in the physical layer signal input and output from the uplink physical port 130 in Embodiment 3 and Embodiment 4 does not include the physical layer load of a complete Ethernet data frame, and thus is applicable to the second type of FIG. 3 or the third form of FIG. Access system.
  • each physical port on a PHY (physical layer) chip has the same bandwidth, and rarely provides more than 8 physical ports on one PHY chip, even if these ports are multiplexed, the bandwidth is insufficient.
  • 100M, long-distance transmission in units of 100M still cannot meet the low-cost requirements of broadband access.
  • the access device can adopt a cascade multiplexing structure, and an example of the logical structure thereof is as shown in FIG.
  • the multiplexing unit is divided into multiple multiplexing levels according to different uplink and downlink bandwidths
  • the downlink interface of the lower-level multiplexing unit 220 is connected to the lower-level downlink physical port unit 210
  • the uplink interface is connected to the intermediate-level multiplexing unit 230
  • the uplink interface of the intermediate stage multiplexing unit 230 is connected to the downlink interface of the uppermost multiplexing unit 240
  • the uplink interface of the uppermost multiplexing unit 240 is connected to the uppermost uplink physical port unit 250.
  • the lower-level downlink physical port unit 210 converts the low-speed physical layer signal of the access node into a low-speed physical layer load, and the lower-speed physical layer load is converted into the high-speed physical layer load by the multiplexing units 220, 230, and 240 step by step.
  • the upper uplink physical port unit 250 carries the high speed physical layer payload in the physical layer signal for output.
  • the upper-level uplink physical port unit 250 parses the high-speed physical layer payload from the received high-speed physical layer signal, and demultiplexes the high-speed physical layer load into the corresponding lower-level downlink physics by the multiplexing units 240, 230, and 220.
  • the low-speed physical layer load of the port unit 210 is output and output; the lowermost downlink The physical port unit 210 converts the low-speed physical layer payload received from the multiplexing unit 220 into a low-speed physical layer signal and outputs it from the access device.
  • the lowermost downlink physical port unit 210 connected to the same multiplexing unit 220 or different multiplexing unit 220 may have different bandwidths.
  • the access device in the present invention supports two-stage and two-stage multiplexing unit cascade multiplexing.
  • each multiplexing unit can have different implementation methods.
  • the fifth embodiment to the ninth embodiment have the same cascading logic structure, and the difference is that the implementation method of the multiplex unit is different.
  • the fifth embodiment to the ninth embodiment will be described below only for the multiplex unit.
  • Each multiplexing unit includes at least two downlink interfaces and one uplink interface, the downlink interface is used for inputting and outputting at least two physical layer loads, and the uplink interface is used for inputting and outputting the multiplexed physical of the downlink interface multiple physical layer loads. Layer load or physical layer load before demultiplexing.
  • the multiplexing unit in the fifth embodiment and the seventh embodiment are multiplexed according to the traffic reaching the downlink interface, and are demultiplexed by the label corresponding to the downlink interface in the physical layer payload; the eighth embodiment and the ninth embodiment are cyclically performed according to the timing cycle. Multiplexing and demultiplexing, the timing period is decomposed into time slots corresponding to the downlink interface, and each time slot is dedicated to the corresponding downlink interface. Even if the corresponding downlink interface does not have network traffic, only the idle signal is used to fill the physical layer. It is not used to transmit traffic of other downlink interfaces.
  • the idle signal may be an idle signal of the Ethernet standard or an idle signal agreed by both parties.
  • the structure of the multiplexing unit in the fifth embodiment of the access device is as shown in FIG. 10, and the downlink interfaces 311, 312 to 31m are respectively connected to the frame multiplexing module 402, and the frame multiplexing module 402 is further connected to the tag storage module 401 and the uplink interface 320 respectively. connection.
  • the multiplexing unit uses a field in the data frame that has a corresponding relationship with the access node as a label, since each access node is connected through a lower-level downlink physical port unit.
  • the downlink interface of the lowermost multiplex unit so for each multiplex unit, the physical layer payload transmitted through any one of the downlink interfaces is different from the physical layer payload passing through other downlink interfaces.
  • the tag may be a field such as a MAC address, an IP address, and the like of the access node in the data frame.
  • the multiplexing unit of each embodiment may adopt the structure in FIG. 10.
  • the frame multiplexing module 402 checks whether the label in the data frame and its corresponding downlink interface have been Saved in the tag storage module 401, if not After saving or changing, the tag storage module 401 is updated to maintain the correspondence between the current tag and the downlink interface in the tag storage module 401, and then converted into a high-speed physical layer payload to output the multiplexing unit from the uplink interface 320.
  • the frame multiplexing module 402 finds a downlink interface corresponding to the mark in the data frame in the tag storage module 401, and converts one physical layer load into a low-speed physical layer load.
  • the downlink interface output multiplexing unit corresponding to the tag.
  • the frame multiplexing module 402 can be implemented by referring to the frame multiplexing module 122 of the access device embodiment 1.
  • the physical layer load carried in the physical layer signal of the input/output access device in the embodiment is in compliance with the Ethernet standard, and can be directly connected to the standard gateway device, and is applicable to the first type shown in FIG. Ethernet access system.
  • the lowermost multiplex unit has a different structure from the other multiplex units.
  • the structure of the lowermost multiplexing unit is as shown in FIG. 11.
  • the marking module 403 is connected to the downlink interfaces 311, 312 to 31m, the label storage module 401 and the code stream multiplexing module 404, respectively, and the code stream multiplexing module 404 is connected to the uplink interface 320.
  • the structure of the other levels of the multiplexing unit is as shown in FIG. 12, and the code stream upper multiplexing module 405 is connected to the downlink interfaces 311, 312 to 31m, the tag storage module 401, and the uplink interface 320, respectively.
  • a flag corresponding to the lowest physical port unit of the access device is used.
  • the tag storage module 401 of the lower-level multiplexing unit stores a flag corresponding to the lowest-level physical port unit to which the lower-level multiplexing unit is connected, and a downlink corresponding to the lower-level physical port unit of the label and the connection.
  • the mark and correspondence may be set in advance in the mark storage module 401.
  • the marking module 403 finds a tag corresponding to the downlink interface in the tag storage module 401, and adds the tag to the received physical layer payload. And outputting to the code stream multiplexing module 404; the code stream multiplexing module 404 performs downlink-to-uplink transmission rate conversion on the added physical layer load, and outputs the rate-converted physical layer load from the uplink interface 320 to the upper level. Use the unit.
  • the code stream multiplexing module 404 For the physical layer load received from the uplink interface 320, the code stream multiplexing module 404 performs uplink to downlink transmission rate conversion on the received physical layer payload, and outputs the rate converted physical layer payload to the marking module 403; the marking module 403 Tag storage module The downlink interface corresponding to the mark in the physical layer payload after the rate conversion is found in 401, and after the mark in the physical layer load is cleared, it is outputted from the downlink interface to the multiplexing unit.
  • the tag storage module 401 of the other levels stores the tag corresponding to the lowermost physical port unit of the multiplex unit connection and the mapping relationship between the tag and the downlink interface corresponding to the lowermost physical port unit. It can be seen that the multiplexing unit with higher multiplexing level tends to have more tags corresponding to the downlink interface.
  • the mark and correspondence can be set in advance in the mark storage module 401.
  • the physical layer payload has been marked by the lowest multiplex unit, and the code stream upper multiplex module 405 performs downlink to uplink.
  • the transmission rate conversion converts the rate converted physical layer load from the uplink interface 320 to the present multiplexing unit.
  • the code stream upper multiplexing module 405 After receiving the physical layer load from the uplink interface 320, the code stream upper multiplexing module 405 searches the tag storage module 401 for the downlink interface corresponding to the tag, and performs uplink to downlink transmission rate conversion on the received physical layer payload.
  • the rate converted physical layer load is output from the downlink interface to the multiplexing unit.
  • the label added by the lower-level multiplexing unit in the physical layer payload is sent out of the access device through each of the upper-level multiplexing units; and the identifier included in the physical layer payload received by the access device from the gateway device side is used for each
  • the upper multiplexing unit determines from which downlink interface the demultiplexed physical layer payload is output, and the flag is cleared before outputting the lowest level multiplexing unit.
  • the marking module 403 of the lowermost multiplexing unit adds or clears the mark in the physical layer load in units of data frames, or may perform a fixed length smaller than the length of the data frame, but all the lower levels.
  • the multiplexing unit must use the same multiplexing method. When a fixed length is used, in order to provide good QoS performance, when the end of the data frame is reached, multiplexing processing can be performed even if the fixed length has not yet been reached.
  • the marker can be added to the set position in the physical layer load, and the marker can also be found and cleared from the set position.
  • the marking module 403 of the lowermost multiplexing unit in this embodiment generally needs to identify the boundary of the data frame in the physical layer payload to determine the code stream range to which the label is applied.
  • the implementation of the code stream multiplexing module 404 of the lowermost multiplex unit and the other stages of the code stream upper multiplex module 405 in this embodiment can be implemented by referring to the frame multiplexing module 122 in the first embodiment of the access multiplexer.
  • the VLAN number that conforms to the Ethernet standard as a flag, so that the physical layer load can be directly processed on the gateway side, and the first type of access shown in FIG. 2 can be used. system.
  • the VLAN number is configured by the service provider on the access device, and the access device adds and strips the VLAN number, and the access node cannot be perceived, so it cannot be accessed like the MAC address and the IP address.
  • the node is modified to cause the access device to repeatedly learn the mark, thereby ensuring stability.
  • the number of VLANs is not a big problem, and the two-layer VLAN can support the access of 4094*4094 nodes.
  • the physical layer signal input and output from the uppermost uplink physical port is not a standard Ethernet physical layer load, and is applicable.
  • the structure of the multiplexing unit in the seventh embodiment of the access device is as shown in FIG. 13, and the offset marking module 407 is connected to the downlink interfaces 311, 312 to 31m, the offset tag storage module 406, and the code stream multiplexing module 404, respectively.
  • Module 404 is coupled to uplink interface 320.
  • the mark used in this embodiment includes an offset mark of each multiplex unit, and the same level offset mark has the same position and the same bit length in the physical layer load, is used by the multiplex unit, and is multiplexed
  • the downlink interface of the unit has a corresponding relationship; the offset labels of the different level units may have different positions and different bit lengths.
  • the offset flag storage module 406 of each multiplexing unit stores the offset flag of the multiplexing unit and its corresponding relationship with the downlink interface of the multiplexing unit.
  • the offset flag module 407 finds an offset flag corresponding to the downlink interface in the offset tag storage module 406, and the received physical layer load The offset flag is added to the position of the offset flag and then output to the code stream multiplexing module 404.
  • the code stream multiplexing module 404 performs a downlink to uplink transmission rate conversion on the physical layer payload added with the offset flag, and the rate is converted. Converted physical layer The payload is output from the present multiplexing unit from the upstream interface 320.
  • the code stream multiplexing module 404 For the physical layer load received from the uplink interface 320, the code stream multiplexing module 404 performs uplink to downlink transmission rate conversion on the received physical layer payload, and outputs the rate converted physical layer payload to the offset label module 407;
  • the shift labeling module 407 extracts the offset flag at the position of the level offset flag in the converted physical layer load, and finds the downlink interface corresponding to the offset flag in the offset label storage module 406, in clearing the physical layer load. After offsetting the flag, it outputs the multiplexing unit from the downstream interface.
  • the uplink physical layer payload adds level offset flags in a step-by-step multiplexing process. Since each offset flag is unique in its multiplexing unit, the combination of the offset flags of each level has a corresponding relationship with the lower-level downstream physical port unit.
  • the offset flag of the stage is cleared step by step, and all offset flags are cleared before the downlink interface of the lowermost multiplexing unit is output.
  • the addition or clearing of the offset flag in the physical layer load in this embodiment may be performed in units of data frames, or may be performed at a fixed length smaller than the length of the data frame, but all the multiplexing units The same multiplexing method must be used. Therefore, the offset marking module 406 of each multiplex unit in this embodiment typically needs to identify the boundaries of the data frames in the physical layer payload to determine the range of code streams to which the labels are applied.
  • the code stream multiplexing module 404 of the multiplexing unit in this embodiment can be implemented by referring to the frame multiplexing module 122 in the first embodiment of the access multiplexing device.
  • VLAN number Or use multi-layer VLAN technology, one layer for each level of offset mark. In this way, the physical layer load can be directly processed on the gateway side.
  • the structure of the multiplexing unit in the eighth embodiment of the access device is as shown in FIG. 14.
  • the fixed length multiplexing module 409 is connected to the downlink interfaces 311, 312 to 31m, the sequential storage module 408, and the uplink interface 320, respectively.
  • each time slot in the timing cycle has the same length.
  • the storage module 408 stores a downlink interface corresponding to each time slot sequentially arranged in the sequence period. For each downlink interface having the same bandwidth, each downlink interface may be corresponding to the same number of time slots; and for different downlink interface bandwidths, each downlink interface may be corresponding to a number of time slots matching its bandwidth.
  • the fixed length multiplexing module 409 writes the physical layer load received from each downlink interface into the buffer area of the downlink interface, and sequentially takes the time slot in the time series, and sequentially
  • the downlink interface buffer corresponding to the slot outputs a physical layer payload of a certain length according to the clock determined by the bandwidth of the uplink interface.
  • the idle signal of the Ethernet standard or the idle signal agreed by the communication party is filled with the corresponding Time slot.
  • the fixed length multiplexing module 409 multiplexes the multiple physical layer loads of the downlink interface into one physical layer load, and the physical layer load of the current multiplexing includes the downlink interfaces corresponding to the multiplexing unit. Stream segment.
  • the fixed length multiplexing module 409 outputs the physical layer load multiplexed by the current level from the uplink interface 320 to the multiplexing unit.
  • the fixed length multiplexing module 409 For the physical layer payload received from the uplink interface 320, the fixed length multiplexing module 409 writes it to the buffer area of the uplink interface 320, in the order of the time slot in the timing cycle, and sequentially determines from the buffer area of the uplink interface 320.
  • the clock of the downlink interface bandwidth outputs a physical layer load of a certain length to the downlink interface corresponding to the time slot.
  • the fixed length multiplexing module 409 demultiplexes one physical layer load of the uplink interface 320 into multiple physical layer loads of the downlink interface, and outputs the multiplexing unit from the corresponding downlink interface.
  • the low-speed physical layer payload transmitted from the downlink interface of the lower-level multiplexing unit is multiplexed by the progressive multiplexing unit to form a high-speed physical layer load, and the high-speed physical layer generated in the timing cycle of each upper-level multiplexing unit
  • the physical layer payload includes a code stream segment corresponding to a downlink interface of each of the lowest multiplex units, and the code stream segments have the same length, and the order of the segments is corresponding to the downlink interface in the timing cycle of each multiplex unit. The order of the gap is determined.
  • the high-speed physical layer load transmitted from the uplink interface of the upper-level multiplexing unit includes a code stream segment corresponding to the downlink interface of each lower-level multiplexing unit in the same order, and the high-speed physical layer load is multiplexed by each level. After the unit is demultiplexed step by step, it is output from the downlink interface of the corresponding lower-level multiplexing unit.
  • variable length multiplexing module 411 is connected to the downlink interfaces 311, 312 to 31m, and the timing and duration storage module 410. And the upstream interface 320.
  • the node users may adopt different access rates. This embodiment is more suitable for the same length that each of the lower-level downlink physical port units may have different bandwidths.
  • the timing and duration module 410 in addition to storing the downlink interface corresponding to each time slot sequentially arranged in the sequence period, the length of the time slot is also stored, and the time slot length matches the bandwidth of the downlink interface.
  • the variable length multiplexing module 411 writes the multiple physical layer loads received from the downlink interface into the buffer area of the downlink interface, and is determined by the downlink interface buffer area corresponding to each time slot according to the time slot sequence in the time series period.
  • the clock of the uplink interface bandwidth outputs a physical layer load of a certain length, and the length of the physical layer payload corresponds to the length of the time slot; for the downlink interface that does not currently have network traffic, the idle standard signal or the communication side of the Ethernet standard on the physical layer The agreed idle signal fills its corresponding time slot.
  • variable length multiplexing module 411 multiplexes the multiple physical layer loads of the downlink interface into one physical layer load of the uplink interface, and the multiplexed physical layer load of the path includes the corresponding downlink interfaces. It is possible to have codestream segments of different lengths.
  • the variable length multiplexing module 411 outputs the multiplexed physical layer load from the uplink interface 320 to the present multiplexing unit.
  • variable length multiplexing module 411 For one physical layer load received from the uplink interface 320, the variable length multiplexing module 411 writes the physical layer load of the path into the uplink interface buffer area, and determines the downlink from the uplink interface buffer area according to the time slot sequence in the time series period.
  • the clock of the interface bandwidth outputs a physical layer load of a certain length to the downlink interface corresponding to the time slot, and the length of the output physical layer payload corresponds to the length of the time slot.
  • the variable length multiplexing module 411 demultiplexes one physical layer load of the uplink interface into multiple physical layer loads of the downlink interface, and outputs the multiplexing unit from the corresponding downlink interface.
  • the time slots in the eighth embodiment and the ninth embodiment can also be regarded as the physical layer payload length processed during multiplexing and demultiplexing, for example, the number of bytes of the physical layer payload.
  • the multiplexing unit does not need to identify or parse data frames in the physical layer payload.
  • the bandwidth allocated to each downlink connection in the uplink bandwidth in the two embodiments is strictly reserved, and the QoS performance is guaranteed from the physical layer.
  • Embodiment 8 and Embodiment 9 apply to the second of FIG. Or the third access system of Figure 4.
