WO2011144028A2 - 矢量dsl的方法、系统及一种单板和dslam设备 - Google Patents
矢量dsl的方法、系统及一种单板和dslam设备 Download PDFInfo
- Publication number
- WO2011144028A2 WO2011144028A2 PCT/CN2011/074253 CN2011074253W WO2011144028A2 WO 2011144028 A2 WO2011144028 A2 WO 2011144028A2 CN 2011074253 W CN2011074253 W CN 2011074253W WO 2011144028 A2 WO2011144028 A2 WO 2011144028A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- board
- user
- data
- processing
- vector processor
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/005—Interface circuits for subscriber lines
- H04M3/007—Access interface units for simultaneous transmission of speech and data, e.g. digital subscriber line [DSL] access interface units
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M11/00—Telephonic communication systems specially adapted for combination with other electrical systems
- H04M11/06—Simultaneous speech and data transmission, e.g. telegraphic transmission over the same conductors
- H04M11/062—Simultaneous speech and data transmission, e.g. telegraphic transmission over the same conductors using different frequency bands for speech and other data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B3/00—Line transmission systems
- H04B3/02—Details
- H04B3/32—Reducing cross-talk, e.g. by compensating
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
- H04L12/2854—Wide area networks, e.g. public data networks
- H04L12/2856—Access arrangements, e.g. Internet access
- H04L12/2869—Operational details of access network equipments
- H04L12/2878—Access multiplexer, e.g. DSLAM
- H04L12/2892—Access multiplexer, e.g. DSLAM characterised by the access multiplexer architecture
- H04L12/2896—Distributed processing, e.g. on line cards
Definitions
- the present invention relates to the field of digital subscriber line DSL data processing, and more particularly to a method and system for vector DSL and a single board and a DSLAM device.
- xDSL is a general term for various types of DSL (Digital Subscribe Line).
- xDSL is a transmission technology for transmitting high-speed data over a telephone twisted pair (unshielded Twist Pair, UTP for short).
- UTP unshielded Twist Pair
- xDSLAM Digital Subscriber Line Access Multiplexer
- NEXT near-end crosstalk
- FEXT far-end crosstalk
- the CO side performs joint reception processing of signals.
- b is used to represent the noise vector of fxi, which is a column vector;
- y represents the channel output vector, which is a column vector of ⁇ .
- the channel output vector y the following is as follows:
- a joint transmission process of signals is performed on the CO side.
- the CO side needs to send a signal to the user to form a column vector X of xi.
- a vector precoding processor is introduced at the CO end, and the vector precoding processor generates a The precoding process matrix P, and P makes HP a diagonal matrix, the signal sent at this time is ⁇ , and its expression is as follows:
- a centralized vector DSL system is proposed in the prior art. As shown in FIG. 4, the system includes N boards, each board controls n users, and users of all boards Data is transferred to a centralized vector processor for data processing to eliminate crosstalk.
- the centralized vector processor combines the functions of a crosstalk cancellation processor and a vector precoding processor.
- each board sends data of n users passing through the first data processing unit to the centralized vector processor, so the centralized vector processor receives data Hx of all N*w users, and Hx is (N). * «)xl column vector; the centralized vector processor generates a (N* «)x(N* «) crosstalk cancellation processing matrix W, and the centralized vector processor needs to perform W*Hx operations due to the components of each vector Both are complex numbers, so 4 (N* «) 2 multiplication operations are required in this process.
- the centralized vector processor returns the processed data to the corresponding board for subsequent data processing.
- the centralized vector processor For the downstream direction, the centralized vector processor generates a precoding processing matrix P and performs a vector operation of ⁇ * ⁇ , which also requires 4 (N* «) 2 multiplication operations and 3(N* «) 2 -(N * «) Addition.
- the existing centralized vector DSL system Since the vector DSL technology performs precoding processing and crosstalk cancellation processing, and has high real-time requirements, the existing centralized vector DSL system is required to process the user data of all the boards in real time. Centralized vector processors need to have high data processing capabilities. The more users there are, the higher the data processing capability of the centralized vector processor is. The data processing capability of the existing centralized vector processor limits the user capacity of the vector DSL system.
- Embodiments of the present invention provide a method, system, and a single board and DSLAM device for vector DSL to increase the user capacity of a vector DSL system.
- a vector DSL method including:
- the at least one board receives and processes the user data of the respective boards in advance; the board transmits the user data of the board to the auxiliary vector processor on the board for internal vectorization processing to obtain internal processing data;
- the board transmits its user data to a centralized vector processor for external vectorization processing. Obtain externally processed data;
- the board receives the external processing data of the board and performs post processing in combination with the internal processing data.
- a single board including:
- An advanced processing unit configured to receive and process user data of the board in advance
- a first transmitting unit configured to transmit user data of the board to an auxiliary vector processor on the board;
- An auxiliary vector processor configured to perform internal vectorization processing on user data of the board, and output internal processing data
- a first receiving unit configured to receive internal processing data output by the auxiliary vector processor
- a second transmitting unit configured to transmit user data of the board to the centralized vector processor for external vectorization processing
- a second receiving unit configured to receive external processing data that is output after being externally vectorized by the centralized vector processor
- a post processing unit configured to post-process the internal processing data of the first receiving unit and the external processing data of the second receiving unit.
- a digital subscriber line access multiplexer DSLAM device includes at least two boards, at least one of which includes:
- An advanced processing unit configured to receive and process user data of the board in advance
- a first transmitting unit configured to transmit user data of the board to an auxiliary vector processor on the board;
- An auxiliary vector processor configured to perform internal vectorization processing on user data of the board, and output internal processing data
- a first receiving unit configured to receive internal processing data output by the auxiliary vector processor
- a second transmitting unit configured to transmit user data of the board to the centralized vector processor for external vectorization processing
- a second receiving unit configured to receive an external vectorization process by the centralized vector processor External processing data
- a post processing unit configured to post-process the internal processing data of the first receiving unit and the external processing data of the second receiving unit.
- a vector DSL system comprising: a DSLAM device; the DSLAM device includes at least two boards, at least one of which includes an auxiliary vector processor, and the auxiliary vector processor is used for user data of a board on which the board is located Internal vectorization processing is performed, and user data of a board including the auxiliary processor is externally vectorized by a centralized vector processor; the centralized vector processor is configured to perform user data on a board including the auxiliary vector processor The external vectorization process performs internal vectorization processing and external vectorization processing on the user data of the board that does not include the auxiliary vector processor.
- the method and system for vector DSL provided by the embodiment of the present invention, and a single board and a DSLAM device, by using an auxiliary vector processor on at least one board, using the auxiliary vector processor to perform user data of the board on which the board is located Internal vectorization processing, and external vectorization processing of the user data of the board by using a centralized vector processor; since the auxiliary vector processor shares a part of the data processing amount, when the number of system users is constant, Reduce the data processing capacity of the centralized vector processor, and increase the user capacity of the vector DSL system when the data processing capability of the centralized vector processor is constant; of course, if the data processing capability of the centralized vector processor can be improved, then The user capacity of the vector DSL system can be further increased by using the solution provided by the present invention.
- Figure 1 is a schematic diagram of near-end crosstalk and far-end crosstalk
- FIG. 2 is a schematic diagram of user uplink data to the CO end
- FIG. 3 is a schematic diagram of downlink data of the CO end to the user; 4 is a schematic diagram of a centralized vector DSL system provided by the prior art; FIG. 5 is a schematic diagram of a vector DSL system provided in Embodiment 1;
- FIG. 6 is a schematic diagram of another vector DSL system provided in Embodiment 1;
- Figure ⁇ is a schematic diagram of a method for vector DSL according to an embodiment of the present invention.
- FIG. 8 is a schematic diagram of another method for vector DSL according to an embodiment of the present invention
- FIG. 9 is a schematic diagram of a method for providing uplink data to a CO end of a user according to Embodiment 2;
- FIG. 10 is a schematic diagram of a method for providing a vector DSL for the case where the downlink data of the CO end is given to the user in the fourth embodiment;
- Figure 1 is a structural block diagram of a single board provided in Embodiment 6;
- FIG. 12 is a structural block diagram of another board provided in Embodiment 6.