  • the multiplexing chip may include the following three types: The first one is a MAC interface multiplexing chip 710, that is, one multiplexing unit 711 is encapsulated in a chip, and the uplink and downlink ports input and output physical layer loads; The second type is a PHY-MAC interface multiplexing chip 720, which encapsulates a multiplexing unit 711, a downlink physical port unit 712 of each downlink interface of the connection multiplexing unit, and a downlink port input and output physical layer load, and a downlink port input.
  • the first one is a MAC interface multiplexing chip 710, that is, one multiplexing unit 711 is encapsulated in a chip, and the uplink and downlink ports input and output physical layer loads
  • the second type is a PHY-MAC interface multiplexing chip 720, which encapsulates a multiplexing unit 711, a downlink physical port unit 712 of each downlink interface of the connection multiplexing unit, and a downlink port input and
  • the physical layer signal is output; the third is the PHY interface multiplexing chip 730, a multiplexing unit 711, a downlink physical port unit 712 of each downlink interface of the connection multiplexing unit, and an uplink physical port unit 713 of the uplink interface of the connection multiplexing unit.
  • the uplink and downlink ports input and output physical layer signals.
  • the uplink port of the lower-level MAC interface multiplexing chip can directly connect to the downlink interface of the adjacent upper-level MAC interface multiplexing chip;
  • the uplink port of the PHY interface of the lower PHY interface can also directly connect to the downlink interface of the adjacent upper PHY interface multiplexing chip;
  • the PHY-MAC interface multiplex chip is used, the lower PHY-MAC interface
  • the uplink port of the multiplex chip needs to connect to the downlink interface of the adjacent upper-level PHY-MAC interface multiplexing chip through the uplink physical port unit that matches the transmission rate of the uplink port.
  • the three multiplexed chips can also be used in combination. In this case, it may be necessary to access the uplink physical port unit or the downlink physical port unit between the upper and lower multiplex chips for signal conversion.
  • the third type of Ethernet access system shown in Figure 4 uses an existing gateway device to connect to the external network, and the access switching device acts as an interface between the access device and the gateway device.
  • the access switching device in the present invention includes at least one access side physical port unit to connect to the access device, and should also include a gateway side physical port unit to connect to the gateway device.
  • the access switching device also needs to employ a multiplexing technique that is compatible with the access device.
  • FIG. 17 to FIG. 19 are respectively schematic structural diagrams of Embodiments 1 to 3 of the access switching apparatus according to the present invention.
  • the multiplexing conversion unit 520 is respectively connected to the access side physical port unit 510 and the gateway side physical port unit 530 (required description
  • the gateway side indicates that it is close to the network. In the off direction, it does not mean that the port of the gateway is directly connected. It can be understood by those skilled in the art that a conventional Ethernet switch or even a Layer 2 Ethernet ring network can still exist between the access switching device and the gateway.
  • the access side physical port unit 510 is connected to the access node through the access device, and parses the multiplexed physical layer payload carried therein when receiving the physical layer signal, and outputs the multiplexed physical layer payload to the multiplex conversion unit 520.
  • the multiplex conversion unit 520 multiplexes
  • the physical layer payload is converted to a physical layer load in units of complete data frames and output to the gateway side physical port unit 530; the gateway side physical port unit 530 transmits the physical layer payload in units of data frames to the physical layer signal to Gateway device.
  • the gateway side physical port unit 530 parses it into a physical layer payload in units of data frames, and outputs it to the multiplexing conversion unit 520.
  • the multiplexing conversion unit 520 will be in units of data frames.
  • the physical layer load is reversely converted to the multiplexed physical layer load and then output to the access physical port unit 510.
  • the access side physical port unit 510 converts the multiplexed physical layer load into a physical layer signal and outputs the signal to the access device.
  • the first embodiment to the third embodiment of the access switching device differ in that the internal implementation of the multiplexing conversion unit 520 is different due to the different multiplexing techniques employed.
  • FIG. 17 is a schematic structural diagram of Embodiment 1 of the access switching device.
  • the multiplex conversion unit 520 includes a multiplex mark module 521 and a multiplex mark processing module 522, and the multiplex mark processing module 522 is connected to the access physical port unit 510, the multiplex mark unit 521, and the gateway physical port unit 530, respectively.
  • the access switching device of this embodiment can be used in conjunction with the access device embodiments 2, 6, and 7.
  • the uplink physical layer payloads of the access devices of the second, sixth and seventh embodiments all include a downlink physical port unit connected to the access node, or a label corresponding to the access node.
  • the access side physical port unit 510 of the access switching device receives the uplink physical layer signal of the access device, the physical layer load carried therein is parsed and output to the multiplexing tag processing module 522.
  • the multiplexing tag processing module 522 extracts the tag in the physical layer payload, generates a physical layer payload in units of complete data frames according to the tag, and checks whether the access node identifier in the tag and the data frame has been saved in the multiplexing tag module 521.
  • the multiplex tag module 521 is updated if the data is not saved or changed, so that the correspondence between the current tag and the access node identifier is maintained in the multiplex tag module 521; the tag processing module 522 is multiplexed and The physical layer payload in units of data frames after the flag is cleared is output to the gateway side physical port unit 530.
  • the multiplex tag processing module 522 can obtain the physical layer payload in units of data frames after clearing the tag; and when the access device uses a fixed length code When a stream is added with a marker in the physical layer payload, the multiplexed token processing module 522 may need to recombine the physical layer payloads with the same token, clearing the tokens therein to obtain the physical layer payload in units of data frames.
  • the access node identifier is a field that is corresponding to the access node in the packet sent by the access node or the packet transmitted to the access node, and may be the MAC address, IP address, and VLAN number of the access node. And other fields. After the access node is powered on, the access node first sends a packet to the gateway device, so that the access switching device can learn the correspondence between the access node identifier and the tag.
  • the multiplexing tag processing module 522 finds a tag corresponding to the access node identifier in the data frame in the multiplexing tag module 521, This flag is added to the physical layer payload and output to the access side physical port unit 510.
  • the multiplexing tag processing module 522 can also connect the access side physical port unit 510 corresponding to the access node identifier. The same is stored in the multiplexing tag module 521.
  • the multiplex tag processing module 522 searches for the corresponding access side physical port unit 510 when searching for the tag corresponding to the access node identifier in the data frame, and The marked downlink physical layer payload is added to the access side physical port unit 510.
  • the access switching device should process the tag in a manner that matches the access device connected thereto. If the access device multiplexes in units of data frames, the access switching device also processes the tag in units of data frames; if the access device multiplexes with a fixed length, the access switching device also performs a fixed-length code stream. Tag processing, and the location of the tag in the data frame or fixed length code stream should also be the same.
  • the marking in this embodiment is added by the lowest multiplex unit of the access device in the code stream and traverses all the upper multiplexes.
  • the marking of the unit; when the access system adopts the access device embodiment seven networking, the marking in this embodiment includes offset markings of all levels of multiplexing units.
  • FIG. 18 is a schematic structural diagram of Embodiment 2 of the access conversion apparatus.
  • the multiplexing conversion unit 520 includes a multiplexing timing module 523, a timing conversion module 524, and a data frame identification module 525.
  • the timing conversion module 524 is connected to the access side physical port unit 510, the multiplexing timing module 523, and the data frame identification module 525, respectively.
  • the data frame identification module 525 is connected to the gateway side physical port unit 530.
  • the access switching device of this embodiment can be used in conjunction with the access device embodiments three and eight.
  • the uplink physical layer payloads of the access devices of Embodiments 3 and 8 are all multiplexed according to the multiplexing timing cycle, and the lengths of the time slots in the multiplexing timing cycle are the same.
  • the multiplexed physical layer payloads each include a downlink physical port unit corresponding to the connected access node, or a code stream segment corresponding to the access node.
  • the timing conversion module 524 For the multiplexed physical layer payload received from the access side physical port unit 510, the timing conversion module 524 writes the multiplexed physical layer payload segment into the slot according to the slot order in the multiplexed sequence period. The upstream buffer area of the time slot.
  • the data frame identification module 525 extracts the physical layer payload in units of complete data frames from the uplink buffer area of each time slot, and checks whether the access node identifier and the corresponding time slot in the data frame have been saved in the multiplexing sequence module 523.
  • the multiplexing timing module 523 is updated to maintain the correspondence between the current time slot and the access node identifier in the multiplexing timing module 523; the data frame identification module 525 and the physical layer in units of data frames
  • the payload is output to the gateway side physical port unit 530.
  • the data frame identification module 525 finds the time slot corresponding to the access node identifier in the data frame in the multiplexing timing module 523, The physical layer load in units of the data frame is written to the downlink buffer of the time slot.
  • the timing conversion module 524 outputs a physical layer load of a certain length from the downlink buffer area of each time slot to the access side physical port unit 510 in units of time slots in the multiplexing timing period in the order of time slots.
  • the access switching apparatus in this embodiment should adopt a time division multiplexing manner matching the access apparatus connected thereto, such as a matched multiplexing timing period and a matching number of slots and lengths to ensure the physical layer.
  • the load is properly multiplexed and decomposed at both ends of the transmission use.
  • FIG. 19 is a schematic structural diagram of Embodiment 3 of the access switching device.
  • the multiplexing conversion unit 520 includes a multiplexing timing and duration module 526, a timing and duration conversion module 527, and a data frame identification module 525.
  • the timing and duration conversion module 527 is connected to the access side physical port unit 510 and the multiplexing timing and duration module 526, respectively.
  • the data frame identification module 525, the data frame identification module 525 is connected to the gateway side physical port unit 530.
  • the access switching device of this embodiment can be used in conjunction with the access device embodiments four and nine.
  • the uplink physical layer payloads of the access devices of Embodiments 4 and 9 are all multiplexed according to the multiplexing timing period, and the lengths of the slots in the multiplexing timing period may be different.
  • the multiplexed physical layer payloads each include an arrangement order corresponding to a downlink physical port unit connected to the access node, or a code stream segment corresponding to the access node, and the length of the code stream segment corresponds to the slot length.
  • the length and arrangement order of the respective time slots in the multiplexed timing cycle are pre-stored in the multiplexed timing and duration module 526.
  • the timing and duration conversion module 527 multiplexes the code stream in the physical layer payload according to the slot order in the multiplexed sequence period.
  • a code stream segment whose length corresponds to the length of the time slot is written into the uplink buffer area of the time slot.
  • the data frame identification module 525 extracts a physical layer load in units of complete data frames from the uplink buffer area of each time slot, and checks whether the access node identifier and the corresponding time slot in the data frame have been saved in the multiplexing timing and duration module.
  • the multiplexing timing and duration module 526 if not saved or changed, the multiplexing timing and duration module 526 is updated to maintain the correspondence between the current slot, the slot length and the access node identifier in the multiplexing timing and duration module 526; the data frame identification module 525 outputs the physical layer payload in units of data frames to the gateway side physical port unit 530.
  • the data frame identification module 525 finds the access node in the data frame in the multiplexing timing and duration module 526. The corresponding time slot is identified, and the physical layer is loaded and written into the downlink buffer area of the time slot.
  • the timing and duration conversion module 527 outputs the code stream length from the downlink buffer area of each time slot to the access side physical port unit 510 in the order of time slots in units of time slots in the multiplexing timing period, corresponding to the time slot. The physical layer load of the length.
  • the access switching device in this embodiment should adopt a time division multiplexing manner matching the access device connected thereto, such as matching multiplexing timing period, matching number of slots and slot length, and matching.
  • the order is arranged to ensure proper multiplexing and demultiplexing of physical layer loads at both ends of the transmission.
  • the multiplexing gateway device needs to implement multiplexing by using a multiplexing technology that cooperates with the access device.
  • the multiplexing gateway device in the present invention can be realized by integrating the main function modules of the above access switching device into an existing gateway device.
  • an embodiment of a multiplexing gateway device includes a physical port unit, a multiplexing interface unit, and a forwarding unit, where:
  • the physical port unit is configured to convert between a physical layer signal of the input/output multiplexing gateway device and a multiplexed MAC layer code stream carried therein;
  • the multiplexing interface unit is configured to convert the multiplexed MAC layer code stream received from the physical port unit into a data frame and output to the forwarding unit, and inversely convert the data frame received from the forwarding unit into a multiplexed MAC layer code stream and output to the Physical port unit;
  • the forwarding unit is used for data frame forwarding.
  • all the physical port units in the access device, the multiplexing gateway device, and the access switching device of the present invention have the functions of converting the physical layer signal and the physical layer load, and the function is different according to different bandwidths. Can be implemented by using existing methods
  • FIG. 20 is a flowchart of an Ethernet uplink multiplexing access method according to the present invention. In the steps
  • the low-speed physical layer signals received from the respective access nodes are respectively parsed from the low-speed physical layer loads corresponding to the respective access nodes.
  • step S12 the low-speed physical layer loads of the respective access nodes are multiplexed into one high-speed physical layer load.
  • step S13 the multiplexed high-speed physical layer payload is carried in the high-speed physical layer signal.
  • step S14 the high speed physical layer signal is uplinked.
  • the low-speed physical layer load is multiplexed into one high-speed physical layer load in step S12, which may be one-level multiplexing or multi-stage multiplexing.
  • Multi-stage multiplexing In a hierarchical manner, the access nodes are grouped first, and the physical layer load of the access nodes in each group is multiplexed into one physical layer load of each group, and then the physical layer loads of each group are regrouped, multiplexed or directly multiplexed. Until the physical layer load of all access nodes is multiplexed into one physical layer load.
  • the uplink multiplexing technology in the foregoing embodiments of the access device may be used in the first-level or multi-stage multiplexing, which is only briefly described herein. For details, refer to the foregoing embodiment.
  • step S12 When the field in the data frame having a one-to-one correspondence with the access node, that is, the access node identifier is used as the mark of the physical layer load, the steps after step S12 are the same as the prior art.
  • the time division multiplexing technique is adopted, in step S12, each time slot corresponding to a physical layer load before multiplexing in the sequence period, the code stream length is matched with the physical layer load output of the time slot length. The physical layer load after use.
  • step S12 before each stage of multiplexing, the physical layer load before multiplexing is added to the corresponding physical layer load of the current multiplexing--the offset flag.
  • FIG. 21 is a flowchart of an Ethernet downlink multiplexing access method according to the present invention.
  • step S21 a high speed physical layer signal of the downlink transmission is received.
  • step S22 the multiplexed physical layer load is parsed out in the received high-speed physical layer signal.
  • step S23 the multiplexed physical layer payload is demultiplexed into a low-speed physical layer payload corresponding to the access node.
  • step S24 the demultiplexed low-speed physical layer bearer bearer is sent to the corresponding access node in the low-speed physical layer signal.
  • the physical layer load corresponding to the access node may be demultiplexed in the step S23 or may be a multi-stage demultiplexing.
  • the multi-stage demultiplexing is performed by first demultiplexing the multiplexed physical layer load into multiple physical layer loads; and then demultiplexing the demultiplexed physical layer loads respectively until demultiplexing corresponds to The physical layer load of the access node.
  • the uplink multiplexing technology in the foregoing embodiments of the foregoing access device may be used for performing the one-level or multi-stage demultiplexing, which is only briefly described herein. For details, refer to the foregoing embodiment.
  • the access node identifier is used as the identifier of the physical layer payload
  • the access node identifier is demultiplexed in step S23.
  • the label in the code stream should be removed before converting the physical layer payload corresponding to the access node to the low-speed physical layer signal.
  • each time slot corresponding to a certain physical layer load before multiplexing in the sequence period matches the code stream length to the pre-demultiplexed physical layer load output of the time slot length.
  • the physical layer load after demultiplexing corresponding to the time slot is adopted.
  • each stage demultiplexing has an offset flag corresponding to each physical layer load after demultiplexing according to the physical layer load before demultiplexing. Performing and clearing the offset flag of the physical layer payload according to this demultiplexing each time demultiplexing.
  • the access node can adopt the existing standard Ethernet physical layer rate.
  • a non-standard Ethernet physical layer transmission rate occurs in the future, such as 2M, 40M, 80M, etc., as long as both ends of the physical link are These non-standard transfer rates are supported.
  • the uplink physical port that outputs the high-speed physical layer signal in the access device may not be the standard Ethernet physical layer rate. Even some physical ports may not be standard Ethernet ports, as long as both ends of the physical link can be simultaneously supported and can be carried.
  • the present invention is equally applicable to Ethernet physical layer loading. In a broader sense, the present invention is not concerned with the specific physical layer processing at both ends of each physical link, as will be readily appreciated by those skilled in the art.
  • the networking structure of the access node to the gateway device in the access system is very simple, and the traffic of the access node is multiplexed by the access device and sent to the upper device for processing.
  • the present invention does not need to perform the layer 2 forwarding process, and only works at one layer, so the reliability is more advantageous than the access through the layer 2 switch, and at the same time, according to the related embodiment of the present invention.
  • the cost of the access device of the present invention is expected to be about one tenth of that of a common switch, and the occupied space is very small, which is suitable for being placed in a residential corridor.
  • access devices are reliable and cost-effective; for users, they can be directly accessed through computer network cards, without the need for modems and dial-up software, providing users with ease of use while reducing user convenience. cost. a large number of users In the case of dense distribution, access devices using cascade multiplexing will be more suitable.
  • the invention provides a networking scheme of an Ethernet broadband access system through the fusion multiplexing technology and the Ethernet technology, which can greatly reduce the broadband access cost, simplify the access network architecture, and provide a high-speed and reliable access rate.

Landscapes

  • Engineering & Computer Science (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Small-Scale Networks (AREA)

Description

以太网接入装置及接入方法 技术领域
本发明涉及以太网接入技术，尤其涉及一种以太网接入装置及接 入方法。 背景技术
随着互联网的高速发展， 尤其是 IPTV ( Internet Protocol Television, 交互式网絡电视）等新业务的兴起， 终端用户最后一公里 的接入问题成为新业务开展的瓶颈。 当前的主流接入技术 ADSL ( Asymmetrical Digital Subscriber Loop, 非对称数字用户线环路 ) 不 能提供足够的带宽支持， 因此业界目前正在研究 VDSL ( Very High Bit-Rate Digital Subscriber Loop,甚高速数字用户线）和 ADSL2+( ITU 标准 G.992.5 )等技术以期解决上述问题， 但目前这些新技术由于尚 不够成熟， 商业应用平均到每个用户成本依然较高， 服务提供商并不 敢贸然采用。 因此， 目前宽带接入市场仍然以 ADSL接入为主。
以太网技术已经发展得非常成熟， 并且随着芯片厂商的技术进 步， 其传输距离大幅上升， 这为以太网应用于宽带接入带来了新的机 会。 现有技术中， 用于接入宽带终端用户的以太网接入系统的典型结 构如图 1所示， 由二层交换机 810用作接入设备、三层交换机 820用 作网关，接入终端通过二层交换机 810接入网关 820。二层交换机 810 在用户和网关 820之间进行二层转发，完成各个接入终端之间的二层 隔离， 其他工作如 QoS ( Quality of Service, 服务质量）等均可以在 网关 820处完成。
在宽带接入领域， 每个用户的平均带宽比较低， 例如 10M (兆） 带宽可以满足 80 %以上的宽带接入应用。 而以太网产品如 MAC ( Media Access Control, 媒介接入控制）芯片、 长距离传输介质的带 宽已经发展到满足企业网应用的 1G (吉）、 10G, 并且其价格相对于 带宽具有相当的优势。但将这些产品直接应用于宽带接入会造成带宽 的巨大浪费， 同时参考图 1来看， 用户侧的设备成本仍然比较高， 服 务提供商按此方案进行网络部署难以降低平均到每用户的接入设备 等成本。 而且在目前实际应用中， 在最靠近用户侧， 服务提供商需要 布置大量的二层交换机，为了保证每个可能的用户需要接入互联网的 服务时都能够立刻为其开通接入服务，这些二层交换机通常具有几十 个端口的接入容量， 一般可能与附近的家庭数量相当， 甚至大于家庭 数量。 但实际上的接入开通率很低， 往往一栋楼只开通不到 10户， 这样就造成一个二层以太网交换机的成本被几个用户分担 ,显然成本 太高，这也是以太网这个成熟技术不能广泛地在互联网接入应用中获 得一席之地的关键原因之一。
如何降低以太网接入设备的成本，这是服务提供商一直关心的问 题。 对于设备提供商来说， 由于以太网的市场巳经相当成熟， 各种商 用芯片的价格已经难有大幅度的下降空间， 简化设备、 降低设备成本 只得另寻途径。
IBM公司创新地利用复用技术在一定程度上达到筒化以太网交 换机的目的，具体可参考其申请的公开号为 20010050921的美国专利 申请第， 该专利申请公开一种物理层时分复用的实现方案， 循环地采 集物理层芯片的多路低速物理端口在一定时间段内的输入信号，将其 复用为一路高速数字信号后传输至 MAC芯片进行二层处理，对 MAC 芯片输出的数字信号其处理过程相反。 由此降低以太网交换机内 MAC芯片的数量， 在以太网交换机内用一个 100M/GE的 MAC芯片 代替多个 10M/100M的 MAC芯片， 总体成本也相应地得到降低。
现有技术中的该专利申请揭示的以太网复用技术与传统的以太 网相关的复用技术最大的不同就在于其利用复用技术解决以太网交 换机的内部筒化问题，而不是像其他以太网技术中利用复用解决以太 网传输物理链路成本的问题， 或者解决由于设备不支持 10GE数据处 理能力而无法直接出 10GE物理通道的问题，例如将 10个 GE通道通 过复用装置复用到一个 10GE的通道进行传输， 在另一端再进行解复 用后恢复出相应的 GE通道。
IBM提交的该专利申请确实在一定程度上解决以太网设备简化 的问题， 从而使得设备的可靠性得到提高， 并且设备整体成本得到降 低。 如果将这样的技术应用于宽带接入，服务提供商的成本问题会得 到一定的緩解， 但这样的优化程度仍然有限， 这是因为， 按照所述现 有技术，接入设备将仍然具备较多的 MAC层及上层功能套片。因此， 要使得以太网的宽带接入能够成为主流接入手段，接入设备的总体成 本必须还要降低。 发明内容
本发明提供一种以太网接入装置及接入方法， 其成本较低， 设备 架构较为简化。
根据本发明的一方面，所述以太网接入装置位于多个接入节点和 网关设备之间， 包括至少两个下行物理端口单元、 复用解复用单元和 上行物理端口单元， 其中：
下行物理端口单元用来在输入输出接入装置的低速物理层信号 与其中承载的低速物理层载荷之间进行转换；
上行物理端口单元用来在输入输出接入装置的高速物理层信号 与其中承载的高速物理层载荷之间进行转换；
复用解复用单元用来在下行物理端口单元的低速物理层载荷与 上行物理端口单元的高速物理层载荷之间进行复用与解复用。
可选地，所述复用解复用单元的复用与解复用根据物理层载荷中 与下行物理端口单元对应的标记进行。
可选地， 所述标记为物理层载荷中数据帧内的字段， 该字段与下 行物理端口单元具有对应关系；
所述复用解复用单元包括标记存储模块和帧复用模块， 其中： 标记存储模块用来存储标记及其对应的下行物理端口单元； 帧复用模块用来将下行物理端口单元输入的低速物理层载荷以 数据帧为单位转换为高速物理层载荷输出至上行物理端口单元，以及 将从上行物理端口单元输入的高速物理层载荷以数据帧为单位转换 为低速物理层载荷并输出至其中标记所对应的下行物理端口单元。
可选地， 所述复用解复用单元包括标记存储模块、标记模块和码 流复用模块， 其中： 标记存储模块用来存储标记及其对应的下行物理端口单元； 标记模块用来在从下行物理端口单元输入的低速物理层载荷中 添加与该下行物理端口单元对应的标记并输出至码流复用模块，以及 在从码流复用模块输入的物理层载荷中去除标记并从与该标记对应 的下行物理端口单元输出；
码流复用模块用来在标记模块的低速物理层载荷与上行物理端 口单元的高速物理层载荷之间进行转换。
可选地，所述标记模块在物理层载荷中添加和去除标记以数据帧 为单位或以固定长度码流为单位进行。
可选地，所述复用解复用单元的复用与解复用才 居时序周期循环 进行， 时序周期中包括对应于各个下行物理端口单元的时隙。
可选地，所述时序周期中对应于各个下行物理端口单元的时隙长 度相同；
所述复用解复用单元包括时序存储模块和定长复用模块， 其中： 时序存储模块用来存储时序周期中时隙对应的下行物理端口单 元；
定长复用模块在各个时隙将从对应的下行物理端口单元接收的 低速物理层载荷转换为一路高速物理层载荷输出至上行物理端口单 元，以及将从上行物理端口单元接收的高速物理层载荷转换为低速物 理层载荷在各个时隙输出至对应的下行物理端口单元。
可选地，所述复用解复用单元包括时序与时长存储模块和变长复 用模块， 其中：
时序与时长存储模块用来存储时序周期中各时隙的长度、该时隙 所对应的下行物理端口单元；
变长复用模块用来在各个时隙将从对应的下行物理端口单元接 收的码流长度匹配于该时隙长度的低速物理层载荷转换为一路高速 物理层载荷并输出至上行物理端口单元，以及将从上行物理端口单元 接收的高速物理层载荷转换低速物理层载荷，在各个时隙将码流长度 匹配于该时隙长度的低速物理层载荷输出至对应的下行物理端口单 元。 根据本发明的另一方面，一种以太网接入装置位于多个接入节点 和网关设备之间， 包括多个最下级下行物理端口单元、 最上级上行物 理端口单元和至少两级复用单元 ,下級复用单元的上行接口与上级复 用单元的下行接口连接， 其中：
最下级下行物理端口单元连接最下级复用单元的下行接口，用来 在输入输出接入装置的低速物理层信号与其中承载的低速物理层载 荷之间进行转换；
最上级上行物理端口单元连接最上级复用单元的上行接口，用来 在输入输出接入装置的高速物理层信号与其中承载的高速物理层载 荷之间进行转换；
各级复用单元在最下级复用单元下行接口的低速物理层载荷与 最上级复用单元上行接口的高速物理层载荷之间进行逐级的复用与 解复用。
可选地，所述各级复用单元的复用与解复用根据物理层载荷中与 最下级复用单元下行接口对应的标记进行。
可选地，所述标记为最下级复用单元下行接口的物理层载荷中数 据帧内区别于其他下行接口的字段；
所述每级复用单元包括标记存储模块和帧复用模块， 其中： 标记存储模块用来存储标记及其对应的本复用单元下行接口； 帧复用模块用来将下行接口输入的多路物理层载荷以数据帧为 单位转换为一路物理层载荷后由上行接口输出，以及将从上行接口输 入的一路物理层载荷以数据帧为单位转换为多路物理层载荷并由其 中标记所对应的下行接口输出。
可选地， 所述最下级复用单元包括标记存储模块、标记模块和码 流复用模块， 其中：
标记存储模块用来存储标记及其对应的本复用单元下行接口； 标记模块用来在从下行接口输入的物理层载荷中添加与该下行 接口对应的标记并输出至码流复用模块，以及在从码流复用模块输入 的物理层载荷中去除标记并从与该标记对应的下行接口输出；
码流复用模块用来在标记模块的多路物理层载荷与上行接口的 一路物理层载荷之间进行转换；
所述其他各级复用单元包括标记存储模块和码流上级复用模块， 其中码流上级复用模块用来将从下行接口输入的多路物理层载荷转 换为一路物理层载荷从上行接口输出，以及将从上行接口输入的一路 物理层载荷转换为多路物理层载荷并从与其中标记对应的下行接口 输出。
可选地，所述最下级复用单元的标记模块在物理层载荷中添加和 去除标记以数据帧为单位或以固定长度码流为单位进行。
可选地， 所述标记包括每级复用单元的偏移标记；
所述各级复用单元包括偏移标记存储模块、偏移标记模块和码流 复用模块， 其中：
偏移标记存储模块用来存储本复用单元的偏移标记及其对应的 本复用单元下行接口；
偏移标记模块用来在从下行接口输入的物理层载荷中添加与下 行接口对应的偏移标记并输出至码流复用模块，以及在从码流复用模 块输入的物理层载荷中去除本复用单元的偏移标记并从与该偏移标 记对应的下行接口输出；
码流复用模块用来在偏移标记模块的多路物理层载荷与上行接 口的一路物理层载荷之间进行转换。
可选地，所述各级复用单元的偏移标记模块在物理层载荷中添加 和去除偏移标记以数据帧为单位或以固定长度码流为单位进行。
可选地 ,所述各级复用单元的复用与解复用根据该复用单元的时 序周期循环进行， 时序周期中包括对应于该复用单元下行接口的时 隙。
可选地， 所述每个复用单元的时序周期中的时隙具有相同的长 度；
所述每级复用单元包括时序存储模块和定长复用模块， 其中： 时序存储模块用来存储该复用单元的时序周期中时隙对应的下 行接口；
定长复用模块在各个时隙将从对应的下行接口接收的多路物理 层载荷转换为一路物理层载荷后从上行接口输出，以及将从上行接口 接收的一路物理层载荷转换为多路物理层载荷后在各个时隙由对应 的下行接口输出。
可选地，所述各级复用单元包括时序与时长存储模块和变长复用 模块， 其中：
时序与时长存储模块用来存储该复用单元的时序周期中时隙的 长度、 该时隙对应的下行接口；
变长复用模块用来在各个时隙将从对应的下行接口接收的码流 长度匹配于该时隙长度的物理层载荷转换为一路物理层载荷后从上 行接口输出，以及将从上行接口接收的一路物理层载荷进行速率转换 后 ,在各个时隙将码流长度匹配于该时隙长度的物理层载荷从对应的 下行接口输出。
优选地，所述接入装置还包括除最下级以外的各级下行物理端口 单元， 分别连接该级复用单元的每个下行接口， 用来进行物理层信号 与物理层载荷之间的转换；每个复用单元与其连接的下行物理端口单 元封装在一个物理层 PHY-MAC接口复用芯片中；
所述接入装置还包括除最上級以外的各级上行物理端口单元，连 接在下级 PHY-MAC接口复用芯片与上级 PHY-MAC接口复用芯片之 间， 用来进行物理层信号与物理层载荷之间的转换。
可选地，所述接入装置还包括除最下级以外的各级下行物理端口 单元， 分别连接该级复用单元的每个下行接口， 用来进行物理层信号 与物理层载荷之间的转换；
所述接入装置还包括除最上级以外的各级上行物理端口单元，分 别连接该级复用单元的每个上行接口，用来进行物理层信号与物理层 载荷之间的转换；
每个复用单元与其连接的下行物理端口单元和上行物理端口单 元封装在一个 PHY接口复用芯片中，下级 PHY接口复用芯片的上行 物理端口单元连接至上级 PHY接口复用芯片的下行物理端口单元。
根据本发明的又一方面， 一种以太网接入方法， 包括以下步骤： 将至少两路接入节点的物理层信号中承载的物理层载荷复用为 一路物理层载荷；
将复用后的物理层载荷转换为高速物理层信号后进行上行传输； 将下行高速物理层信号中承载的一路复用物理层载荷解复用为 与至少两个接入节点对应的物理层载荷；
将解复用后的物理层载荷转换为低速物理层信号后进行下行传 输。
可选地，所述将接入节点的物理层载荷复用为一路物理层载荷具 体为： 将同组接入节点的多路物理层载荷复用为每组一路物理层载 荷； 将每组一路物理层载荷再次分組后复用或直接复用， 直至复用为 一路物理层载荷；
所述将一路复用物理层载荷解复用为对应于接入节点的物理层 载荷具体为： 将复用物理层载荷解复用为多路物理层载荷； 将解复用 后的各路物理层载荷分别再次解复用，直至解复用为对应于接入节点 的物理层载荷。
可选地， 所述方法在每次复用前还包括： 在复用前的物理层载荷 中添加与本次复用的各路物理层载荷——对应的偏移标记；
所述每次解复用根据解复用前物理层载荷中与解复用后的各路 物理层载荷具有——对应关系的偏移标记进行；
所述方法在每次解复用时还包括：去除物理层载荷中据以进行本 次解复用的偏移标记。
可选地， 所述方法在进行物理层载荷复用前还包括： 在接入节点 的物理层载荷中添加与接入节点具有对应关系的标记；
所述将一路物理层载荷解复用为多路物理层载荷根据物理层载 荷中与接入节点具有对应关系的标记进行；
所述方法在将对应于接入节点的物理层载荷转换为低速物理层 信号前还包括： 在对应于接入节点的物理层载荷中去除所述标记。
可选地 ,所述将多路物理层载荷复用为一路物理层载荷按照时序 周期循环进行；所述时序周期包括对应于复用前各路物理层载荷的时 隙，在每个时隙将对应的复用前物理层载荷输出为复用后的一路物理 层载荷， 所述复用前物理层载荷的长度匹配于该时隙的长度； 所述将一路物理层载荷解复用为多路物理层载荷按照时序周期 循环进行； 所述时序周期包括对应于解复用后各路物理层载荷的时 隙，在每个时隙将解复用前码流长度匹配于该时隙长度的物理层载荷 输出为对应的解复用后物理层载荷。
可选地，所述将一路物理层载荷解复用为多路物理层载荷根据物 理层载荷的数据帧中与接入节点具有对应关系的字段进行。
根据本发明的再一方面， 一种以太网接入转换装置，位于接入装 置和网关设备之间； 包括接入侧物理端口单元、 复用转换单元和网关 侧物理端口单元； 其中，
所述接入侧物理端口单元用来在接入转换装置和接入装置之间 传输的物理层信号和其中承载的复用物理层载荷之间进行转换； 所述网关侧物理端口单元用来在接入转换装置和网关设备之间 传输的物理层信号及其中承载的以数据帧为单位的物理层载荷之间 进行转换；
所述复用转换单元用来在接入侧物理端口单元的复用物理层载 荷和网关侧物理接口单元的以数据帧为单位的物理层载荷之间进行 转换。
可选地，所述复用转换单元包括复用标记模块和复用标记处理模 块； 其中，
所述复用标记模块用于存储标记与接入节点标识的对应关系； 所述复用标记处理模块用于取出接入侧物理端口单元输入的物 理层载荷中的标记， 居标记生成以数据帧为单位的物理层载荷， 将 清除标记后的以数据帧为单位的物理层载荷输出至网关侧物理端口 单元； 对从网关侧物理端口单元接收的以数据帧为单位的物理层载 荷，在复用标记模块中查找与该数据帧中的接入节点标识相对应的标 记 , 在物理层载荷中添加该标记后输出至接入侧物理端口单元。
可选地， 当接入装置以数据帧为单位在物理层载荷中添加标记 时，所述复用标记处理模块在清除标记后即可得到以数据帧为单位的 物理层载荷； 当接入装置以固定长度码流为单位在物理层载荷中添加 标记时，所述复用标记处理模块将具有相同标记的物理层载荷重新组 合， 清除其中的标记后得到以数据帧为单位的物理层载荷。
可选地， 所述复用转换单元包括复用时序模块、 时序转换模块和 数据帧识别模块； 其中，
所述复用时序模块用来存储时序周期中时隙与接入节点标识的 对应关系；
所述时序转换模块用来以复用时序周期中的时隙为单位，按照时 隙的排列顺序将从接入侧物理端口单元接收的复用物理层载荷分段 写入该时隙的上行緩存区； 以复用时序周期中的时隙为单位， 按照时 隙的排列顺序从每个时隙的下行緩存区中向接入侧物理端口单元输 出一定长度的物理层载荷；