- Embodiment 1 is a diagrammatic representation of Embodiment 1:
- an embodiment of the present invention provides a vector DSL system, including: a DSLAM device; the DSLAM device includes at least two boards, at least one of the boards includes an auxiliary vector processor, and the auxiliary vector
- the processor is configured to perform internal vectorization processing on the user data of the board on which the board is located, and the user data of the board including the auxiliary processor is externally vectorized by the centralized vector processor;
- the centralized vector processor is configured to perform external vectorization processing on user data of a board including the auxiliary vector processor, and perform internal vectorization processing and external vectorization processing on user data of the board not including the auxiliary vector processor.
- the centralized vector processor is disposed in the DSLAM device or is independent of the DSLAM device.
- FIG. 5 is an example in which the centralized vector processor is independent of the DSLAM device, but the location of the centralized vector processor is not limited to that shown in FIG. 5; for example, the centralized vector processor may be configured as a single board inserted in a slot of the DSLAM device.
- the number of boards in the DSLAM device is N (N ⁇ 2), which are respectively boards.
- the board 2 is the board N; the number of users of each board is n (n ⁇ 2), and the user of the board 1 is ⁇ ⁇ », and the user of the board 2 is ⁇ supervise + 1 ⁇ 2 propel The user of the single board is ⁇ ( ⁇ - 1 +1 ⁇ .
- at least one of the boards includes an auxiliary vector processor; in the vector DSL system shown in Figure 5, the board 1
- the board 2 includes an auxiliary vector processor as an example.
- the vector DSL system shown in FIG. 6 is an optimal case.
- Each board of the vector DSL system includes an auxiliary vector processor, The more vector processors, the more data processing is shared. If the data processing capability of the centralized vector processor is constant, the user capacity of the vector DSL system can be increased.
- the vectoring process includes precoding processing and/or crosstalk canceling processing.
- the vectorization process refers to crosstalk cancellation processing on the uplink user data.
- the internal vectorization process refers to crosstalk cancellation processing performed by the auxiliary vector processor on each board to eliminate crosstalk generated between the user and the user on the board.
- the internal vectorization process is also referred to as an internal crosstalk cancellation process;
- the external vector process refers to a crosstalk cancellation process performed by the centralized vector processor to eliminate crosstalk generated between users of each board and users of other boards.
- the external vectorization process is also referred to as an alien crosstalk cancellation process.
- the internal vectorization processing refers to the crosstalk cancellation processing for eliminating the crosstalk generated between the user ⁇ ⁇
- the external vectorization processing refers to eliminating the user ⁇ ⁇ other Crosstalk cancellation processing performed by crosstalk generated between users of the board.
- the vectorization processing refers to precoding processing on the downlink user data.
- the internal vectorization process refers to precoding processing performed by the auxiliary vector processor on each board to eliminate crosstalk generated between the user and the user on the board.
- the external vector processing refers to a precoding process performed by the centralized vector processor to eliminate crosstalk generated between users of each board and users of other boards.
- the external vectorization processing is also called For external precoding processing.
- the internal vectorization processing refers to the precoding processing for eliminating the crosstalk that may occur between the user, and the external vectorization processing refers to eliminating the user ⁇ ⁇ ⁇ other orders. Precoding processing performed by crosstalk between users of the board.
- the internal vectorization processing and the external vectorization processing of the user data of the board are all performed by a centralized vector processor.
- the auxiliary vector processor on each board in the vector DSL system provided by the present invention shares a part of the data processing amount, the data processing amount of the centralized vector processor can be reduced when the number of system users is constant. In the case where the data processing capability of the centralized vector processor is constant, the user capacity of the vector DSL system can be increased.
- the board that includes the auxiliary vector processor can also be used to select a subscriber line with other boards on the board. Strongly related users, and transmit user data on the board that is strongly related to the subscriber lines of other boards to the centralized vector processor.
- the two user lines are strong.
- the user and any other board if the board 1 on a user X; (n + l ⁇ i ⁇ N * n) line will produce crosstalk, the user ⁇ ⁇ ⁇ , strong line correlation.
- the board 1 has nl (nl ⁇ n) users, for example, X ⁇ Xw and any user X of other boards ; (n+l ⁇ i ⁇ N*n) lines may cause crosstalk, then the board The user who is strongly associated with the subscriber line of other boards is X ⁇ Xw.
- the user data on the board 1 that is strongly related to the subscriber lines of other boards is the data of the users X to X nl .
- the user who selects the card on the board that is strongly related to the user line of the other board may be a single The board is obtained based on experience or other means.
- the selection function of the foregoing board may also be completed by a centralized vector processor, and the centralized vector processor is further configured to select a user that is strongly related to the user lines of other boards on each board that includes the auxiliary vector processor. And controlling the board where the user is located to transmit the user data of the user to the centralized vector processor.
- the centralized vector processor may be integrated on any one of the at least two boards.
- the centralized vector processor may also be independent of the At least two veneers exist separately.
- the data transmitted between the board and the centralized vector processor may be user data that is strongly related to the user lines of other boards on each board, that is, may be part of the users of each board. Data, so the use of data transmission bandwidth can be reduced.
- this embodiment also provides a method of vector DSL. As shown in Figure 7, the method includes the following steps:
- At least one board receives and processes user data of the respective boards in advance;
- the method in this embodiment is for a board that includes an auxiliary vector processor, that is, the at least one board refers to at least one board that includes an auxiliary vector processor; and does not include an auxiliary vector processor.
- the processing method of the user data of the board is the same as the prior art, and the vector processing and the external vector processing are performed on the user data by the centralized vector processor, so
- the board transmits the user data of the board to an auxiliary vector processor on the board for internal vectorization processing to obtain internal processing data.
- the board transmits its user data to a centralized vector processor for external vectorization processing to obtain external processing data.
- the board receives external processing data of the board, and performs post processing in combination with the internal processing data.
- the vectoring process includes precoding processing and/or crosstalk canceling processing.
- the method of vector DSL provided by the present invention is due to the auxiliary vector at least one board
- the processor shares a part of the data processing capacity, so that the amount of data processing of the centralized vector processor can be reduced when the number of system users is constant, and the vector can be increased when the data processing capability of the centralized vector processor is constant.
- User capacity of the DSL system Of course, if the data processing capability of the centralized vector processor can be improved, the user capacity of the vector DSL system can be further increased by using the solution provided by the present invention.
- the method further includes:
- Step 705 The board selects a user that is strongly related to the user line of the other board on the board; or the centralized vector processor selects a strong relationship between the board and the user line of the other board. user.
- the board 703 may specifically: the board sends the user data strongly related to the user lines of the other boards to the centralized vector processor for external vectorization processing to obtain external processing data.
- the data transmitted by the at least one board to the centralized vector processor can be part of the user data of the board, bandwidth saving can be achieved.
- the user data of some boards is not particularly large. Instead of using this preferred method, all user data of these boards can still be transmitted to the centralized vector processor.
- Embodiment 2 is a diagrammatic representation of Embodiment 1:
- the method for the above vector DSL will be further described in detail in the embodiments of the present invention.
- the method provided by the embodiment of the present invention is still based on the vector DSL system shown in FIG. 6.
- the number of the boards is N (N ⁇ 2), which are respectively the board 1 and the board 2 board N; each board The number of users is n (n>2), and the user of the board 1 is ⁇ ⁇ ⁇ », the user of the board 2 is ⁇ note + 1 ⁇ 2 resort , and the user of the board X is X( w — 1 + 1; and each board contains an auxiliary vector processor 9, where the user data for uplink to the CO end, the present embodiment provides a method comprising:
- Step 901 The user sends the data to the board where the user is located.
- the user ⁇ ⁇ ⁇ » sends its data to the board 1
- the user ignore + 1 ⁇ 2 drives its data to the board 2
- the user X( w — 1) 3 ⁇ 4+1 ⁇ sends its data to the board ⁇ .
- Step 902 The pre-processing unit on each board performs pre-processing on the user data received by the board.