所述数据帧识别模块用来从各个时隙上行缓存区以数据帧为单 位取出物理层载荷，将以数据帧为单位的物理层载荷输出至网关侧物 理端口单元；将从网关侧物理端口单元接收的以该数据帧为单位的物 理层载荷后写入该时隙的下行緩存区。
可选地， 所述复用转换单元包括复用时序与时长模块、 时序与时 长转换模块和数据帧识别模块； 其中，
所述复用时序与时长模块用来存储复用时序周期中各个时隙的 长度及排列顺序， 以及时隙与接入节点标识的对应关系；
所述时序与时长转换模块用来以复用时序周期中的时隙为单位， 按照时隙的排列顺序将从接入侧物理端口单元接收的复用物理层载 荷中码流长度对应于该时隙长度的码流段写入该时隙的上行緩存区； 以复用时序周期中的时隙为单位 ,按时隙的排列顺序从每个时隙的下 行緩存区中向接入侧物理端口单元输出码流长度对应于该时隙长度 的物理层载荷；
所述数据帧识别模块用来从各个时隙上行緩存区中以数据帧为 单位取出一段物理层载荷，将以数据帧为单位的物理层载荷输出至网 关侧物理端口单元；将从网关侧物理端口单元接收的发送至接入节点 的以数据帧为单位的物理层载荷写入该时隙的下行緩存区。
根据本发明的又一方面. 一种复用网关设备包括物理端口单元、 复用接口单元和转发单元， 其中： 物理端口单元用来在输入输出复用网关设备的物理层信号与其 中承载的复用 MAC层码流之间进行转换；
复用接口单元用来将从物理端口单元接收的复用 MAC层码流转 换为数据帧并输出至转发单元，以及将从转发单元接收的数据帧逆向 转换为复用 MAC层码流并输出至物理端口单元；
转发单元用来进行数据帧转发。
本发明对多路物理层信号的载荷，即其中承载的物理层载荷进行 复用， 并将复用后的物理层载荷经过物理层处理后再进行传输，使得 本发明相对于现有技术， 能够更大程度地降低以太网接入设备的成 本， 将不必要的 MAC层及上层功能套片去除， 接入装置的架构得到 简化的同时具备网关到用户之间的接入设备所必须的功能； 同时， 由 于接入设备简化， 其可靠性也相应得到提高。 附图说明
图 1为现有技术中以太网接入系统的结构示例图；
图 2 为根据本发明的实施例的第一种以太网接入系统的结构示 意图；
图 3 为才 据本发明的实施例的第二种以太网接入系统的结构示 意图；
图 4 为根据本发明的实施例的第三种以太网接入系统的结构示 意图；
图 5为本发明中接入装置实施例一的结构示意图；
图 6为本发明中接入装置实施例二的结构示意图；
图 7为本发明中接入装置实施例三的结构示意图；
图 8为本发明中接入装置实施例四的结构示意图；
图 9 为根据本发明实施例的采用级联复用的接入装置的逻辑结 构示例图；
图 10为本发明中接入装置实施例五复用单元的结构示意图； 图 11为本发明中接入装置实施例六的最下级复用单元结构示意 图；
图 12 为本发明中接入装置实施例六的上级复用单元结构示意 图；
图 13为本发明中接入装置实施例七复用单元的结构示意图； 图 14为本发明中接入装置实施例八复用单元的结构示意图； 图 15为本发明中接入装置实施例九复用单元的结构示意图； 图 16为根据本发明实施例的复用芯片的逻辑结构及连接方式示 意图；
图 17为本发明中接入转换装置实施例一的结构示意图； 图 18为本发明中接入转换装置实施例二的结构示意图； 图 19为本发明中接入转换装置实施例三的结构示意图； 图 20为本发明实施例所述以太网上行接入方法的流程图； 图 21为本发明实施例所述以太网下行接入方法的流程图。 具体实施方式
与企业网应用不同，接入网中的接入节点主要通过网关设备与外 部网络通信， 较少有接入节点之间直接通信的情况。 这样在网关设备 与接入节点之间的接入设备只要能够提供上下行传输功能，即能够将 接入节点发送的信号上行传输至网关设备、将网关设备发送的信号下 行传输给目的接入节点就能够满足接入网的需要了，而二层交换机中 在各个下行端口之间转发报文的功能在接入网中基本上处于闲置状 态。 即使存在少量接入节点之间需要直接通信的情况， 也可以由网关 设备来完成转发功能。 因此， 在接入节点一侧进行复用， 将复用后的 信号远距离传输至网关设备，只在网关设备上实现转发功能的方案正 是适合于接入系统的方案， 并且具有相当低的接入成本。
另一个与企业网应用的不同之处是，宽带接入领域的节点用户分 布极为广泛， 并且所需的接入速率、传输介质也往往各不相同。 在复 用技术中， 进行复用的数据在 OSI ( Open System Interconnection, 开 放系统互连）七层模型中的层次越高， 将复用后的数据返回物理层传 输所需的实现复杂度和成本也越高。综合考虑接入系统应用环境的多 样性和实现， 对物理层载荷进行复用既能够满足应用需求， 又具有尽 可能低的接入成本。 如前所述， 本发明中对物理层载荷进行复用， 本申请文件中所称 的码流为物理层载荷或物理层载荷的一部分。
在本发明中，根据所采用的复用技术的不同， 以太网接入系统可 能具有图 2、 图 3或图 4所示的结构， 其中的接入装置 920将上行物 理层载荷复用后转换为物理层信号传输至网关侧，以及将网关侧的下 行信号传输至目的接入节点 910。
在图 2的第一种以太网接入系统中 ,接入装置 920直接与网关设 备 901连接， 这种系统结构中不需对现有的网关设 ^故改动； 图 3所 示的第二种以太网接入系统中，接入装置 920与复用网关设备 902连 接，复用网关设备 902与接入装置 920釆用相配合的复用技术来实现 接入； 图 4所示的第三种以太网接入系统中，接入装置 920与网关设 备 901之间增加接入转换装置 930,作为接入装置 920与网关设备 901 间的接口， 使得接入装置 920的复用技术对网关设备 901透明。
图 5至图 8分别为本发明接入装置实施例一至实施例四的结构示 意图， 下行物理端口单元 111、 112至 Ilk分别连接至复用解复用单 元 120， 上行物理端口单元 130也与复用解复用单元 120连接。 需要 说明的是， 尽管图示了 k个物理端口单元， 上述四个实施例均支持两 个及两个以上的下行物理端口单元。
在这四个实施例中， 下行物理端口单元 111、 112至 Ilk可以各 与一个接入节点连接，分别从输入接入装置的各路低速物理层信号中 解析出其中承载的低速物理层载荷， 并输出至复用解复用单元 120; 复用解复用单元 120将各路低速物理层载荷复用为一路高速物理层 载荷， 输出到上行物理端口单元 130; 上行物理端口单元 130将高速 物理层载荷承载在高速物理层信号中发送出接入装置。上行物理端口 单元 130接收到高速物理层信号后，将从中解析出的高速物理层载荷 输出到复用解复用单元 120; 复用解复用单元 120将高速物理层载荷 解复用为对应于各个下行物理端口单元的低速物理层载荷，并将各路 低速物理层载荷输出到对应的下行物理端口单元，由接收低速物理层 载荷的下行物理端口单元将其转换为低速物理层信号后输出接入装 置。 实施例一至实施例四的不同之处在于因采用的复用技术不同，导 致复用解复用单元 120的内部实现不同。 其中， 实施例一与实施例二 按照到达下行物理端口单元的流量进行复用，通过物理层载荷中与下 行物理端口单元对应的标记进行解复用； 实施例三与实施例四按照时 序周期循环进行复用与解复用 ,时序周期被分解为对应于下行物理端 口单元的时隙， 每个时隙专用于对应的下行物理端口单元， 即使当时 对应的下行物理端口单元没有网絡流量也只用 idle (等待）信号填充 而不会用于对其他的下行物理端口单元的物理层载荷进行复用， idle 信号可以是符合以太网标准的 idle信号或通信双方约定的 idle信号。
本发明接入装置实施例一的结构示意图如图 5所示，复用解复用 单元 120包括标记存储模块 121和帧复用模块 122, 帧复用模块 122 分别与各个下行物理端口单元、标记存储模块 121以及上行物理端口 单元 130连接。
在本实施例中釆用数据帧中与接入节点具有——对应关系的字 段作为标记，由于每个接入节点通过一个下行物理端口单元连接接入 装置 ,因此通过每个下行物理端口单元传输的物理层载荷中该标记均 不相同。 本发明中数据帧指以太网标准的 MAC层数据帧， 包括在物 理层载荷中。标记可以是数据帧中接入节点的 MAC地址、 IP( Internet Protocol, 网际协议）地址等字段。
某个下行物理端口单元将接收的物理层信号转换为物理层载荷 发送至帧复用模块 122; 帧复用模块 122查看其中数据帧里的标记及 其对应的下行接口是否已经保存在标记存储模块 121中，如果未保存 或者发生变化则更新标记存储模块 121 , 使标记存储模块 121中维持 当前标记与下行物理端口单元的对应关系，之后将其转换为高速物理 层载荷并输出至上行物理端口单元 130; 上行物理端口单元 130将高 速物理层载荷承载在高速物理信号中发送出接入装置。
上行物理端口单元 130将接收的高速物理层信号转换为物理层 载荷发送至帧复用模块 122; 帧复用模块 122在标记存储模块 121中 查找到与高速物理层载荷里数据帧中的标记对应的下行物理端口单 元，将高速物理层载荷转换为低速物理层载荷后输出至该下行物理端 口单元；下行物理端口单元将低速物理层载荷承载在低速物理层信号 中发送出接入装置。
帧复用模块 122具有从物理层载荷中解析数据帧的功能，以及进 行标记学习的功能， 这两种功能可以采用现有的 MAC芯片中已有的 实现方法。对帧复用模块 122的高低速物理层载荷转换功能， 可以在 帧复用模块 122 中分别为每个下行物理端口单元和上行物理端口单 元 130开辟一段緩存区，通过高低速时钟以包括完整数据帧的物理层 载荷为单位进行传输速率转换。
采用本实施例中的这种复用方法，输入输出接入装置的物理层信 号中承载包括完全数据帧的物理层载荷， 符合以太网标准， 可以直接 与标准的网关设备连接， 适用于图 2所示的第一种以太网接入系统。 但这种复用方法需要接入装置解析数据帧， 而这一功能需要在 MAC 层实现， 因此接入装置要植入一些简单的二层处理能力。
图 6为本发明接入装置实施例二的结构示意图，复用解复用单元 120包括标记存储模块 121、 标记模块 123和码流复用模块 124， 标 记模块 123分别连接标记存储模块 121、 码流复用模块 124和各个下 行物理端口单元， 码流复用模块 124与上行物理端口单元 130连接。
标记存储模块 121 中存储着标记及与标记——对应的下行物理 端口单元，标记及其与下行物理端口单元的对应关系可以预先在标记 存储模块 121中设置。
某个下行物理端口单元将接收的低速物理层信号转换为低速物 理层载荷发送至标记模块 123; 标记模块 123在标记存储模块 121中 查找到与该下行物理端口单元对应的标记，在低速物理层载荷中添加 该标记后将其输出至码流复用模块 124; 码流复用模块 124将添加标 记后的低速物理层载荷转换为高速物理层载荷并输出至上行物理端 口单元 130; 上行物理端口单元 130将高速物理层载荷转换为高速物 理层信号后发送出接入装置。
上行物理端口单元 130将接收的高速物理层信号转换为物理层 载荷输出至码流复用模块 124; 码流复用模块 124将高速物理层载荷 转换为低速物理层载荷后输出到标记模块 123; 标记模块 123在标记 存储模块 121 中查找到与低速物理层载荷中标记对应的下行物理端 口单元，在清除低速物理层载荷中的标记后将其输出至该下行物理端 口单元；下行物理端口单元将低速物理层载荷承载在低速物理层信号 中发送出接入装置。
本实施例中，标记模块 123在物理层载荷中添加或清除标记可以 以数据帧为单位进行， 也可以以小于数据帧长度的某个固定长度进 行。在采用固定长度时， 考虑到来自或去向相同下行物理端口单元的 数据帧可能间隔一定时间， 为提供良好的 QoS ( Quanlity of Service, 服务质量）性能， 当到达数据帧末尾时， 即使还没有达到固定长度， 也可以进行复用处理。 标记可以添加在物理层载荷中的设定位置， 同 样也从该设定位置查找和清除标记即可。
因此， 本实施例中通常需要识别物理层载荷中数据帧的边界， 以 确定标记所应用的码流范围。但本实施例中不需要解析物理层载荷中 数据帧的字段； 另外， 由于某个过大的数据帧可能造成其他下行物理 端口单元处于等待状态较长时间，采用固定长度复用可以比以数据帧 为单位复用提供更好的 QoS性能。 需要补充说明的是： 一般来说对 于帧复用的情况， 都是采用每个低速端口轮流复用一帧的方式， 但为 了在一定程度上实现差异化的 QoS,可以改变复用的次序， 比如说对 于 A, B, C三个端口， A和 B轮流复用两帧后， 再复用 C的一帧， 而固定长度复用的情况也可以利用同样的方法， 改善效果更加明显。
本领域技术人员知道，码流复用模块 124的高低速物理层载荷转 换功能可以参照帧复用模块 122的方式实现。
在以数据帧为单位进行复用的情况下，推荐采用符合以太网标准 的 VLAN号作为标记， 这样在网关侧可以直接对物理层载荷进行处 理， 可以采用图 2所示的第一种接入系统。 同时， 由于 VLAN号配 置在接入装置上， 与接入节点无关， 不会因接入节点的 MAC地址或 IP地址变更而导致接入装置反复学习标记，可以保证接入装置工作的 稳定性。 并且随着多层 VLAN技术的成熟， VLAN的数目也不会成 为问题， 两层 VLAN就可以支持 4094*4094个节点的接入。
除以数据帧为单位复用并且釆用 VLAN号作为标记的情况以夕卜， 本实施例中的其他方案从上行物理端口 130输入输出的物理层信号 中承载的并非包括一个完整以太网数据帧的物理层载荷，因此适用于 图 3的第二种或图 4的第三种接入系统。
本发明接入装置实施例三的结构示意图如图 7所示，复用解复用 单元 120包括时序存储模块 125和定长复用模块 126, 定长复用模块 126分别与各个下行物理端口单元、 时序存储模块 125以及上行物理 端口单元 130连接。
本实施例时序周期中的每个时隙具有相同的长度。在时序存储模 块 125中，存储着时序周期中顺序排列的各个时隙所对应的下行物理 端口单元。 对各个下行物理端口单元具有相同带宽的情形， 可以令每 个物理端口单元对应于相同数量的时隙；而对下行物理端口单元带宽 不同的情形，可以令每个下行物理端口单元对应于与其带宽相匹配数 量的时隙。
某个下行物理端口单元将接收的低速物理层信号转换为低速物 理层载荷发送至定长复用模块 126。 定长复用模块 126将每一路输入 的低速物理层载荷写入对应于该下行物理端口单元的緩存区，同时以 时序周期中的时隙为单位，顺序从与时隙对应的下行物理端口单元缓 存区以高速时钟输出一定长度的物理层载荷；对当前没有网络流量的 下行物理端口单元， 以以太网标准的 idle信号或通信双方约定的 idle 信号填充其对应的时隙。 这样按照时序周期循环顺序， 定长复用模块 126将低速物理层载荷复用为一路高速物理层载荷， 其中包括对应于 各个下行物理端口单元的码流段。复用后的高速物理层载荷被输出到 上行物理端口单元 130, 由其将高速物理层载荷转换为高速物理层信 号后发送出接入装置。
上行物理端口单元 130将接收的高速物理层信号转换为物理层 载荷输出至定长复用模块 126。 定长复用模块 126将高速物理层载荷 写入高速緩存区， 同时以时序周期中的时隙为单位， 顺序从高速緩存 区以低速时钟向与该时隙对应的下行物理端口单元输出一定长度的 物理层载荷。 这样按照时序周期循环顺序， 定长复用模块 126将一路 高速物理层载荷解复用为多路低速物理层载荷。下行物理端口单元将 接收的低速物理层载荷承载在低速物理层信号中发送出接入装置。 本发明接入装置实施例四的结构示意图如图 8所示，复用解复用 单元 120包括时序与时长存储模块 127和变长复用模块 128, 变长复 用模块 128分別与各个下行物理端口单元、 时序与时长存储模块 127 以及上行物理端口单元 130连接。
在接入系统中节点用户可能采用不同的接入速率，本实施例比较 适用于这种各个下行物理端口单元可能具有不同带宽的情形。本实施 例与实施例三的不同之处是引入了时序周期中每个时隙的长度。在时 序与时长模块 127中，除了存储时序周期中顺序排列的各个时隙所对 应的下行物理端口单元，还要存储该时隙的长度， 时隙长度匹配于该 下行物理端口单元的带宽。
某个下行物理端口单元将接收的低速物理层信号转换为低速物 理层载荷发送至变长复用模块 128。 变长复用模块 128将每一路输入 的低速物理层载荷写入对应于该下行物理端口单元的緩存区，同时按 照时序周期中的时隙顺序，从每个时隙对应的下行物理端口单元緩存 区以高速时钟输出某个长度的物理层载荷，物理层载荷的长度对应于 该时隙的长度； 对当前没有网络流量的下行物理端口单元 , 以以太网 标准的 idle信号或通信双方约定的 idle信号填充其对应的时隙。这样 按照时序周期循环顺序，变长复用模块 128将低速物理层载荷复用为 一路高速物理层载荷，其中包括对应于各个下行物理端口单元的可能 具有不同长度的码流段。复用后的高速物理层载荷被输出到上行物理 端口单元 130, 由其将高速物理层载荷转换为高速物理层信号后发送 出接入装置。