- the prior processing includes performing a fast Fourier transform on the received user data.
- the user data after fast Fourier transform on each board is expressed as follows:
- Step 903 The pre-processing unit of each board transmits the user data of the board to the auxiliary vector processor on the board.
- Step 904 The auxiliary vector processor performs internal crosstalk cancellation processing on the user data of the board to obtain internal processing data.
- the internal crosstalk cancellation process is specifically: the auxiliary vector processor of each board generates a "crosstalk cancellation processing matrix W1", and the matrix W1 performs matrix multiplication with the column vector of w 1 formed by the user data of the board.
- the column vector of w 1 is obtained as the internal processing data of the board.
- the matrix W1 generated by the auxiliary vector processor of each board is generated to eliminate crosstalk between users of the board. Therefore, the matrix W1 is different for each single board.
- Step 905 The auxiliary vector processor outputs internal processing data to the post processing unit.
- Step 907 The centralized vector processor performs external crosstalk cancellation processing on the user data obtained in step 906 to obtain external processing data.
- the user data in this step is the data of all users of each board, and can be represented by a Nw x l ⁇ 'J vector.
- the external crosstalk cancellation process is: the centralized vector processor generates an Nw x Nw crosstalk cancellation processing matrix W2, which is illustrated, because the auxiliary vector processor has processed the user between the users of the auxiliary vector processor Far-end crosstalk, so the diagonal position of the matrix W2
- the squares of N ") ⁇ (from top left to bottom right) are N zero matrices, and do not participate in the subsequent matrix multiplication; the matrix of Nw X 1 formed by the matrix W2 and the user data of each board The vector is subjected to matrix multiplication, and a column vector of Nw X 1 is obtained as externally processed data.
- Step 908 The centralized vector processor returns the obtained external processing data to the post processing unit of each board.
- steps 903 to 905 and steps 906 to 908 may be two processes in parallel, and there is no order.
- Step 909 The post-processing unit of each board receives the external processing data of the board, and performs post-processing with the internal processing data.
- the post-processing means that the post-processing unit of each board adds the internal processing data of each user of the board and the external processing data to obtain user data after the user performs the crosstalk cancellation processing; of course, the post-processing still
- the method may include: decoding user data after each user performs crosstalk cancellation processing.
- the process needs to perform at most 4 (N*". 2 _ 4 *" 2 multiplication operations and ⁇ _3w * « 2 addition operations, so that the amount of calculation of the ensemble vector processor can be reduced compared with the prior art. In other words.
- the data processing capability of the centralized vector processor is constant, the user capacity of the vector DSL system can be increased.
- a method of vector DSL is provided for the system in this embodiment.
- the above steps are performed only for the user data of at least one board including the auxiliary vector processor, and the auxiliary data is not included.
- the user data of the card of the vector processor can be implemented by the prior art method, and therefore will not be described here; for a vector DSL system in which only part (at least one) of the boards includes the auxiliary vector processor, the concentration is concentrated.
- the vector processor performs external vectorization processing on the user data of the board including the auxiliary vector processor, and performs internal vectorization processing and external vectorization processing on the user data of the board not including the auxiliary vector processor; for example, FIG.
- the board containing the auxiliary vector processor is only the board 1 and the board 2
- the board not including the auxiliary vector processor is the N-2 except the board 1 and the board 2.
- the auxiliary processor of the single board and the single board 2 performs internal vectorization processing on the user data of the boards.
- the centralized vector processor in the system only performs user data on the board 1 and board 2.
- the internal vectorization process performs internal vectorization processing and external vectorization processing on user data of other N-2 boards.
- At least one vector DSL system including an auxiliary vector processor in the single board can also share the data processing amount of the centralized vector processor through the auxiliary vector processor, thereby processing the data in the centralized vector processor.
- the user capacity of the vector DSL system can be increased.
- Embodiment 3 is a diagrammatic representation of Embodiment 3
- the embodiment of the present invention further optimizes the scheme for the vector DSL system shown in FIG. 6 in the second embodiment.
- the solution provided by the present invention further includes step S1 or step S2 before performing step 906 to learn the users who are strongly related to other single-board lines on each board, and only these The user's data is transferred to the centralized vector processor for the purpose of reducing bandwidth.
- Step S1 and step S2 are two parallel schemes, and one of them can be selected.
- Step Sl Each board selects users on the board that are strongly related to the subscriber lines of other boards. Users who are strongly related to the subscriber lines of other boards.
- Step S2 The centralized vector processor selects users who are strongly related to user lines of other boards on each board;
- This step can be that the centralized vector processor selects each order based on empirical values or other means. Users on the board that are strongly related to the user lines of other boards. If this step is used, each board is controlled by the centralized vector processor, and the user data strongly related to other board lines on the board is transmitted to the centralized vector processor for subsequent data processing.
- steps 906 to 908 are different from the second embodiment.
- steps 906 to 908 are different from the second embodiment.
- steps 906 to 908 refer to the description in the second embodiment.
- Step 906 Each board transmits user data that is strongly related to user lines of other boards to the centralized vector processor.
- the board 1 transmits the data of the user to the centralized vector processor; and a strong user associated with other boards (boards other than the board 2) of the subscriber line are "2, the board 2 these" user 2 data transfer vector processing to focus on the plate 2
- the board N transmits the data of the user to the centralized vector until there is one user who is strongly associated with the subscriber line of the other board (the board other than the board N).
- Processor and ⁇ ⁇ ⁇ ⁇ ⁇ > ⁇ 2 ⁇ ⁇ , ... , ⁇ ⁇ ⁇ ⁇ .
- the symbol rate is 4000/s, then the data bandwidth required by each user is
- m l +m 1 +--- + m N 524 *( ⁇ ⁇ + ⁇ 1 + ⁇ + ⁇ ⁇ ) Mbits/s using the prior art centralized vector DSL system, the required ingress bandwidth 52 * N * n Mbits / s; the vector DSL system provided by the embodiment of the present invention and the prior art Compared to centralized vector DSL systems, due to? ⁇ ⁇ + ⁇ + ... + ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ , so that as long as at least one user of any single board is not strongly related to the users of other boards, the vector DSL system provided by the embodiment of the present invention can be compared with the existing The centralized vector DSL system in the technology uses less bandwidth.
- Step 907 The centralized vector processor performs external vectorization processing on the user data obtained in step 906 to obtain external processed data.
- the number of obtained user data is less than or equal to Nn, if obtained If the user data is less than Nw, the user data that is not acquired is set to 0, so that all the user data can still be represented by a column vector of Nwxl.
- the external vectorization process is: the centralized vector processor generates a NW X NW crosstalk cancellation processing matrix W2, which is illustrated, because the auxiliary vector processor has processed the user between the users of the auxiliary vector processor Far-end crosstalk, so the square matrix of N ") ⁇ of the diagonal position (upper left to the lower right position) of the matrix W2 is N zero matrices, and does not participate in the subsequent matrix multiplication operation; the matrix W2 and each The column vector of Nw X 1 formed by the user data of the board is matrix multiplied to obtain a column vector of Nwxl as external processing data. Since the user data set to 0 indicates that the user and any user of other boards are between It is not strongly related to the line. Therefore, when the matrix operation is performed, the calculation related to the user data set to 0 can be omitted, so that the calculation amount of the centralized vector processor can be reduced without ensuring that the final result is not affected.
- Step 908 The centralized vector processor returns the obtained external processing data to the post processing unit of each board.
- the centralized vector processor returns non-zero external processing data to the post processing unit of each board.
- the egress bandwidth of the uplink data has the same bandwidth requirement as the ingress bandwidth, which is the same as the beneficial effect of the step 906 of the embodiment of the present invention. This step can reduce the data transmission bandwidth between the centralized vector processor and the board.
- auxiliary vector processing is included in only a part (at least one) of the boards.
- Vector DSL system such as the vector DSL system shown in Figure 5,
- step S1 is that the board that includes the auxiliary vector processor is further configured to select a user on the board that is strongly related to user lines of other boards.
- the step is specifically that the board 1 in the system is also used to select other boards on the board 1 (N-1 boards except the board 1)
- the user of the user line is strongly related.