上行物理端口单元 130将接收的高速物理层信号转换为物理层 载荷输出至变长复用模块 128。 变长复用模块 128将高速物理层载荷 写入高速緩存区， 同时按照时序周期中的时隙顺序， 从高速緩存区以 低速时钟向与该时隙对应的下行物理端口单元输出某个长度的物理 层载荷， 输出物理层载荷的长度对应于该时隙的长度。这样按照时序 周期循环顺序，变长复用模块 128将一路高速物理层载荷解复用为多 路低速物理层载荷。下行物理端口单元将接收的低速物理层载荷承载 在低速物理层信号中发送出接入装置。
实施例三和实施例四中的时隙实际上也可以看成是对物理层载 荷长度的衡量， 例如可以是物理层载荷的字节数。
实施例三和实施例四中，复用解复用单元 120不需要了解物理层 载荷的内容， 既不需识别也不需解析其中的数据帧。 另外， 与实施例 一和实施例二中上行带宽由所有下行连接共享不同，这两个实施例中 的上行带宽中分配给每个下行连接的部分带宽都得到严格保留，从物 理层上保证了 QoS 性能。 但是， 这两个实施例对上行物理端口 130 至网关设备侧的物理层传输有较高的要求，以避免一旦发生码流错位 可能导致的数据重传。
实施例三和实施例四中从上行物理端口 130输入输出的物理层 信号中承载的并非包括一个完整以太网数据帧的物理层载荷，因此适 用于图 3的第二种或图 4的第三种接入系统。
在实际应用中， 通常一个 PHY (物理层） 芯片上的各个物理端 口都具有相同的带宽， 同时很少在一个 PHY芯片上提供超过 8个的 物理端口， 即使将这些端口复用之后带宽也不足 100M, 以 100M为 单位进行远距离传输仍旧不能满足宽带接入的低成本需求。 此时，接 入装置可以采用级联复用结构， 其逻辑结构的一种示例如图 9所示。
图 9中， 复用单元按照其上下行带宽的不同分为多个复用级别， 最下级复用单元 220的下行接口连接最下级下行物理端口单元 210, 上行接口连接中间级复用单元 230; 中间級复用单元 230的上行接口 连接最上级复用单元 240的下行接口；最上级复用单元 240的上行接 口连接最上級上行物理端口单元 250。
最下级下行物理端口单元 210将接入节点的低速物理层信号转 换为低速物理层载荷， 将由各级复用单元 220、 230和 240逐级将低 速物理层载荷转换为高速物理层载荷后， 最上级上行物理端口单元 250将高速物理层载荷承载在物理层信号中输出。 最上级上行物理端 口单元 250从接收的高速物理层信号中解析出高速物理层载荷，经各 级复用单元 240、 230和 220逐级将高速物理层载荷解复用为对应于 最下级下行物理端口单元 210的低速物理层载荷并输出；最下级下行 物理端口单元 210将从复用单元 220接收的低速物理层载荷转换为低 速物理层信号后从接入装置输出。连接在同一复用单元 220或不同复 用单元 220上的最下级下行物理端口单元 210可以具有不同的带宽。
本发明中的接入装置支持两级及两级以上的复用单元级联复用。 根据所采用的复用技术不同， 各复用单元可以有不同的实现方法。接 入装置实施例五至实施例九具有相同的级联逻辑结构，其不同在于复 用单元的实现方法不同， 简便起见， 以下只针对复用单元来对实施例 五至实施例九进行说明。每个复用单元均包括至少两个下行接口和一 个上行接口， 下行接口用来进行至少两路物理层载荷的输入输出， 上 行接口用来输入输出下行接口多路物理层载荷的复用后物理层载荷 或解复用前物理层载荷。
实施例五与实施例七中的复用单元按照到达下行接口的流量进 行复用， 通过物理层载荷中与下行接口对应的标记进行解复用； 实施 例八与实施例九按照时序周期循环进行复用与解复用，时序周期被分 解为对应于下行物接口的时隙，每个时隙专用于对应的下行接口， 即 使当时对应的下行接口没有网絡流量也只用 idle信号在物理层填充 而不会用于传输其他下行接口的流量， 其中， idle信号可以是以太网 标准的 idle信号或通信双方约定的 idle信号。
接入装置实施例五中复用单元的结构如图 10 所示， 下行接口 311、 312至 31m分别与帧复用模块 402连接， 帧复用模块 402还与 标记存储模块 401和上行接口 320分别连接。
与接入装置实施例一中相同，本实施例中复用单元采用数据帧中 与接入节点具有——对应关系的字段作为标记，由于每个接入节点通 过一个最下级下行物理端口单元连接最下级复用单元的下行接口，因 此对每级复用单元而言，通过任一个下行接口传输的物理层载荷中该 标记均与通过其他下行接口的物理层载荷不同。标记可以是数据帧中 接入节点的 MAC地址、 IP地址等字段。
本实施例中各级复用单元均可以采用图 10中的结构， 对从下行 接口输入的多路物理层载荷，帧复用模块 402查看其中数据帧里的标 记及其对应的下行接口是否已经保存在标记存储模块 401中，如果未 保存或者发生变化则更新标记存储模块 401 , 使标记存储模块 401中 维持当前的标记与下行接口的对应关系 ,之后将其转换为高速物理层 载荷自上行接口 320输出复用单元。
对从上行接口 320接收的一路物理层载荷，帧复用模块 402在标 记存储模块 401中查找到与其中数据帧里的标记对应的下行接口，将 一路物理层载荷转换为低速物理层载荷后从与标记对应的下行接口 输出复用单元。
帧复用模块 402可以参照接入装置实施例一的帧复用模块 122实 现。 经过逐级复用， 本实施例中输入输出接入装置的物理层信号中承 载的物理层载荷均符合以太网标准， 可以直接与标准的网关设备连 接， 适用于图 2所示的第一种以太网接入系统。
在接入装置实施例六中，最下级复用单元与其他各级复用单元具 有不同的结构。 最下级复用单元的结构如图 11 所示， 标记模块 403 分别与下行接口 311、 312至 31m、 标记存储模块 401和码流复用模 块 404连接， 码流复用模块 404与上行接口 320连接。 其他各级复用 单元的结构如图 12所示，码流上级复用模块 405分别与下行接口 311、 312至 31m、 标记存储模块 401和上行接口 320连接。 本实施例中采 用与接入装置的最下级物理端口单元具有——对应关系的标记。
请参见图 11， 最下级复用单元的标记存储模块 401 中存储着该 最下级复用单元所连接的最下级物理端口单元所对应的标记、以及该 标记与连接对应最下级物理端口单元的下行接口的映射关系。标记及 对应关系可以预先设置于标记存储模块 401中。
在最下级复用单元中， 对从某个下行接口接收的物理层载荷， 标 记模块 403在标记存储模块 401中查找到与该下行接口对应的标记， 在接收的物理层载荷中添加该标记后将其输出至码流复用模块 404; 码流复用模块 404对添加标记后的物理层载荷进行下行至上行的传 输速率转换，将速率转换后的物理层载荷自上行接口 320输出至上级 复用单元。 对从上行接口 320接收的物理层载荷， 码流复用模块 404 对接收的物理层载荷进行上行至下行的传输速率转换，将速率转换后 的物理层载荷后输出到标记模块 403; 标记模块 403在标记存储模块 401中查找到与速率转换后物理层载荷中标记对应的下行接口， 在清 除物理层载荷中的标记后将其从该下行接口输出复用单元。
请参见图 12, 其他各级的标记存储模块 401 中存储着该复用单 元连接的最下级物理端口单元所对应的标记以及该标记与连接对应 最下级物理端口单元的下行接口的映射关系。 可见， 复用级别越高的 复用单元， 其下行接口所对应的标记往往越多。 标记及对应关系可以 预先设置于标记存储模块 401中。
在其他各級复用单元中， 对从某个下行接口接收的物理层载荷， 该物理层载荷已经由最下级复用单元添加了标记，码流上级复用模块 405对其进行下行至上行的传输速率转换， 将速率转换后的物理层载 荷自上行接口 320输出本复用单元。从上行接口 320接收到物理层载 荷后，码流上级复用模块 405在标记存储模块 401中查找到与其中标 记对应的下行接口，并对接收的物理层载荷进行上行至下行的传输速 率转换， 将速率转换后的物理层载荷从该下行接口输出本复用单元。
本实施例中，最下级复用单元在物理层载荷中添加的标记将穿越 各个上级复用单元发送出接入装置；而接入装置从网关设备侧接收的 物理层载荷中包括的标识供各个上级复用单元确定将解复用后的物 理层载荷从哪个下行接口输出，在输出最下级复用单元前标记才被清 除。
本实施例中 ,最下级复用单元的标记模块 403在物理层载荷中添 加或清除标记可以以数据帧为单位进行，也可以以小于数据帧长度的 某个固定长度进行，但所有的最下级复用单元必须采用同样的复用方 式。 在采用固定长度时， 为提供良好的 QoS性能， 当到达数据帧末 尾时， 即使还没有达到固定长度， 也可以进行复用处理。 标记可以添 加在物理层载荷中的设定位置，同样也从该设定位置查找和清除标记 即可。
因此，本实施例中最下级复用单元的标记模块 403通常需要识别 物理层载荷中数据帧的边界， 以确定标记所应用的码流范围。但本实 施例中不需要解析物理层载荷中 t据帧的字段； 另外， 由于某个过大 的数据帧可能造成其他下行物理端口单元处于等待状态较长时间，采 用固定长度复用可以比以数据帧为单位复用提供更好的 QoS性能。 本实施例中最下级复用单元的码流复用模块 404和其他各级码 流上级复用模块 405 的实现可以参照接入复用装置实施例一中的帧 复用模块 122实现。
在以数据帧为单位进行复用的情况下，推荐采用符合以太网标准 的 VLAN号作为标记， 这样在网关侧可以直接对物理层载荷进行处 理， 可以采用图 2所示的第一种接入系统。 同时， VLAN号是服务提 供商在接入装置上配置的， 接入装置对 VLAN号的添加和剥除， 接 入节点是无法感知的，因此不会像 MAC地址和 IP地址那样可以被接 入节点修改从而导致接入装置反复学习标记， 从而可以保证稳定性。 并且随着多层 VLAN技术的成熟， VLAN数目也不是大的问题， 两 层 VLAN就可以支持 4094*4094个节点的接入。
除以数据帧为单位复用并且采用 VLAN号作为标记的情况以外， 本实施例的其他方案中从最上级上行物理端口输入输出的物理层信 号中承载的并非标准的以太网物理层载荷，适用于图 3的第二种或图 4的第三种接入系统。
接入装置实施例七中复用单元的结构如图 13所示， 偏移标记模 块 407分别连接下行接口 311、 312至 31m、 偏移标记存储模块 406 和码流复用模块 404， 码流复用模块 404与上行接口 320连接。 本实 施例中采用的标记包括每级复用单元的偏移标记，同级偏移标记在物 理层载荷中具有相同的位置和相同的位长度， 由该级复用单元使用， 与该复用单元的下行接口具有——对应关系；不同級单元的偏移标记 可以具有不同的位置和不同的位长度。
每个复用单元的偏移标记存储模块 406 中存储着本复用单元的 偏移标记及其与本复用单元下行接口的对应关系。
对从某级复用单元某个下行接口接收的物理层载荷，偏移标记模 块 407在偏移标记存储模块 406中查找到与该下行接口对应的偏移标 记，在接收的物理层载荷中本级偏移标记的位置添加该偏移标记后将 其输出至码流复用模块 404; 码流复用模块 404对添加偏移标记后的 物理层载荷进行下行至上行的传输速率转换，将速率转换后的物理层 载荷自上行接口 320从本复用单元输出。对从上行接口 320接收的物 理层载荷，码流复用模块 404对接收的物理层载荷进行上行至下行的 传输速率转换， 将速率转换后的物理层载荷后输出到偏移标记模块 407; 偏移标记模块 407在转换后物理层载荷中本级偏移标记的位置 取出偏移标记 ,在偏移标记存储模块 406中查找到与该偏移标记对应 的下行接口，在清除物理层载荷中的偏移标记后将其从该下行接口输 出复用单元。
本实施例中，上行的物理层载荷在逐级复用过程中逐级添加了各 级偏移标记。 由于每个偏移标记在其复用单元具有唯一性， 由各级偏 移标记組合而成的标记与最下级下行物理端口单元具有——对应关 系。而在下行物理层载荷的逐级解复用过程中逐级清除该级的偏移标 记， 至输出最下级复用单元的下行接口前清除所有的偏移标记。
与实施例六中相同，本实施例中在物理层载荷中添加或清除偏移 标记可以以数据帧为单位进行，也可以以小于数据帧长度的某个固定 长度进行， 但所有的复用单元必须采用同样的复用方式。 因此， 本实 施例中各級复用单元的偏移标记模块 406通常需要识别物理层载荷 中数据帧的边界， 以确定标记所应用的码流范围。
本实施例中复用单元的码流复用模块 404可以参照接入复用装 置实施例一中的帧复用模块 122实现。
在以数据帧为单位进行复用的情况下，推荐采用如下的方式来设 置各级偏移标记，使所有各级偏移标记的组合形成符合以太网标准的
VLAN号。 或者采用多层 VLAN技术， 每一级偏移标记使用一层。 这样在网关侧可以直接对物理层载荷进行处理，此时本实施例采用图
2所示的第一种接入系统。
除上述各级偏移标记的组合形成 VLAN号的情况以外， 本实施 例中的其他方案适用于图 3的第二种或图 4的第三种接入系统。
接入装置实施例八中复用单元的结构如图 14所示， 定长复用模 块 409分别与下行接口 311、 312至 31m、 时序存储模块 408和上行 接口 320连接。
本实施例中， 时序周期中的每个时隙具有相同的长度。在时序存 储模块 408中，存储着时序周期中顺序排列的各个时隙所对应的下行 接口。对各个下行接口具有相同带宽的情形， 可以令每个下行接口对 应于相同数量的时隙； 而对下行接口带宽不同的情形， 可以令每个下 行接口对应于与其带宽相匹配数量的时隙。
在每个复用单元中，定长复用模块 409将从每个下行接口接收到 的物理层载荷写入该下行接口的緩存区，同时以时序周期中的时隙为 单位，顺序从与时隙对应的下行接口緩存区以决定于上行接口带宽的 时钟输出一定长度的物理层载荷； 对当前没有网络流量的下行接口， 以以太网标准的 idle信号或通信双方约定的 idle信号填充其对应的时 隙。 这样按照时序周期循环顺序， 定长复用模块 409将下行接口的多 路物理层载荷复用为一路物理层载荷，本级复用后的物理层载荷包括 对应于本复用单元各个下行接口的码流段。定长复用模块 409将本级 复用后的物理层载荷自上行接口 320输出本复用单元。
对从上行接口 320接收的物理层载荷，定长复用模块 409将其写 入上行接口 320的緩存区， 同时以时序周期中的时隙为单位， 顺序从 上行接口 320 的緩存区以由决定于下行接口带宽的时钟向与该时隙 对应的下行接口输出一定长度的物理层载荷。这样按照时序周期循环 顺序，定长复用模块 409将上行接口 320的一路物理层载荷解复用为 下行接口的多路物理层载荷， 并从对应的下行接口输出本复用单元。
这样，从最下级复用单元下行接口上行传输的低速物理层载荷经 过逐级复用单元的复用后形成一路高速物理层载荷，每个最上级复用 单元的时序周期中生成的该路高速物理层载荷中包括对应于每个最 下級复用单元下行接口的码流段， 且这些码流段具有相同的长度， 其 排列顺序由每級复用单元的时序周期中对应于下行接口的时隙的顺 序决定。从最上级复用单元上行接口下行传输的一路高速物理层载荷 中 ,包括同样排列顺序的对应于每个最下级复用单元下行接口的码流 段， 该路高速物理层载荷由各级复用单元逐级解复用后， 从对应的最 下级复用单元下行接口输出。
接入装置实施例九中复用单元的结构如图 15所示， 变长复用模 块 411分别连接下行接口 311、 312至 31m、时序与时长存储模块 410 和上行接口 320。
在接入系统中节点用户可能采用不同的接入速率，本实施例比较 适用于这种各个最下级下行物理端口单元可能具有不同带宽的情 同的长度。 在时序与时长模块 410中， 除了存储时序周期中顺序排列 的各个时隙所对应的下行接口，还要存储该时隙的长度， 时隙长度匹 配于该下行接口的带宽。
变长复用模块 411 将自下行接口接收的多路物理层载荷写入该 下行接口的緩存区, 同时按时序周期中的时隙顺序， 从每个时隙对应 的下行接口緩存区以决定于上行接口带宽的时钟输出某个长度的物 理层载荷， 物理层载荷的长度对应于该时隙的长度； 对当前没有网络 流量的下行接口，在物理层上以以太网标准的 idle信号或通信双方约 定的 idle信号填充其对应的时隙。 这样按照时序周期循环顺序， 变长 复用模块 411 将下行接口的多路物理层载荷复用为上行接口的一路 物理层载荷，复用后的该路物理层载荷中包括对应于各个下行接口的 可能具有不同长度的码流段。变长复用模块 411将复用后的一路物理 层载荷从上行接口 320输出本复用单元。
对从上行接口 320接收的一路物理层载荷，变长复用模块 411将 该路物理层载荷写入上行接口緩存区，同时按照时序周期中的时隙顺 序，从上行接口緩存区以决定于下行接口带宽的时钟向与该时隙对应 的下行接口输出某个长度的物理层载荷，输出物理层载荷的长度对应 于该时隙的长度。 这样按照时序周期循环顺序， 变长复用模块 411将 上行接口的一路物理层载荷解复用为下行接口的多路物理层载荷，并 从对应的下行接口输出本复用单元。