- the same board 2 is also used to select users on the board 2 that are strongly related to the subscriber lines of other boards (excluding N-1 boards other than the board 2).
- This step is to transfer only the user data that is strongly related to the user lines of other boards on the board containing the auxiliary vector processor to the centralized vector processor, thus saving the board and the centralized vector processor.
- Step S2 is: the centralized vector processor is further configured to select a user that is strongly related to user lines of other boards on each board that includes the auxiliary vector processor;
- the step is specifically that the centralized vector processor in the system is also used to select other boards on the board 1 (excluding the single board 1 and N-1 singles).
- the user who is strongly associated with the subscriber line of the board is also used to select the user on the board 2 that is strongly related to the subscriber line of other boards (excluding N-1 boards other than the board 2).
- the purpose of this step is to allow the centralized vector processor to control the selected user's board to transfer the user's user data to the centralized vector processor. That is, the centralized vector processor needs to acquire the auxiliary vector processor.
- the user data on the board that is strongly related to the user lines of other boards, so that the bandwidth of data transmission between the board and the centralized vector processor can be saved to some extent.
- Embodiment 4 is a diagrammatic representation of Embodiment 4:
- the method provided by the embodiment of the present invention is still based on the vector DSL system shown in FIG. 6.
- the number of the boards is N (N ⁇ 2), which are respectively the board 1 and the board 2 board N; each board The number of users is n (n>2), and the user of the board 1 is X ⁇ X", the user of the board 2 is X Tra + 1 ⁇ X 2 volunteer, and the user of the board N is X( w — 1 + 1; and each board contains an auxiliary vector processor 10, for the case of downlink data to the CO end user, the present embodiment mentioned embodiment.
- the methods provided include:
- Step 1001 The CO end sends the data that needs to be downlinked to the user to the board where each user is located;
- the co-end will send the data sent to the user ⁇ » to the board 1.
- the data that needs to be downlinked to the user ⁇ réelle + 1 ⁇ ⁇ 2 consider will be sent to the board 2, which will need to be downlinked to the user X ( W —
- the data of 1 +1 ⁇ is sent to the board N. Specifically, it is the pre-processing unit sent to each board.
- Step 1002 The pre-processing unit on each board performs pre-processing on the user data sent by the CO end received by the board;
- the pre-processing in this step includes: the pre-processing unit on each board passes the user data received by the board through the encoding process of the encoder.
- Step 1003 The pre-processing unit of each board transmits the user data of the board to the auxiliary vector processor on the board.
- Step 1004 The auxiliary vector processor performs internal precoding processing on the user data of the board to obtain internal processing data.
- the internal precoding process in this step is specifically that the auxiliary vector processor of each board generates an X precoding processing matrix P 1 , and the matrix P 1 and the user data of the board constitute a column vector of X 1 .
- a matrix multiplication operation is performed to obtain a column vector of w X 1 as internal processing data of the board.
- the matrix P1 generated by the auxiliary vector processor of each board is generated to eliminate the crosstalk that may occur between users of the board, so the matrix P1 is different for each board. .
- Step 1005 the auxiliary vector processor outputs internal processing data to the post-processing unit; the processor;
- Step 1007 The centralized vector processor performs external data obtained by step 1006. Part vectorization processing, obtaining external processing data;
- the user data in this step is the data of all users of each board. You can use one.
- the external pre-coding process is: the centralized vector processor generates an NW X NW pre-coding processing matrix P 2 , which is illustrated, because the auxiliary vector processor has processed the user between the boards of the auxiliary vector processor Far-end crosstalk, so the square of the N"" ⁇ of the diagonal position (upper left to the lower right position) of the matrix P 2 is N zero matrices, and does not participate in the subsequent matrix multiplication; 2 Do matrix multiplication with the column vector of Nw x1 formed by the user data of each board, and obtain a column vector of Nw X 1 as external processing data.
- Step 1008 The centralized vector processor returns the obtained external processing data to the post processing unit of each board.
- steps 1003 to 1005 and steps 1006 to 1008 may be two processes in parallel, and there is no order.
- Step 1009 The post-processing unit of each board receives the external processing data of the board, and performs post-processing in combination with the internal processing data.
- the post-processing in this step is that the post-processing unit of each board adds the internal processing data of each user of the board and the external processing data to obtain user data after the user performs pre-coding processing;
- the processing may further include performing inverse Fourier transform on the user data pre-coded by each user, and then transmitting the data to each user.
- the process needs to perform at most 4 (N * 2 multiplications and ⁇ _3w * «2 addition operations, as compared with the prior art can reduce the amount of calculation centralized vector processor" * 2 _ 4). " In other words. When the data processing capability of the centralized vector processor is constant, the user capacity of the vector DSL system can be increased.
- Embodiment 5 is a diagrammatic representation of Embodiment 5:
- the embodiment of the present invention further optimizes the scheme for the vector DSL system shown in Fig. 6 in the fourth embodiment.
- the solution provided by the present invention further includes step Q1 or step Q2 before performing step 1006 to learn the users who are strongly related to other single-board lines on each board, and only these The user's data is transferred to the centralized vector processor for the purpose of reducing bandwidth.
- Step Q1 and step Q2 are two parallel schemes, and one of them can be selected.
- Step Q1 Each board selects users on the board that are strongly related to the subscriber lines of other boards. The user line of his board is strongly related to the user.
- Step Q2 The centralized vector processor selects users who are strongly related to the user lines of other boards on each board;
- the centralized vector processor selects users who are strongly related to the user lines of other boards on each board according to experience values or other methods. If this step is used, each board is controlled by the centralized vector processor, and the user data strongly related to other board lines on the board is transmitted to the centralized vector processor for subsequent data processing.
- steps 1006 to 1008 are different from the fourth embodiment.
- steps 1006 to 1008 are different from the fourth embodiment.
- steps 1006 to 1008 refer to the description in the fourth embodiment.
- Step 1006 Each board transmits user data that is strongly related to user lines of other boards to the centralized vector processor.
- the board 1 transmits the data of the user to the centralized vector processor; and a strong user associated with other boards (boards other than the board 2) of the subscriber line are "2, the board 2 these" user 2 data transfer vector processing to focus on the plate 2 Device
- the board N transmits the data of the user to the centralized vector processor until there is one user who is strongly associated with the subscriber line of the other board (the board other than the board N).
- the bandwidth requirement between the vector board and the centralized vector processor of the vector DSL system provided by the embodiment of the present invention is obtained by referring to the calculation method in the third embodiment.
- the vector DSL system provided by the embodiment of the present invention can use a smaller bandwidth than the centralized vector DSL system in the prior art, as long as at least one user of any single board is not strongly related to the users of other boards. .
- Step 1007 The centralized vector processor performs external vectorization processing on the user data obtained in step 1006 to obtain external processed data.
- the number of obtained user data is less than or equal to one, if obtained If the user data is less than one, the user data that has not been acquired is set to 0, so that all the user data can still be represented by a column vector of Nwxl.
- the external vectorization process is: the centralized vector processor generates a NWXNW crosstalk cancellation processing matrix P2.
- the auxiliary vector processor since the auxiliary vector processor has processed the remote end between the users of the board where the auxiliary vector processor is located Crosstalk, so the square matrix of the N"" ⁇ of the diagonal position (upper left to the lower right position) of the matrix P2 is N zero matrices, and does not participate in the subsequent matrix multiplication operation; the matrix P2 and the individual boards
- the column vector of Nwxl formed by the user data is matrix multiplied to obtain a column vector of Nwxl as external processing data. Since the user data set to 0 indicates that the user and any user of other boards are not strong.
- Step 1008 The centralized vector processor returns the obtained external processing data to the post processing unit of each board.
- the centralized vector processor can transmit non-zero external processing data to the post processing unit of each board.
- the egress bandwidth of the downlink data has the same bandwidth requirement as the ingress bandwidth, which is the same as the beneficial effect of the step 1006 of the embodiment of the present invention. This step can reduce the data transmission bandwidth between the centralized vector processor and the board.
- a vector DSL system vector DSL system including only an auxiliary vector processor in a part (at least one) of the boards shown in FIG. 5,
- the board including the auxiliary vector processor is further used to select a user on the board that is strongly related to the user lines of other boards.
- the centralized vector processor is further configured to select a user that is strongly related to the user lines of other boards on each board that includes the auxiliary vector processor.
- the bandwidth of data transmission between the board and the centralized vector processor can also be reduced.
- the embodiment of the present invention further provides a board applied to the vector DSL system shown in FIG. 5 or FIG. 6, the board includes:
- the pre-processing unit 1101 is configured to receive and process user data of the board in advance;
- the first transmitting unit 1102 is configured to transmit user data of the board to an auxiliary vector processor on the board;
- An auxiliary vector processor 1103, configured to perform internal vectorization processing on user data of the board, and output internal processing data;
- a first receiving unit 1104 configured to receive internal processing data output by the auxiliary vector processor
- a second transmitting unit 1105 configured to transmit user data of the board to a centralized vector
- the processor performs external vectorization processing
- a second receiving unit 1106, configured to receive external processing data that is output after being externally vectorized by the centralized vector processor
- the post-processing unit 1107 is configured to post-process the internal processing data of the first receiving unit and the external processing data of the second receiving unit.
- the block diagram of the single board shown in FIG. 11 is applied to the single board 1 and the single board 2 in the vector DSL system shown in FIG. 5; in the vector DSL system shown in FIG. 6, it is applied to all N singles. board.
- the vectoring process includes precoding processing and/or crosstalk canceling processing.
- the vectorization processing refers to crosstalk cancellation processing; in the case where the CO side downlink data is sent to the user, the vectorization processing refers to precoding processing. That is, the board can be used only for uplink or only for downlink. Of course, it can be used for both uplink and downlink (in the case of user uplink data to the CO end, but the board is used for uplink; In the case where the downlink data is sent to the user, the board is used for downlink).
- the centralized vector processor may be integrated on the single board, or the centralized vector processor may exist separately from the single board substrate.
- the board shown in FIG. 12 may further include:
- the selecting unit 1108 is configured to select a user on the board that is strongly related to user lines of other boards;
- the second transmitting unit 1105 is specifically configured to transmit user data strongly related to user lines of other boards on the board to the centralized vector processor for external vectorization processing;
- the second receiving unit 1106 is configured to receive external processing data that is output by the centralized vector processor after external vectorization processing is performed on user data that is strongly related to user lines of other boards on the board.
- the centralized vector processor can be reduced when the number of system users is constant.
- the user capacity of the vector DSL system can be further increased.
- the embodiment of the present invention further provides a DSLAM device, where the device includes at least two boards, and at least one of the boards is as shown in FIG.
- the pre-processing unit 1101 is configured to receive and process user data of the board in advance;
- the first transmitting unit 1102 is configured to transmit user data of the board to an auxiliary vector processor on the board;
- An auxiliary vector processor 1103, configured to perform internal vectorization processing on user data of the board, and output internal processing data;
- a first receiving unit 1104 configured to receive internal processing data output by the auxiliary vector processor
- a second transmitting unit 1105 configured to transmit user data of the board to a centralized vector processor for external vectorization processing