实施例八和实施例九中的时隙实际上也可以看成是复用和解复 用时处理的物理层载荷长度， 例如可以是物理层载荷的字节数。
实施例八和实施例九中，复用单元既不需识别也不需解析物理层 载荷中的数据帧。 另外， 与实施例三和实施例四相同， 这两个实施例 中的上行带宽中分配给每个下行连接的部分带宽都得到严格保留，从 物理层上保证了 QoS性能。 实施例八和实施例九适用于图 3的第二 种或图 4的第三种接入系统。
如前所述， 级联复用主要针对采用复用芯片的情况。 参见图 16, 一般来说， 复用芯片可能包括以下三种： 第一种是 MAC接口复用芯 片 710, 即将一个复用单元 711封装在芯片中， 其上下行端口均输入 输出物理层载荷； 第二种是 PHY-MAC接口复用芯片 720, 将一个复 用单元 711、 连接复用单元每一个下行接口的下行物理端口单元 712 封装在芯片中， 上行端口输入输出物理层载荷， 下行端口输入输出物 理层信号； 第三种是 PHY接口复用芯片 730, 将一个复用单元 711、 连接复用单元每一个下行接口的下行物理端口单元 712 和连接复用 单元上行接口的上行物理端口单元 713封装在芯片中 ,上下行端口均 输入输出物理层信号。
当接入装置实施例五至实施例九中的复用单元采用 MAC接口复 用芯片时， 下级 MAC接口复用芯片的上行端口可以直接连接相邻的 上级 MAC接口复用芯片的下行接口； 当采用 PHY接口复用芯片时， 下级 PHY接口复用芯片的上行端口也可以直接连接相邻的上级 PHY 接口复用芯片的下行接口； 当采用 PHY-MAC接口复用芯片时， 下级 PHY-MAC接口复用芯片的上行端口需要通过与其上行端口的传输速 率匹配的上行物理端口单元才可以连接相邻的上级 PHY-MAC接口 复用芯片的下行接口。
当然， 这三种复用芯片也可以混合使用， 此时可能需要在上下级 复用芯片之间接入上行物理端口单元或下行物理端口单元进行信号 转换。
在图 4 所示的第三种以太网接入系统采用现有的网关设备进行 与外部网络的连接，由接入转换装置来作为接入装置与网关设备间的 接口。本发明中的接入转换装置包括至少一个接入侧物理端口单元来 连接接入装置， 还应包括一个网关侧物理端口单元来连接网关设备。 接入转换装置也需要采用与接入装置相配合的复用技术。
图 17至图 19分别为本发明中接入转换装置实施例一至实施例三 的结构示意图， 复用转换单元 520分别与接入侧物理端口单元 510、 网关侧物理端口单元 530相连接（需要说明的是， 网关侧表示靠近网 关方向， 并非表示一定和网关的端口直接连接， 本领域技术人员可以 理解的是，接入转换装置和网关之间仍然可以存在传统的以太网交换 机甚至一个二层以太环网）。 接入侧物理端口单元 510通过接入装置 连接接入节点，在接收到物理层信号时解析出其中承载的复用物理层 载荷， 输出至复用转换单元 520; 复用转换单元 520将复用物理层载 荷转换为以完整的数据帧为单位的物理层载荷后输出至网关侧物理 端口单元 530; 网关侧物理端口单元 530将以数据帧为单位的物理层 载荷承载在物理层信号中发送至网关设备。 网关侧物理端口单元 530 从网关设备接收到物理层信号后，将其解析为以数据帧为单位的物理 层载荷， 输出至复用转换单元 520; 复用转换单元 520将以数据帧为 单位的物理层载荷逆向转换为复用物理层载荷后输出至接入物理端 口单元 510; 接入侧物理端口单元 510将复用物理层载荷转换为物理 层信号后输出至接入装置。
接入转换装置实施例一至实施例三的不同之处在于因采用的复 用技术不同， 导致复用转换单元 520的内部实现不同。
图 17所示为接入转换装置实施例一的结构示意图。 复用转换单 元 520包括复用标记模块 521和复用标记处理模块 522， 复用标记处 理模块 522分別连接接入物理端口单元 510、 复用标记模块 521和网 关物理端口单元 530。
本实施例的接入转换装置可以与接入装置实施例二、六和七配合 使用。 实施例二、 六和七的接入装置的上行物理层载荷中均包括与连 接接入节点的下行物理端口单元，或者说与接入节点具有 对应关 系的标记。
当接入转换装置的接入侧物理端口单元 510接收到接入装置的 上行物理层信号时，解析出其中承载的物理层载荷输出至复用标记处 理模块 522。 复用标记处理模块 522取出物理层载荷中的标记， 根据 标记生成以完整的数据帧为单位的物理层载荷，并查看标记与数据帧 中的接入节点标识是否已经保存在复用标记模块 521中，如果未保存 或者发生变化则更新复用标记模块 521， 使复用标记模块 521中维持 当前的标记与接入节点标识的对应关系；复用标记处理模块 522并将 清除标记后以数据帧为单位的物理层载荷输出至网关侧物理端口单 元 530。
当接入装置以数据帧为单位在物理层载荷中添加标记时，复用标 记处理模块 522在清除标记后即可得到以数据帧为单位的物理层载 荷； 而当接入装置以固定长度码流为单位在物理层载荷中添加标记 时，复用标记处理模块 522可能需要将具有相同标记的物理层载荷重 新组合， 清除其中的标记后得到以数据帧为单位的物理层载荷。
接入节点标识为由该接入节点发送的报文或传输至该接入节点 的报文中与该接入节点具有对应关系的字段， 可以是接入节点的 MAC地址、 IP地址、 VLAN号等字段。 接入系统中接入节点在上电 后首先向网关设备侧发送报文，这样接入转换装置可以学习到接入节 点标识与标记的对应关系。
对从网关侧物理端口单元 530接收的以数据帧为单位的物理层 载荷，复用标记处理模块 522在复用标记模块 521中查找到与该数据 帧中的接入节点标识相对应的标记，在物理层载荷中添加该标记后输 出至接入侧物理端口单元 510。
对有超过 1个接入侧物理端口单元 510与复用标记处理模块 522 连接的接入转换装置，复用标记处理模块 522还可以将与接入节点标 识对应的接入侧物理端口单元 510—同存储在复用标记模块 521中。 对从网关侧物理端口单元 530接收的下行物理层载荷，复用标记处理 模块 522在查找与数据帧中接入节点标识对应的标记时，同时查找对 应的接入侧物理端口单元 510, 并且将添加了标记的下行物理层载荷 输出至该接入侧物理端口单元 510。
需要说明的是，本实施例中接入转换装置应采用与其连接的接入 装置相匹配的方式对标记进行处理。如果接入装置以数据帧为单位进 行复用，接入转换装置也以数据帧为单位处理标记； 如果接入装置以 固定长度进行复用，则接入转换装置也对固定长度的码流进行标记处 理， 并且， 标记在数据帧或固定长度码流中的位置也应相同。
当接入系统采用接入装置实施例六组网时，本实施例中的标记为 由接入装置的最下级复用单元在码流中添加并且穿越所有上级复用 单元的标记； 当接入系统采用接入装置实施例七组网时， 本实施例中 的标记包括所有级别复用单元的偏移标记。
图 18所示为接入转换装置实施例二的结构示意图。 复用转换单 元 520包括复用时序模块 523、 时序转换模块 524和数据帧识別模块 525, 时序转换模块 524分别连接接入侧物理端口单元 510、 复用时 序模块 523和数据帧识别模块 525 , 数据帧识别模块 525连接网关侧 物理端口单元 530。
本实施例的接入转换装置可以与接入装置实施例三和八配合使 用。实施例三和八的接入装置的上行物理层载荷均按照复用时序周期 复用， 复用时序周期中时隙的长度相同。 复用物理层载荷中均包括排 列顺序对应于连接接入节点的下行物理端口单元，或者说对应于接入 节点的码流段。
对从接入侧物理端口单元 510接收的复用物理层载荷，时序转换 模块 524以复用时序周期中的时隙为单位，按照时隙的排列顺序将复 用物理层载荷分段写入该时隙的上行緩存区。数据帧识别模块 525从 各个时隙上行緩存区的以完整的数据帧为单位取出物理层载荷，查看 数据帧中的接入节点标识与对应的时隙是否已经保存在复用时序模 块 523中， 如果未保存或者发生变化则更新复用时序模块 523 , 使复 用时序模块 523中维持当前的时隙与接入节点标识的对应关系；数据 帧识别模块 525 并将以数据帧为单位的物理层载荷输出至网关侧物 理端口单元 530。
对从网关侧物理端口单元 530接收的以数据帧为单位的物理层 载荷，数据帧识别模块 525在复用时序模块 523中查找到与该数据帧 中接入节点标识相对应的时隙，将以该数据帧为单位的物理层载荷后 写入该时隙的下行緩存区。时序转换模块 524以复用时序周期中的时 隙为单位，按照时隙的排列顺序从每个时隙的下行緩存区中向接入侧 物理端口单元 510输出一定长度的物理层载荷。
同样，本实施例中的接入转换装置应采用与其连接的接入装置相 匹配的时分复用方式，例如相匹配的复用时序周期和相匹配的时隙数 量和长度， 以保证对物理层载荷在传输两端进行正确的复用和解复 用。
图 19所示为接入转换装置实施例三的结构示意图。 复用转换单 元 520包括复用时序与时长模块 526、 时序与时长转换模块 527和数 据帧识别模块 525， 时序与时长转换模块 527分别连接接入侧物理端 口单元 510、 复用时序与时长模块 526和数据帧识别模块 525， 数据 帧识别模块 525连接网关侧物理端口单元 530。
本实施例的接入转换装置可以与接入装置实施例四和九配合使 用。实施例四和九的接入装置的上行物理层载荷均按照复用时序周期 复用， 复用时序周期中时隙的长度可能不同。 复用物理层载荷中均包 括排列顺序对应于连接接入节点的下行物理端口单元，或者说对应于 接入节点的码流段， 并且码流段的长度对应于时隙长度。
在复用时序与时长模块 526 中预先存储复用时序周期中各个时 隙的长度及排列顺序。
对从接入侧物理端口单元 510接收的复用物理层载荷，时序与时 长转换模块 527以复用时序周期中的时隙为单位，按照时隙的排列顺 序将复用物理层载荷中码流长度对应于该时隙长度的码流段写入该 时隙的上行緩存区。数据帧识别模块 525从各个时隙上行緩存区中以 完整的数据帧为单位取出一段物理层载荷，查看数据帧中的接入节点 标识与对应的时隙是否已经保存在复用时序与时长模块 526中，如果 未保存或者发生变化则更新复用时序与时长模块 526， 使复用时序与 时长模块 526中维持当前的时隙、时隙长度与接入节点标识的对应关 系；数据帧识别模块 525并将以数据帧为单位的物理层载荷输出至网 关侧物理端口单元 530。
对从网关侧物理端口单元 530接收的发送至接入节点的以数据 帧为单位的物理层载荷，数据帧识别模块 525在复用时序与时长模块 526中查找到与该数据帧中接入节点标识相对应的时隙， 将该物理层 载荷后写入该时隙的下行緩存区。时序与时长转换模块 527以复用时 序周期中的时隙为单位 ,按时隙的排列顺序从每个时隙的下行缓存区 中向接入侧物理端口单元 510输出码流长度对应于该时隙长度的物 理层载荷。 同样，本实施例中的接入转换装置应采用与其连接的接入装置相 匹配的时分复用方式， 例如相匹配的复用时序周期、相匹配的时隙数 量和时隙长度、 以及相匹配的排列顺序， 以保证对物理层载荷在传输 两端进行正确的复用和解复用。
在图 3所示的第二种以太网接入系统中，复用网关设备需要釆用 与接入装置相配合的复用技术来实现接入。本发明中的复用网关设备 可以通过将上述接入转换装置主要功能模块集成在现有的网关设备 中实现。 例如， 一种复用网关设备的实施例包括物理端口单元、 复用 接口单元和转发单元， 其中：
物理端口单元用来在输入输出复用网关设备的物理层信号与其 中承载的复用 MAC层码流之间进行转换；
复用接口单元用来将从物理端口单元接收的复用 MAC层码流转 换为数据帧并输出至转发单元，以及将从转发单元接收的数据帧逆向 转换为复用 MAC层码流并输出至物理端口单元；
转发单元用来进行数据帧转发。
简而言之，只需要把接入转换装置的两个物理端口单元替换为分 别替换为网关的物理端口单元和网关的转发单元即可，具体属于现有 技术部分的改动和配合， 如接口的变化等， 此处不再赘述。
需要说明的是， 本发明上述接入装置、 复用网关设备和接入转换 装置中所有的物理端口单元都具有转换物理层信号和物理层载荷的 功能， 这一功能对应于不同的带宽有不同的实现方法， 均可采用现有
PHY芯片中的技术实现。
图 20为本发明所述以太网上行复用接入方法的流程图。 在步驟
S11 , 在接入节点侧， 对从各个接入节点接收的低速物理层信号， 分 别从中解析出对应于各个接入节点的低速物理层载荷。
在步骤 S12, 将各个接入节点的低速物理层载荷复用为一路高速 物理层载荷。
在步骤 S13 , 将复用后的高速物理层载荷承载在高速物理层信号 中。
在步骤 S14, 将高速物理层信号进行上行传输。
上述上行复用接入流程中，步骤 S12中将低速物理层载荷复用为 一路高速物理层载荷可以是一级复用， 也可以是多级复用。 多级复用 分级进行， 先将接入节点分组， 将每个组中接入节点的物理层载荷复 用为每組一路物理层载荷，然后将每组一路的物理层载荷再次分组后 复用或直接复用，直到将所有接入节点的物理层载荷复用为一路物理 层载荷。
步骤 S12中在进行一级或多级复用时，可以釆用前述接入装置的 各个实施例中的上行复用技术， 此处仅作简单说明， 详细内容请参见 前述实施例。
当采用数据帧中与接入节点具有一一对应关系的字段，即接入节 点标识作为物理层载荷的标记时，步骤 S12之后的各个步骤都与现有 技术相同。 当采用时分复用技术时， 步驟 S12中在时序周期中每个与 复用前某路物理层载荷对应的时隙，将码流长度匹配于该时隙长度的 该路物理层载荷输出为复用后的一路物理层载荷。
偏移标记仅适用于多级复用的情况，步骤 S12中每级复用前在复 用前的物理层载荷中添加与本次复用的各路物理层载荷——对应的 偏移标记。
图 21为本发明所述以太网下行复用接入方法的流程图。 在步骤 S21 , 接收下行传输的高速物理层信号。
在步驟 S22, Μ矣收的高速物理层信号中解析出复用物理层载荷。 在步骤 S23， 将复用物理层载荷解复用为对应于接入节点的低速 物理层载荷。
在步骤 S24， 将解复用后的低速物理层载荷承载在低速物理层信 号中发送至对应的接入节点。
上述下行复用接入流程中，步骤 S23中将复用物理层载荷解复用 为对应于接入节点的物理层载荷可以是一级解复用 ,也可以是多级解 复用。 多级解复用分级进行， 先将复用物理层载荷解复用为多路物理 层载荷； 再将解复用后的各路物理层载荷分別再次解复用， 直至解复 用为对应于接入节点的物理层载荷。
步骤 S23中在进行一级或多级解复用时，可以采用前述接入装置 的各个实施例中的上行复用技术， 此处仅作简单说明， 详细内容请参 见前述实施例。 当采用数据帧中与接入节点具有一一对应关系的字段，即接入节 点标识作为物理层载荷的标记时，步驟 S23中 居接入节点标识来进 行解复用。当根据复用物理层载荷中与接入节点——对应的标记进行 解复用时，在将对应于接入节点的物理层载荷转换为低速物理层信号 前应去除码流中的标记。 当采用时分复用技术时， 步驟 S23中在时序 周期中每个与复用前某路物理层载荷对应的时隙 ,将码流长度匹配于 该时隙长度的解复用前物理层载荷输出为对应于该时隙的一路解复 用后物理层载荷。
偏移标记仅适用于多级复用的情况，步骤 S23中每级解复用根据 解复用前物理层载荷中与解复用后的各路物理层载荷具有一一对应 关系的偏移标记进行，并在每次解复用时清除物理层载荷中据以进行 本次解复用的偏移标记。
对本发明而言， 接入节点可以采用现有的标准以太网物理层速 率， 当然若未来出现非标准的以太网物理层传输速率 , 如 2M、 40M、 80M等，只要该物理链路两端均支持这些非标准的传输速率即可。 同 样接入装置中输出高速物理层信号的上行物理端口也可以不是标准 的以太网物理层速率， 甚至部分物理端口都可能不是标准以太网口， 只要物理链路两端能够同时支持， 并能够承载以太网物理层载荷， 那 么本发明同样适用， 从更广意义上讲， 本发明并不关心每一段物理链 路的两端的具体物理层处理，相信这一点这对本领域技术人员是容易 理解的。
本发明中， 接入系统中接入节点到网关设备的组网结构非常简 单，接入节点的流量经过接入装置复用后送到上层设备处理。相对于 现有的接入系统， 本发明不需要做二层转发处理， 仅仅工作在一层， 因此可靠性上比通过二层交换机接入更有优势，同时随着基于本发明 相关实施方式的商用规模的扩大，本发明接入装置的成本有望做到一 台普通交换机的十分之一左右， 并且占用空间非常小，很适合放在小 区楼道内。 对运营商而言， 接入设备使用可靠且成本非常低廉； 对用 户而言， 可以通过计算机网卡直接接入， 不需要调制解调设备和拨号 软件， 在提供使用便利性的同时降低了用户的成本。对用户数量多且 密集分布的情况， 则采用级联复用的接入装置将更为适用。
本发明通过融合复用技术和以太网技术，提供了一种以太网宽带 接入系统的组网方案， 能够大幅降低宽带接入成本， 简化接入网络架 构， 提供高速可靠的接入速率。
以上所述的本发明实施方式， 并不构成对本发明保护范围的限 定。任何在本发明的精神和原则之内所作的任何修改、等同替换和改 进等， 均应包含在本发明的权利要求保护范围之内。