- a second receiving unit 1106, configured to receive external processing data that is output after being externally vectorized by the centralized vector processor
- the post-processing unit 1107 is configured to post-process the internal processing data of the first receiving unit and the external processing data of the second receiving unit.
- the board may further include:
- the selecting unit 1108 is configured to select a user on the board that is strongly related to user lines of other boards; At this time, the second transmitting unit 1105 is specifically configured to: transmit user data strongly related to user lines of other boards on the board to the centralized vector processor for external vectorization processing;
- the second receiving unit 1106 is configured to receive external processing data that is output by the centralized vector processor after external vectorization processing is performed on user data that is strongly related to user lines of other boards on the board.
- the centralized vector processor is configured to externally vectorize the user data of the board including the auxiliary vector processor, and the user data of the board that does not include the auxiliary vector processor. Internal vectorization processing and external vectorization processing are performed.
- the board containing the auxiliary vector processor in the DSLAM device provided by the embodiment of the present invention shares a part of the data processing amount by the auxiliary vector processor, so that the centralized vector processor can be reduced when the number of system users is constant.
- the user capacity of the vector DSL system can be further increased.
- the present invention can be implemented by means of software plus necessary general hardware, and of course, by hardware, but in many cases, the former is a better implementation. .
- the technical solution of the present invention which contributes in essence or to the prior art, can be embodied in the form of a software product based on the hardware architecture of the vector DSL system, and the computer software product is stored in a readable form.
- the storage medium such as a floppy disk, hard disk or optical disk of a computer, includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments of the present invention.
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Telephonic Communication Services (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Description
矢量 DSL的方法、 系统及一种单板和 DSLAM设备 技术领域
本发明涉及数字用户线路 DSL数据处理领域, 尤其涉及矢量 DSL的 方法、 系统及一种单板和 DSLAM设备。
背景技术
xDSL是各种类型 DSL ( Digital Subscribe Line, 数字用户线路) 的总 称, xDSL是一种在电话双绞线(即无屏蔽双绞线, Unshielded Twist Pair, 简称为 UTP ) 上传输高速数据的传输技术。 在 xDSL 系统的局端 (CO , 全称为 Central Office )有提供多路 xDSL信号接入的装置 DSLAM( Digital Subscriber Line Access Multiplexer, DSL接入复用器)。
由于电磁感应现象的存在, DSLAM接入的多路信号之间, 会产生相 互干扰, 称为串扰 ( Crosstalk )。 如图 1 所示, 存在两种串扰: 近端串扰 ( NEXT, 全称为 Near-end crosstalk )和远端串扰 ( FEXT, 全称为 Far-end crosstalk )。 由于 xDSL上下行信道采用频分复用, 所以 NEXT对系统的 性能不产生太大的危害; 但 FEXT则会严重影响线路的传输性能。 例如, 当一捆电缆内有多路用户都要求开通 xDSL业务时, 会因为 FEXT使一些 线路速率低、 性能不稳定、 甚至不能开通等问题。 为消除 FEXT , 目前业 界提出了矢量 DSL ( Vectored - DSL ) 技术, 该技术主要利用 DSLAM联 合收发的功能, 使用信号处理的方法来消除 FEXT干扰。
如图 2所示, 对于上行方向而言, CO端做信号的联合接收处理。 首 先, 设 ^个用户发送的信号组成信道输入向量, 为一个 f xl的列向量 x= ( ι, 2, · · · , ^)Γ ;传输信道 Η表示在频率域的一个子载波上各 DSL之间 的串扰信道, Η是一个 f 的信道传输矩阵, 表达式如下:
H = (公式 1 )
其中, 表示从信道 j到信道 i的传输方程。 另外, b用来表示 f x i的 噪声向量, 为一个 的列向量; y表示信道输出向量, 为一个 χ ΐ的列 向量。 对于信道输出向量 y的表达式, 如下:
y = Hx + b (公式 2 ) 为了消除串扰, 在 CO端做信号的联合接收处理, 在接收端引入一个 串扰消除处理器, 该串扰消除处理器生成一 X 的串扰消除处理矩阵 w, 且 w使得 WH为一个对角矩阵, 之后按照公式 3对信号输出向量 y进行 处理, 得到处理后的信道输出向量^
y = WHx + Wb (公式 3 ) 由于 WH为一个对角矩阵时, 故串扰可消除。
如图 3所示, 对于下行方向而言, 在 CO端做信号的联合发送处理。 同 样地, 设有 个用户, CO端需要向 个用户发送信号组成一 x i的列向量 X , 为了消除传输信道的串扰, 在 CO端引入一个向量预编码处理器, 该向 量预编码处理器生成一 的预编码处理矩阵 P, 且 P使得 HP为一个对角 矩阵, 此时发送的信号为 Ϊ , 其表达式如下:
ϊ = Ρχ 用户的接收单元接收到的信号为:
y = HPx + b (公式 5 ) 由于 HP为一个对角矩阵, 故串扰可消除。
基于上述原理, 现有技术中提出了一种集中式矢量 DSL系统, 如图 4 所示, 该系统包括 N个单板, 每个单板控制 n个用户, 且所有单板的用户
数据都传送到集中矢量处理器上进行数据处理以消除串扰。该集中矢量处 理器兼有串扰消除处理器和向量预编码处理器的功能。
对于上行方向, 每个单板将经过第一数据处理单元的 n个用户的数据 发送到集中矢量处理器中, 故集中矢量处理器接收到所有 N*w个用户的数 据 Hx, Hx为(N*«)xl的列向量; 集中矢量处理器生成一(N*«)x(N*«)的串扰 消除处理矩阵 W, 集中矢量处理器需要进行 W*Hx的运算, 由于每个向量 的分量都是复数, 故在此过程中需要进行 4(N*«)2个乘法运算及
3(N*«)2 -(N*«)个加法运算; 然后, 集中矢量处理器将处理完成的数据返回 至相应单板, 进行后续的数据处理。
对于下行方向,集中矢量处理器生成一预编码处理矩阵 P,并进行 Ρ*χ 的向量运算, 同样也需要进行 4(N*«)2个乘法运算及 3(N*«)2 -(N*«)个加法 运算。
由于利用矢量 DSL技术进行预编码处理及串扰消除处理, 有较高的实时 性要求, 所以使用现有的集中式矢量 DSL 系统, 要求集中矢量处理器能 够实时处理所有单板上用户数据, 则该集中矢量处理器需要具有很高的数 据处理能力。 用户数越多, 对集中矢量处理器的数据处理能力的要求也就 越高; 依据现有集中矢量处理器的数据处理能力, 就对矢量 DSL 系统的 用户容量造成了限制。