Claims

权 利 要 求
1. 一种以太网接入装置， 位于多个接入节点和网关设备之间， 其特征在于， 包括至少两个下行物理端口单元、 复用解复用单元和上 行物理端口单元， 其中：
所述下行物理端口单元用来在输入输出接入装置的低速物理层 信号与其中承载的低速物理层载荷之间进行转换；
所述上行物理端口单元用来在输入输出接入装置的高速物理层 信号与其中承载的高速物理层载荷之间进行转换；
所述复用解复用单元用来在下行物理端口单元的低速物理层载 荷与上行物理端口单元的高速物理层载荷之间进行复用与解复用。
2. 如权利要求 1所述的以太网接入装置， 其特征在于， 所述复 用解复用单元的复用与解复用根据物理层载荷中与下行物理端口单 元对应的标记进行。
3. 如权利要求 2所述的以太网接入装置， 其特征在于： 所述标 记为物理层载荷中数据帧内的字段，该字段与下行物理端口单元具有 对应关系；
所述复用解复用单元包括标记存储模块和帧复用模块， 其中： 所述标记存储模块用来存储标记及其与下行物理端口单元的对 应关系；
所述帧复用模块用来将下行物理端口单元输入的低速物理层载 荷以数据帧为单位转换为高速物理层载荷输出至上行物理端口单元， 以及将从上行物理端口单元输入的高速物理层载荷以数据帧为单位 转换为低速物理层载荷并输出至其中标记所对应的下行物理端口单 元。
4. 如权利要求 2所述的以太网接入装置， 其特征在于： 所述复 用解复用单元包括标记存储模块、 标记模块和码流复用模块， 其中： 所述标记存储模块用来存储标记及其与下行物理端口单元的对 应关系；
所述标记模块用来在从下行物理端口单元输入的 4氏速物理层载 荷中添加与该下行物理端口单元对应的标记并输出至码流复用模块， 以及在从码流复用模块输入的物理层载荷中去除标记并输出至与该 标记对应的下行物理端口 元；
所述码流复用模块用来在标记模块的低速物理层载荷与上行物 理端口单元的高速物理层载荷之间进行转换。
5. 如权利要求 4所述的以太网接入装置， 其特征在于： 所述标 记模块在物理层载荷中添加和去除标记以数据帧为单位或以固定长 度码流为单位进行。
6. 如权利要求 1所述的以太网接入装置， 其特征在于： 所述复 用解复用单元的复用与解复用根据时序周期循环进行，时序周期中包 括对应于各个下行物理端口单元的时隙。
7. 如权利要求 6所述的以太网接入装置， 其特征在于： 所述时 序周期中对应于各个下行物理端口单元的时隙长度相同；
所述复用解复用单元包括时序存储模块和定长复用模块， 其中： 所述时序存储模块用来存储时序周期中时隙与下行物理端口单 元的对应关系；
所述定长复用模块用来在各个时隙将从对应的下行物理端口单 元接收的低速物理层载荷转换为一路高速物理层载荷输出至上行物 理端口单元，以及将从上行物理端口单元接收的高速物理层载荷转换 为低速物理层载荷在各个时隙输出至对应的下行物理端口单元。
8. 如权利要求 6所述的以太网接入装置， 其特征在于： 所述复 用解复用单元包括时序与时长存储模块和变长复用模块， 其中： 所述时序与时长存储模块用来存储时序周期中各时隙的长度、该 时隙与下行物理端口单元的对应关系；
所述变长复用模块用来在各个时隙将从对应的下行物理端口单 元接收的码流长度匹配于该时隙长度的低速物理层载荷转换为一路 高速物理层载荷并输出至上行物理端口单元，以及将从上行物理端口 单元接收的高速物理层载荷转换为低速物理层载荷，在各个时隙将码 流长度匹配于该时隙长度的低速物理层载荷输出至对应的下行物理 端口单元。
9. 一种以太网接入装置， 位于多个接入节点和网关设备之间， 其特征在于， 包括多个最下级下行物理端口单元、 最上级上行物理端 口单元和至少两级复用单元，下级复用单元的上行接口与上级复用单 元的下行接口连接， 其中：
所述最下级下行物理端口单元连接最下级复用单元的下行接口， 用来在输入输出接入装置的低速物理层信号与其中承载的低速物理 层载荷之间进行转换；
所述最上级上行物理端口单元连接最上级复用单元的上行接口， 用来在输入输出接入装置的高速物理层信号与其中承载的高速物理 层载荷之间进行转换；
所述至少两级复用单元在最下级复用单元下行接口的低速物理 层载荷与最上级复用单元上行接口的高速物理层载荷之间进行逐级 的复用与解复用。
10. 如权利要求 9所述的以太网接入装置， 其特征在于： 所述各 级复用单元的复用与解复用根据物理层载荷中与最下级复用单元下 行接口对应的标记进行。
11. 如权利要求 10所述的以太网接入装置， 其特征在于： 所述 标记为最下级复用单元下行接口的物理层载荷中数据帧内区别于其 他下行接口的字段；
所述每级复用单元包括标记存储模块和帧复用模块， 其中： 所述标记存储模块用来存储标记及其与本复用单元下行接口的 对应关系；
所述帧复用模块用来将下行接口输入的多路物理层载荷以数据 帧为单位转换为一路物理层载荷后由上行接口输出，以及将从上行接 口输入的一路物理层载荷以数据帧为单位转换为多路物理层载荷并 由其中标记所对应的下行接口输出。
12. 如权利要求 10所述的以太网接入装置， 其特征在于： 所述 最下级复用单元包括标记存储模块、标记模块和码流复用模块，其中： 所述标记存储模块用来存储标记及其与本复用单元下行接口的 对应关系； 所述标记模块用来在从下行接口输入的物理层载荷中添加与该 下行接口对应的标记并输出至码流复用模块，以及在从码流复用模块 输入的物理层载荷中去除标记并从与该标记对应的下行接口输出； 所述码流复用模块用来在标记模块的多路物理层载荷与上行接 口的一路物理层载荷之间进行转换；
所述最下级复用单元之外的其他各级复用单元包括标记存储模 块和码流上级复用模块，其中所述码流上级复用模块用来将从下行接 口输入的多路物理层载荷转换为一路物理层载荷从上行接口输出，以 及将从上行接口输入的一路物理层载荷转换为多路物理层载荷并从 与其中标记对应的下行接口输出。
13. 如权利要求 12所述的以太网接入装置， 其特征在于： 所述 最下级复用单元的标记模块在物理层载荷中添加和去除标记以数据 帧为单位或以固定长度码流为单位进行。
14. 如权利要求 10所述的以太网接入装置， 其特征在于： 所述 标记包括每级复用单元的偏移标记；
所述各級复用单元包括偏移标记存储模块、偏移标记模块和码流 复用模块， 其中：
所述偏移标记存储模块用来存储本复用单元的偏移标记及其与 本复用单元下行接口的对应关系；
所述偏移标记模块用来在从下行接口输入的物理层载荷中添加 与下行接口对应的偏移标记并输出至码流复用模块，以及在从码流复 用模块输入的物理层载荷中去除本复用单元的偏移标记并从与该偏 移标记对应的下行接口输出；
所述码流复用模块用来在偏移标记模块的多路物理层载荷与上 行接口的一路物理层载荷之间进行转换。
15. 如权利要求 14所述的以太网接入装置， 其特征在于： 所述 各级复用单元的偏移标记模块在物理层载荷中添加和去除偏移标记 以数据帧为单位或以固定长度码流为单位进行。
16. 如权利要求 9所述的以太网接入装置， 其特征在于： 所述各 级复用单元的复用与解复用根据该复用单元的时序周期循环进行，时 序周期中包括对应于该复用单元下行接口的时隙。
17. 如权利要求 16所述的以太网接入装置， 其特征在于： 所述 每个复用单元的时序周期中的时隙具有相同的长度；
所述每级复用单元包括时序存储模块和定长复用模块， 其中： 所述时序存储模块用来存储该复用单元的时序周期中时隙与下 行接口的对应关系；
所述定长复用模块在各个时隙将从对应的下行接口接收的多路 物理层载荷转换为一路物理层载荷后从上行接口输出，以及将从上行 接口接收的一路物理层载荷转换为多路物理层载荷后在各个时隙由 对应的下行接口输出。
18. 如权利要求 16所述的以太网级联复用装置， 其特征在于， 所述各级复用单元包括时序与时长存储模块和变长复用模块， 其中： 所述时序与时长存储模块用来存储该复用单元的时序周期中时 隙的长度、 该时隙与下行接口的对应关系；
所述变长复用模块用来在各个时隙将从对应的下行接口接收的 码流长度匹配于该时隙长度的物理层载荷转换为一路物理层载荷后 从上行接口输出，以及将从上行接口接收的一路物理层载荷进行速率 转换后，在各个时隙将码流长度匹配于该时隙长度的物理层载荷从对 应的下行接口输出。
19. 如权利要求 9至 18所述的以太网接入装置， 其特征在于： 所述接入装置还包括除最下级以外的各级下行物理端口单元 ,分别连 接该级复用单元的每个下^ f亍接口，用来进行物理层信号与物理层载荷 之间的转换；每个复用单元与其连接的下行物理端口单元封装在一个 物理层-媒介接入控制层 PHY-MAC接口复用芯片中；
所述接入装置还包括除最上级以外的各级上行物理端口单元，连 接在下级 PHY-MAC接口复用芯片与上级 PHY-MAC接口复用芯片之 间， 用来进行物理层信号与物理层载荷之间的转换。
20. 如权利要求 9至 18所述的以太网接入装置， 其特征在于： 所述接入装置还包括除最下级以外的各级下行物理端口单元，分别连 接该级复用单元的每个下行接口 ,用来进行物理层信号与物理层载荷 之间的转换；
所述接入装置还包括除最上级以外的各级上行物理端口单元，分 别连接该级复用单元的每个上行接口，用来进行物理层信号与物理层 载荷之间的转换；
每个复用单元与其连接的下行物理端口单元和上行物理端口单 元封装在一个 PHY接口复用芯片中，下级 PHY接口复用芯片的上行 物理端口单元连接至上级 PHY接口复用芯片的下行物理端口单元。
21. 一种以太网接入方法， 其特征在于， 包括以下步驟： 将至少两路接入节点的物理层信号中承载的物理层载荷复用为 一路物理层载荷；
将复用后的物理层载荷转换为高速物理层信号后进行上行传输； 将下行高速物理层信号中承载的一路复用物理层载荷解复用为 与至少两个接入节点对应的物理层载荷；
将解复用后的物理层载荷转换为低速物理层信号后进行下行传 输。
22. 如权利要求 21所述的以太网接入方法, 其特征在于， 所述 将接入节点的物理层载荷复用为一路物理层载荷具体为：将来自接入 节点的多路物理层载荷分组复用为每组一路物理层载荷；将每组一路 物理层载荷再次分组后复用或直接复用， 直至复用为一路物理层载 荷；
所述将一路复用物理层载荷解复用为对应于接入节点的物理层 载荷具体为： 将复用物理层载荷解复用为多路物理层载荷； 将解复用 后的各路物理层载荷分别再次解复用，直至解复用为对应于接入节点 的物理层载荷。
23. 如权利要求 22所述的以太网接入方法， 其特征在于， 所述 方法在每次复用前还包括：在复用前的物理层载荷中添加与本次复用 的各路物理层载荷——对应的偏移标记；
所述每次解复用根据解复用前物理层载荷中与解复用后的各路 物理层载荷具有——对应关系的偏移标记进行；
所述方法在每次解复用时还包括：去除物理层载荷中据以进行本 次解复用的偏移标记。
24. 如权利要求 21或 22所述的以太网接入方法， 其特征在于， 所述方法在进行物理层载荷复用前还包括：在接入节点的物理层载荷 中添加与接入节点具有对应关系的标记；
所述将一路物理层载荷解复用为多路物理层载荷根据物理层载 荷中与接入节点具有对应关系的标记进行；
所述方法在将对应于接入节点的物理层载荷转换为低速物理层 信号前还包括： 在对应于接入节点的物理层载荷中去除所述标记。
25. 如权利要求 21或 22所述的以太网接入方法， 其特征在于， 所述将多路物理层载荷复用为一路物理层载荷按照时序周期循环进 行； 所述时序周期包括对应于复用前各路物理层载荷的时隙， 在每个 时隙将对应的复用前物理层载荷输出为复用后的一路物理层载荷，所 述复用前物理层载荷的长度匹配于该时隙的长度；
所述将一路物理层载荷解复用为多路物理层载荷按照时序周期 循环进行； 所述时序周期包括对应于解复用后各路物理层载荷的时 隙，在每个时隙将解复用前码流长度匹配于该时隙长度的物理层载荷 输出为对应的解复用后物理层载荷。
26. 如权利要求 21或 22所述的以太网接入方法， 其特征在于： 所述将一路物理层载荷解复用为多路物理层载荷根据物理层载荷的 数据帧中与接入节点具有对应关系的字段进行。
27. 一种以太网接入转换装置， 其特征在于， 位于接入装置和网 关设备之间； 包括接入侧物理端口单元、 复用转换单元和网关侧物理 端口单元； 其中，
所述接入侧物理端口单元用来在接入转换装置和接入装置之间 传输的物理层信号和其中承载的复用物理层载荷之间进行转换； 所述网关侧物理端口单元用来在接入转换装置和网关设备之间 传输的物理层信号及其中承载的以数据帧为单位的物理层载荷之间 进行转换；
所述复用转换单元用来在接入侧物理端口单元的复用物理层载 荷和网关侧物理接口单元的以数据帧为单位的物理层载荷之间进行 转换。
28. 如权利要求 27所述的以太网接入转换装置， 其特征在于， 所述复用转换单元包括复用标记模块和复用标记处理模块； 其中， 所述复用标记模块用于存储标记与接入节点标识的对应关系； 所述复用标记处理模块用于取出接入侧物理端口单元输入的物 理层载荷中的标记，才艮据标记生成以数据帧为单位的物理层载荷， 将 清除标记后的以数据帧为单位的物理层载荷输出至网关侧物理端口 单元； 对从网关侧物理端口单元接收的以数据帧为单位的物理层载 荷，在复用标记模块中查找与该数据帧中的接入节点标识相对应的标 记， 在物理层载荷中添加该标记后输出至接入侧物理端口单元。
29. 如权利要求 27所述的以太网接入转换装置， 其特征在于， 当接入装置以数据帧为单位在物理层载荷中添加标记时 ,所述复用标 记处理模块在清除标记后即可得到以数据帧为单位的物理层载荷； 当: 接入装置以固定长度码流为单位在物理层载荷中添加标记时，所述复 用标记处理模块将具有相同标记的物理层载荷重新组合，清除其中的 标记后得到以数据帧为单位的物理层载荷。
30. 如权利要求 27所述的以太网接入转换装置， 其特征在于， 所述复用转换单元包括复用时序模块、时序转换模块和数据帧识别模 块； 其中，
所述复用时序模块用来存储时序周期中时隙与接入节点标识的 对应关系；
所述时序转换模块用来以复用时序周期中的时隙为单位，按照时 隙的排列顺序将从接入侧物理端口单元接收的复用物理层载荷分段 写入该时隙的上行緩存区； 以复用时序周期中的时隙为单位，按照时 隙的排列顺序从每个时隙的下行緩存区中向接入侧物理端口单元输 出一定长度的物理层载荷；
所述数据帧识别模块用来从各个时隙上行緩存区以数据帧为单 位取出物理层载荷，将以数据帧为单位的物理层载荷输出至网关侧物 理端口单元；将从网关侧物理端口单元接收的以该数据帧为单位的物 理层载荷后写入该时隙的下行緩存区。
31. 如权利要求 27所述的以太网接入转换装置， 其特征在于， 所述复用转换单元包括复用时序与时长模块、 Erf序与时长转换模块和 数据帧识别模块； 其中，
所述复用时序与时长模块用来存储复用时序周期中各个时隙的 长度及排列顺序， 以及时隙与接入节点标识的对应关系；
所述时序与时长转换模块用来以复用时序周期中的时隙为单位， 按照时隙的排列顺序将从接入侧物理端口单元接收的复用物理层载 荷中码流长度对应于该时隙长度的码流段写入该时隙的上行緩存区； 以复用时序周期中的时隙为单位，按时隙的排列顺序从每个时隙的下 行緩存区中向接入侧物理端口单元输出码流长度对应于该时隙长度 的物理层载荷；
所述数据帧识别模块用来从各个时隙上行緩存区中以数据帧为 单位取出一段物理层载荷，将以数据帧为单位的物理层载荷输出至网 关侧物理端口单元；将从网关侧物理端口单元接收的发送至接入节点 的以数据帧为单位的物理层载荷写入该时隙的下行緩存区。
32. 一种复用网关设备， 其特征在于， 包括物理端口单元、 复用 接口单元和转发单元， 其中：
物理端口单元用来在输入输出复用网关设备的物理层信号与其 中承载的复用 MAC层码流之间进行转换；
复用接口单元用来将从物理端口单元接收的复用 MAC层码流转 换为数据帧并输出至转发单元，以及将从转发单元接收的数据帧逆向 转换为复用 MAC层码流并输出至物理端口单元；
转发单元用来进行数据帧转发。
PCT/CN2006/001138 2005-05-31 2006-05-29 Ethernet access device and method thereof WO2006128369A1 (en)

Priority Applications (3)

Application Number Priority Date Filing Date Title
US11/921,234 US7940812B2 (en) 2005-05-31 2006-05-29 Ethernet access device and Ethernet access method
JP2008513897A JP4819883B2 (ja) 2005-05-31 2006-05-29 イーサネット（登録商標）アクセス装置及びそのアクセス方法
EP20060742026 EP1887722A4 (en) 2005-05-31 2006-05-29 ETHERNET ACCESS DEVICE AND METHOD THEREFOR

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN200510073308.X 2005-05-31
CN200510073308 2005-05-31
CNB2006100786876A CN100477642C (zh) 2005-05-31 2006-04-30 以太网接入装置及其接入方法
CN200610078687.6 2006-04-30

Publications (1)

Publication Number Publication Date
WO2006128369A1 true WO2006128369A1 (en) 2006-12-07

Family

ID=37481228

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2006/001138 WO2006128369A1 (en) 2005-05-31 2006-05-29 Ethernet access device and method thereof

Country Status (5)

Country Link
US (1) US7940812B2 (zh)
EP (1) EP1887722A4 (zh)
JP (1) JP4819883B2 (zh)
CN (1) CN100477642C (zh)
WO (1) WO2006128369A1 (zh)

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5691543B2 (ja) * 2011-01-18 2015-04-01 富士通株式会社 光伝送装置
US9172597B2 (en) * 2011-04-28 2015-10-27 Invensys Systems, Inc. Data combiner and splitter
US8582437B2 (en) * 2011-06-21 2013-11-12 Broadcom Corporation System and method for increasing input/output speeds in a network switch
JP6113861B2 (ja) * 2012-12-05 2017-04-12 華為技術有限公司Ｈｕａｗｅｉ Ｔｅｃｈｎｏｌｏｇｉｅｓ Ｃｏ．，Ｌｔｄ． データ処理方法、通信ボードおよびデバイス
GB2509151B (en) * 2012-12-21 2016-07-20 Canon Kk Communication devices in a communication network and methods for processing data in such devices
US9485199B2 (en) * 2013-02-28 2016-11-01 Broadcom Corporation System and method for data flow identification and alignment in a 40/100 gigabit ethernet gearbox
ES2913444T3 (es) * 2013-03-21 2022-06-02 Huawei Tech Co Ltd Aparato de transmisión, método de transmisión de datos y soporte de grabación legible por ordenador no transitorio
WO2017166139A1 (zh) * 2016-03-30 2017-10-05 广东欧珀移动通信有限公司 中继传输的方法和装置
CN109698732B (zh) 2017-10-23 2021-07-09 华为技术有限公司 传输数据的方法和装置
CN110460455B (zh) * 2018-05-07 2022-05-31 华为技术有限公司 一种通信控制方法及控制器
CN110858811B (zh) * 2018-08-24 2022-01-18 华为技术有限公司 测量时延的方法和网络设备
CN109765867A (zh) * 2019-02-02 2019-05-17 宁波吉利汽车研究开发有限公司 生产控制及车身制造系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6188699B1 (en) * 1997-12-11 2001-02-13 Pmc-Sierra Ltd. Multi-channel encoder/decoder
CN1474550A (zh) * 2002-08-05 2004-02-11 北京润光泰力科技发展有限公司 100M以太网与2Mb/s电路的复用方法及其装置
KR100460672B1 (ko) * 2002-12-10 2004-12-09 한국전자통신연구원 10 기가비트 이더넷 회선 정합 장치 및 그 제어 방법
WO2005018120A1 (en) * 2003-08-12 2005-02-24 Teknovus, Inc. Method and apparatus for multiplexing ethernet channels

Family Cites Families (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5251210A (en) * 1991-11-01 1993-10-05 Ibm Corporation Method and apparatus for transforming low bandwidth telecommunications channels into a high bandwidth telecommunication channel
JPH0888623A (ja) * 1994-09-20 1996-04-02 Mitsubishi Electric Corp 通信装置における回線制御方式
JP2000224213A (ja) * 1999-01-28 2000-08-11 Mitsubishi Electric Corp 通信ネットワーク、および該通信ネットワークを構成するマスタ装置、スレーブ装置、多重化装置、並びに交換装置
JP2000224263A (ja) * 1999-01-29 2000-08-11 Ando Electric Co Ltd 通信システムの遠隔監視制御方法
US6690682B1 (en) * 1999-03-12 2004-02-10 Lucent Technologies Inc. Bit multiplexing of packet-based channels
US6498798B1 (en) * 1999-04-23 2002-12-24 Lucent Technologies Inc. Priority-based statistical multiplexer-hub
US6577649B1 (en) * 1999-11-12 2003-06-10 International Business Machines Corporation Multiplexer for asynchronous data
US6980563B2 (en) * 2000-04-13 2005-12-27 International Business Machines Corporation Method and system for fast ethernet serial port multiplexing to reduce I/O pin count
JP4168574B2 (ja) * 2000-06-02 2008-10-22 株式会社日立製作所 パケット転送装置、パケット転送制御方法、及びパケット転送装置の設定方法
US7095758B2 (en) * 2000-06-16 2006-08-22 Nippon Telegraph And Telephone Corporation Multiplexing and transmission apparatus
US7013338B1 (en) * 2000-07-28 2006-03-14 Prominence Networks, Inc. Multiplexing several individual application sessions over a pre-allocated reservation protocol session
US6788682B1 (en) * 2000-08-02 2004-09-07 3Com Corporation Mapping of packets between links of trunk groups using Mux/Demux devices
US6836486B2 (en) * 2000-12-21 2004-12-28 Nortel Networks Limited Switching of low order data structures using a high order switch
JP3654230B2 (ja) * 2001-10-18 2005-06-02 日本電気株式会社 通信システム
JP2003224574A (ja) * 2002-01-31 2003-08-08 Kyushu Ando Denki Kk 論理回線フロー制御方式
US7483432B2 (en) * 2002-09-23 2009-01-27 Alcatel Lucent Usa Inc. Packet transport arrangement for the transmission of multiplexed channelized packet signals
CN100446450C (zh) * 2002-11-07 2008-12-24 诺基亚公司 传输格式数据传输
US7711006B2 (en) * 2003-08-15 2010-05-04 Napatech A/S Data merge unit, a method of producing an interleaved data stream, a network analyser and a method of analysing a network
KR100576715B1 (ko) * 2003-12-23 2006-05-03 한국전자통신연구원 10기가비트 이더넷 프레임 다중화/역다중화 장치
US7590118B2 (en) * 2003-12-23 2009-09-15 Agere Systems Inc. Frame aggregation format

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6188699B1 (en) * 1997-12-11 2001-02-13 Pmc-Sierra Ltd. Multi-channel encoder/decoder
CN1474550A (zh) * 2002-08-05 2004-02-11 北京润光泰力科技发展有限公司 100M以太网与2Mb/s电路的复用方法及其装置
KR100460672B1 (ko) * 2002-12-10 2004-12-09 한국전자통신연구원 10 기가비트 이더넷 회선 정합 장치 및 그 제어 방법
WO2005018120A1 (en) * 2003-08-12 2005-02-24 Teknovus, Inc. Method and apparatus for multiplexing ethernet channels

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP1887722A4 *

Also Published As

Publication number Publication date
JP2008543187A (ja) 2008-11-27
US7940812B2 (en) 2011-05-10
CN1917471A (zh) 2007-02-21
EP1887722A1 (en) 2008-02-13
CN100477642C (zh) 2009-04-08
JP4819883B2 (ja) 2011-11-24
US20090232133A1 (en) 2009-09-17
EP1887722A4 (en) 2015-04-08

Similar Documents

Publication Publication Date Title
WO2006128369A1 (en) Ethernet access device and method thereof
US8289962B2 (en) Multi-component compatible data architecture
US8588209B2 (en) Multi-network compatible data architecture
US6847644B1 (en) Hybrid data transport scheme over optical networks
US6963561B1 (en) Facility for transporting TDM streams over an asynchronous ethernet network using internet protocol
US6587476B1 (en) Ethernet frame encapsulation over VDSL using HDLC
KR101290413B1 (ko) 인피니밴드 네트워크의 물리적 범위를 확장하기 위한 방법
US6771663B1 (en) Hybrid data transport scheme over optical networks
US20020083190A1 (en) Apparatus and method for GFP frame transfer
JP2008543187A5 (zh)
EP2019519B1 (en) Method for addressing ethernet streams with a structured GPON GEM Port ID
US20070121624A1 (en) Method and system of network clock generation with multiple phase locked loops
JP4476819B2 (ja) 放送／通信融合システム及び放送／通信融合システムにおけるイーサネット（登録商標、以下同じ）データの送受信方法
CN101179556B (zh) 一种光纤通道业务的传送方法和装置
US20070121638A1 (en) Method and system of communicating superframe data
US6959008B2 (en) Alignment of TDM-based signals for packet transmission using framed and unframed operations
US20070121623A1 (en) Method and system for establishing narrowband communications connections using virtual local area network identification
CN101072141B (zh) 以太网接入转换装置、复用网关设备及接入转换方法
JP2003078540A (ja) 多重化伝送システム及びそのデータ転送方法
CN101977186B (zh) 一种实现STM-1接口汇聚多路Ethernet over E1协议转换的装置
JP6154595B2 (ja) 回線エミュレーションシステム、装置および方法
WO2024032269A1 (zh) 通信方法、相关装置及计算机可读存储介质
EP3664371B1 (en) Switching method and apparatus
WO2008071007A1 (en) Method and system for multiplexing data for transmission
JP2003143152A (ja) 通信方法及びシステム、伝送装置、終端装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application
WWE Wipo information: entry into national phase

Ref document number: 2008513897

Country of ref document: JP

WWE Wipo information: entry into national phase

Ref document number: 11921234

Country of ref document: US

Ref document number: 2006742026

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Ref document number: DE

NENP Non-entry into the national phase

Ref country code: RU

WWW Wipo information: withdrawn in national office

Ref document number: RU

WWP Wipo information: published in national office

Ref document number: 2006742026

Country of ref document: EP