发明内容
本发明的实施例提供矢量 DSL 的方法、 系统及一种单板和 DSLAM 设备, 用以增大矢量 DSL系统的用户容量。
为达到上述目的, 本发明的实施例采用如下技术方案:
一种矢量 DSL的方法, 包括:
至少一个单板接收并先期处理各自单板的用户数据; 所述单板将该单板的用户数据传送给该单板上的辅助矢量处理器进 行内部矢量化处理, 得到内部处理数据;
所述单板将其用户数据传送给集中矢量处理器进行外部矢量化处理,
得到外部处理数据;
所述单板接收该单板的外部处理数据,并结合内部处理数据进行后期 处理。
一种单板, 包括:
先期处理单元, 用于接收并先期处理所述单板的用户数据;
第一传送单元,用于将所述单板的用户数据传送给该单板上的辅助矢 量处理器;
辅助矢量处理器, 用于对所述单板的用户数据进行内部矢量化处理, 并输出内部处理数据;
第一接收单元, 用于接收由辅助矢量处理器输出的内部处理数据; 第二传送单元,用于将所述单板的用户数据传送给集中矢量处理器进 行外部矢量化处理;
第二接收单元,用于接收由集中矢量处理器进行外部矢量化处理后输 出的外部处理数据;
后期处理单元,用于将所述第一接收单元的内部处理数据和所述第二 接收单元的外部处理数据进行后期处理。
一种数字用户线路接入复用器 DSLAM设备, 包括至少两个单板, 其 中, 至少一个所述单板包括:
先期处理单元, 用于接收并先期处理所述单板的用户数据;
第一传送单元,用于将所述单板的用户数据传送给该单板上的辅助矢 量处理器;
辅助矢量处理器, 用于对所述单板的用户数据进行内部矢量化处理, 并输出内部处理数据;
第一接收单元, 用于接收由辅助矢量处理器输出的内部处理数据; 第二传送单元,用于将所述单板的用户数据传送给集中矢量处理器进 行外部矢量化处理;
第二接收单元,用于接收由集中矢量处理器进行外部矢量化处理后输
出的外部处理数据;
后期处理单元,用于将所述第一接收单元的内部处理数据和所述第二 接收单元的外部处理数据进行后期处理。
一种矢量 DSL系统, 包括: DSLAM设备; 所述 DSLAM设备包括至 少两个单板, 其中至少一个所述单板包含一辅助矢量处理器, 该辅助矢量 处理器用于对其所在单板的用户数据进行内部矢量化处理,并且包含所述 辅助处理器的单板的用户数据由集中矢量处理器进行外部矢量化处理; 所述集中矢量处理器用于对包含辅助矢量处理器的单板的用户数据 进行外部矢量化处理,并对不包含辅助矢量处理器的单板的用户数据进行 内部矢量化处理和外部矢量化处理。
本发明实施例提供的矢量 DSL的方法、 系统及一种单板和 DSLAM设备, 通过在至少一个单板上设置一辅助矢量处理器,利用该辅助矢量处理器对 其所在单板的用户数据进行内部矢量化处理,并且利用集中矢量处理器对 所述单板的用户数据进行外部矢量化处理;由于辅助矢量处理器分担了一 部分的数据处理量, 故在系统用户数量不变的情况下, 可以减少集中矢量 处理器的数据处理量, 在集中矢量处理器的数据处理能力一定的情况下, 可以增大矢量 DSL 系统的用户容量; 当然, 若集中矢量处理器的数据处 理能力可以得到提升, 那么运用本发明提供的方案可以进一步增大矢量 DSL系统的用户容量。
附图说明
实施例或现有技术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员 来讲, 在不付出创造性劳动性的前提下, 还可以根据这些附图获得其他的 附图。
图 1为近端串扰和远端串扰的示意图;
图 2为用户上行数据给 CO端的示意图;
图 3为 CO端下行数据给用户的示意图;
图 4为现有技术提供的一种集中式矢量 DSL系统示意图; 图 5为实施例一提供的一种矢量 DSL系统示意图;
图 6为实施例一提供的另一种矢量 DSL系统示意图;
图 Ί为本发明实施例提供的一种矢量 DSL的方法示意图;
图 8为本发明实施例提供的另一种矢量 DSL的方法示意图; 图 9为实施例二针对用户上行数据给 CO端的情况,提供的一种矢量
DSL的方法示意图;
图 10为实施例四针对 CO端下行数据给用户的情况, 提供的一种矢 量 DSL的方法示意图;
图 1 1为实施例六提供的一种单板的结构框图;
图 12为实施例六提供的另一种单板的结构框图。
具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进 行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例, 而不是全部的实施例。 基于本发明中的实施例, 本领域普通技术人员在没 有作出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的 范围。
实施例一:
如图 5所示,本发明实施例提供了一种矢量 DSL系统,包括: DSLAM 设备;所述 DSLAM设备包括至少两个单板其中至少一个所述单板包含一 辅助矢量处理器,该辅助矢量处理器用于对其所在单板的用户数据进行内 部矢量化处理,并且包含所述辅助处理器的单板的用户数据由集中矢量处 理器进行外部矢量化处理;
所述集中矢量处理器用于对包含辅助矢量处理器的单板的用户数据 进行外部矢量化处理,并对不包含辅助矢量处理器的单板的用户数据进行 内部矢量化处理和外部矢量化处理。
所述集中矢量处理器设置在所述 DSLAM 设备中, 或者独立于所述
DSLAM设备。 图 5以集中矢量处理器独立于 DSLAM设备为例, 但集中 矢量处理器的位置并不局限于图 5所示; 例如, 集中矢量处理器可以做成 单板插在 DSLAM设备的插槽中。
在本实施例中, DSLAM设备中单板的个数为 N ( N≥2 ), 分别是单板
1、 单板 2 单板 N; 每个单板的用户数为 n ( n≥2 ), 且单板 1的用户 为 Χ^ Χ» , 单板 2的用户为 Χ„ + 1〜Χ2„, 单板 Ν的用户为 Χ(ΛΓ— 1 +1〜 。 其中, Ν个单板中至少一个单板包含一辅助矢量处理器即可; 图 5所示的 矢量 DSL系统中, 以单板 1、 单板 2各包含一辅助矢量处理器为例; 图 6 所示的矢量 DSL系统是一种最佳的情况, 该矢量 DSL系统的每个单板中 各包含一辅助矢量处理器, 由于辅助矢量处理器越多, 则分担的数据处理 量就越多, 那么在集中矢量处理器的数据处理能力一定的情况下, 可以矢 量 DSL系统的用户容量也就越多。
所述矢量化处理包括预编码处理和 /或串扰消除处理。
其中, 针对用户上行数据给 CO端的情况, 所述矢量化处理指的是对 上行的用户数据的串扰消除处理。 这种情况下, 所述内部矢量化处理是指 每个单板上的辅助矢量处理器为消除该单板上用户和用户之间所产生的 串扰而进行的串扰消除处理, 此时, 所述内部矢量化处理也称为内部串扰 消除处理;所述外部矢量处理是指集中矢量处理器为消除各个单板的用户 和其他单板的用户之间所产生的串扰而进行的串扰消除处理, 此时, 所述 外部矢量化处理也称为外部串扰消除处理。 例如, 对于单板 1而言, 内部 矢量化处理是指为消除用户 Χ ^ Χ»之间所产生的串扰而进行的串扰消除 处理, 外部矢量化处理是指为消除用户 Χ ^Χ^ρ其他单板的用户之间所产 生的串扰而进行的串扰消除处理。
另外, 对于 CO端下行数据给用户而言, 所述矢量化处理指的是对下 行的用户数据的预编码处理。 在这种情况下, 所述内部矢量化处理是指每 个单板上的辅助矢量处理器为消除该单板上用户和用户之间会产生的串 扰而进行的预编码处理, 此时, 所述内部矢量化处理也称为内部预编码处
理;所述外部矢量处理是指集中矢量处理器为消除各个单板的用户和其他 单板的用户之间会产生的串扰而进行的预编码处理, 此时, 所述外部矢量 化处理也称为外部预编码处理。 例如, 对于单板 1而言, 内部矢量化处理 是指为消除用户 χ^χ»之间会产生的串扰而进行的预编码处理, 外部矢量 化处理是指为消除用户 Χ Χ ^Ρ其他单板的用户之间会产生的串扰而进 行的预编码处理。
需要说明的是,对于矢量 DSL系统的不包含辅助处理器的单板而言, 该单板的用户数据的内部矢量化处理和外部矢量化处理都由集中矢量处 理器完成。
由于本发明提供的矢量 DSL 系统中的每个单板上的辅助矢量处理器 分担了一部分的数据处理量, 故在系统用户数量不变的情况下, 可以减少 集中矢量处理器的数据处理量,在集中矢量处理器的数据处理能力一定的 情况下, 可以增大矢量 DSL系统的用户容量。
在本实施例中, 为减少单板和集中矢量处理器之间的数据传输带宽, 所述包含所述辅助矢量处理器的单板还可以用于选择单板上的与其他单 板的用户线路强相关的用户,并将该单板上的与其他单板的用户线路强相 关的用户数据传送给所述集中矢量处理器。
本实施例中, 对于线路强相关做如下解译:
在 CO端和用户之间通过线路进行数据传输(包括数据上行和数据下 行)的过程中, 若一个用户的线路对另一用户的线路上传输的数据会产生 串扰, 则这两个用户线路强相关。 例如, 若单板 1 上的用户 和其他单 板的任一用户 X; ( n+l<i<N*n )的线路会产生串扰, 则用户 Χ ^ Χ,线路强 相关。 若单板 1有 nl ( nl<n ) 个用户, 例如为 X^ Xw和其他单板的任一 用户 X; ( n+l<i<N*n ) 的线路会产生串扰, 则该单板 1上的与其他单板的 用户线路强相关的用户为 X ^Xw ,该单板 1上的与其他单板的用户线路强 相关的用户数据为用户 X 〜 Xnl的数据。
所述选择该单板上的与其他单板的用户线路强相关的用户,可以是单
板根据经验值或者其他方式得到。
上述单板所具有的选择功能也可以由集中矢量处理器完成,所述集中 矢量处理器还用于选择每个包含辅助矢量处理器的单板上的与其他单板 的用户线路强相关的用户,并控制该用户所在单板将该用户的用户数据传 送给所述集中矢量处理器。
若所述集中矢量处理器设置在 DSLAM设备上,则集中矢量处理器可 以集成在所述至少两个单板中的任一单板上, 当然, 所述集中矢量处理器 也可以独立于所述至少两个单板而单独存在。
由于在本发明实施例中,单板和集中矢量处理器之间传输的数据可以 是每个单板上与其他单板的用户线路强相关的用户数据,即可以是每个单 板的部分用户数据, 故可以减少对数据传输带宽的使用。
基于上述矢量 DSL系统, 本实施例还提供了一种矢量 DSL的方法。 如图 7所示, 该方法包括以下步骤:
701、 至少一个单板接收并先期处理各自单板的用户数据;
本实施例中的方法是针对包含辅助矢量处理器的单板而言的,即所述 至少一个单板是指至少一个包含有辅助矢量处理器的单板;对于不包含有 辅助矢量处理器的单板的用户数据的处理方法,与现有技术相同都是由集 中矢量处理器对其用户数据进行内部矢量处理和外部矢量处理,故在此不
702、 所述单板将该单板的用户数据传送给该单板上的辅助矢量处理 器进行内部矢量化处理, 得到内部处理数据;
703、 所述单板将其用户数据传送给集中矢量处理器进行外部矢量化 处理, 得到外部处理数据;
704、 所述单板接收该单板的外部处理数据, 并结合内部处理数据进 行后期处理。
所述矢量化处理包括预编码处理和 /或串扰消除处理。
本发明提供的矢量 DSL 的方法, 由于至少一个单板上的辅助矢量处
理器分担了一部分的数据处理量, 故在系统用户数量不变的情况下, 可以 减少集中矢量处理器的数据处理量,在集中矢量处理器的数据处理能力一 定的情况下, 可以增大矢量 DSL 系统的用户容量; 当然, 若集中矢量处 理器的数据处理能力可以得到提升,那么运用本发明提供的方案可以进一 步增大矢量 DSL系统的用户容量。
优选地, 为了减少单板和集中矢量处理器之间的数据传输带宽, 如图 8所示在步骤 703之前还包括:
步骤 705、 所述单板选择该单板上的与其他单板的用户线路强相关的 用户; 或者, 所述集中矢量处理器选择所述单板上的与其他单板的用户线 路强相关的用户。
此时步骤 703具体可以为,所述单板将该单板上的与其他单板的用户 线路强相关的用户数据传送给集中矢量处理器进行外部矢量化处理,得到 外部处理数据。
由于至少一个单板传送给集中矢量处理器中的数据可以是该单板的 部分用户数据, 故可以达到节省带宽的目的, 当然, 在某些单板全部的用 户数据本身就不是特别大的情况, 可以不选用此优选的方式, 仍将这些单 板全部的用户数据传送给集中矢量处理器。
实施例二:
针对上述矢量 DSL的方法, 本发明实施例将进一步地详述。 本发明 实施例提供的方法仍是基于图 6所示的矢量 DSL系统, 其中单板的个数 为 N ( N≥2 ), 分别是单板 1、 单板 2 单板 N; 每个单板的用户数为 n ( n>2 ), 且单板 1 的用户为 Χ ^ Χ» , 单板 2的用户为 Χ„ + 1〜Χ2„ , 单 板 Ν的用户为 X(w— 1 +1〜 ; 且每一单板都包含一辅助矢量处理器。 参考图 9 , 针对用户上行数据给 CO端的情况, 本实施例提供的方法包括:
步骤 901、 用户将其数据发送给该用户所在的单板上;
即用户 Χ ^ Χ»将其数据发送给单板 1 , 用户 „ + 1〜 2„将其数据发送给 单板 2 , 用户 X(w— 1)¾+1〜 将其数据发送给单板 Ν。 具体地, 是发送给
该用户所在单板的先期处理单元中。
步骤 902、 各单板上的先期处理单元将其单板接收到的用户数据进行 先期处理;
所述先期处理包括对接收到的用户数据进行快速傅里叶变换。各个单 板上经过快速傅里叶变换后的用户数据做如下表示: 对于单板 1 的用户 Χ ^ Χ»的数据表示为 = {;^, ,...,;^}; 对于单板 2的用户 X„ + 1〜X2 数据表 示为 2 = {x„+1, x„+2, ..., „} ; 同样地, 直至对于单板 N 的用户 X^^ X^的数 据表示为 v = {^( -1)«+1 '
步骤 903、 每个单板的先期处理单元将该单板的用户数据传送给该单 板上的辅助矢量处理器;
步骤 904、 辅助矢量处理器对该单板的用户数据进行内部串扰消除处 理, 得到内部处理数据;
所述内部串扰消除处理具体为, 每个单板的辅助矢量处理器生成一 "串扰消除处理矩阵 W1 ,该矩阵 W1与该单板的用户数据所构成的 w 1的 列向量进行矩阵乘法运算,得到 w 1的列向量作为该单板的内部处理数据。 需要说明的是, 每个单板的辅助矢量处理器所生成的矩阵 W1是为了消除 该单板的用户之间的串扰而生成的, 故对于每个单板的矩阵 W1是不同的。
步骤 907、 集中矢量处理器对通过步骤 906得到的用户数据进行外部 串扰消除处理, 得到外部处理数据;
在此步骤中的用户数据是各个单板的所有用户的数据, 可以用一 Nw x l的歹 'J向量来表示。
所述外部串扰消除处理具体为,集中矢量处理器生成一 Nw x Nw串扰消 除处理矩阵 W2 , 需要说明的是, 由于辅助矢量处理器已经处理了该辅助 矢量处理器所在单板的用户之间的远端串扰, 故该矩阵 W2的对角线位置
(左上至右下的位置) 的 N个") ^的方阵为 N个零矩阵, 且不参与之后 的矩阵乘法运算; 将矩阵 W2和各个单板的用户数据所构成的 Nw X 1的列向 量做矩阵乘法运算, 得到一 Nw X 1的列向量作为外部处理数据。
步骤 908、 集中矢量处理器将得到的外部处理数据回传至各个单板的 后期处理单元;
需要说明的是上述步骤 903〜905和步骤 906〜908可以是并行的两个过 程, 而没有先后顺序之分。
步骤 909、 每个单板的后期处理单元接收该单板的外部处理数据, 并 结合内部处理数据进行后期处理。
所述后期处理是指,每个单板的后期处理单元将该单板的各个用户的 内部处理数据和外部处理数据相加,得到该用户进行串扰消除处理后的用 户数据; 当然, 后期处理还可以包括, 将每个用户进行串扰消除处理后的 用户数据进行解码。
由于在本实施例提供的集中矢量处理器的矩阵乘法运算中,排除了矩 阵 W2中对角线位置的 N个《χ «的方阵的矩阵运算, 此过程具体最多需要 进行 4(N * ")2 _4 *"2个乘法运算及 Ν _3w * «2个加法运算, 故与现有技 术相比可以减少集中矢量处理器的运算量。 换言之。 集中矢量处理器的数 据处理能力一定的情况下, 可以增大矢量 DSL系统的用户容量。
需要说明的是, 由于图 6所示的矢量 DSL系统是一种优选的方案, 故在本实施例中针对该系统提供了一种矢量 DSL 的方法。 而对于只有部 分 (至少一个) 单板中包含辅助矢量处理器的矢量 DSL 系统而言, 上述 各步骤都只是针对至少一个包含辅助矢量处理器的单板的用户数据进行 的,而对于不包含辅助矢量处理器的单板的用户数据可以采用现有技术的 方法就可以实现, 故在此不加贅述; 针对只有部分(至少一个)单板中包 含辅助矢量处理器的矢量 DSL 系统而言, 集中矢量处理器对包含辅助矢 量处理器的单板的用户数据进行外部矢量化处理,并对不包含辅助矢量处 理器的单板的用户数据进行内部矢量化处理和外部矢量化处理; 例如图 5
所示的矢量 DSL系统中包含辅助矢量处理器的单板只有单板 1、 单板 2, 不包含辅助矢量处理器的单板为除去单板 1、 单板 2之外剩下的 N-2个单 板, 单板 1、 单板 2的辅助处理器只对各自单板的用户数据进行内部矢量 化处理, 该系统中的集中矢量处理器对单板 1、 单板 2的用户数据只进行 内部矢量化处理,对其他 N-2个单板的用户数据进行内部矢量化处理和外 部矢量化处理。容易得到对于图 5所示的这种只有部分单板中包含辅助矢 量处理器的矢量 DSL 系统而言, 其集中矢量处理器需要进行大于 4(N*")2_4N*"2且小于 4(W*„)2个乘法运算, 需要进行大于 3(N*«)2-3N*«2且 小于 3(N*«)2— (N*«)个力口法运算。
故而较现有技术而言, 至少一个单板中包含辅助矢量处理器的矢量 DSL 系统, 也可以通过辅助矢量处理器来分担集中矢量处理器的数据处 理量, 从而在集中矢量处理器的数据处理能力一定的情况下, 可以增大矢 量 DSL系统的用户容量。
实施例三:
为了减少单板和集中矢量处理器之间的数据传输带宽,本发明实施例 对实施例二中针对图 6所示的矢量 DSL系统的方案进一步地优化。 相较 于实施二中的方案而言,本发明提供的方案在进行步骤 906之前还包括步 骤 S1或步骤 S2, 以得知各个单板上与其他单板线路强相关的用户, 并只 将这些用户的数据传送给集中矢量处理器, 以达到减少带宽的目的。
步骤 S1和步骤 S2是两个并列的方案, 选择其一即可。
步骤 Sl、 每个单板选择该单板上的且与其他单板的用户线路强相关 的用户。 与其他单板的用户线路强相关的用户。
步骤 S2、 集中矢量处理器选择每个单板上与其他单板的用户线路强 相关的用户;
此步骤可以为,集中矢量处理器根据经验值或者其他方式选择每个单
板上与其他单板的用户线路强相关的用户。 若利用此步骤, 则各个单板在 集中矢量处理器控制下,将该单板上与其他单板线路强相关的用户数据传 送给集中矢量处理器进行后续的数据处理。
此后, 进行步骤 906〜908与实施例二有所不同, 其他步骤可参考实施 例二中的描述。
步骤 906、 各个单板将该单板上的且与其他单板的用户线路强相关的 用户数据传送给集中矢量处理器;
例如, 单板 1上的且与其他单板 (除单板 1之外的单板)的用户线路 强相关的用户有 个, 则单板 1将这 个用户的数据传送给集中矢量处理 器; 单板 2上的的且与其他单板(除单板 2之外的单板)的用户线路强相 关的用户有《2个, 则单板 2将这《2个用户的数据传送给集中矢量处理器; 同样地, 直至单板 N上的且与其他单板 (除单板 N之外的单板) 的用户 线路强相关的用户有 个, 则单板 N将这 个用户的数据传送给集中矢 量处理器; 且 ηλ < η ·> η2 <η , … , ηΝ≤η。
则本发明实施例提供的矢量 DSL 系统在每个单板与集中矢量处理器 之间的带宽需要如下:
假设每一个用户可用的子载波为 4096, 每个子载波的信号数据为
32bit, 符号率为 4000/s, 那么每个用户需要的数据带宽为
4096 X 32 X 4000 = 524 Mbits/S
则单板 1需要的数据带宽为 Wl =524 Mbits/s;
单板 2需要的数据带宽为《¾=524*«2 Mbits/s;
依次类推, 直至
单板 N需要的数据带宽为? =524* Mbits/s;
故对于集中矢量处理器而言, 其需要的入口带宽为
ml +m1 +--- + mN = 524 *(ηι +η1 +··· + ηΝ) Mbits/s 而利用现有技术中的集中式矢量 DSL 系统, 所需要的入口带宽为 52 *N*n Mbits/s; 将本发明实施例提供的矢量 DSL系统与现有技术中的
集中式矢量 DSL系统相比, 由于 ?^ ^ + ^ +…+ ^ ?^ ^ ^ , 故只要任 一单板的至少一个用户与其他单板的用户不是线路强相关,则利用本发明 实施例提供的矢量 DSL系统就可以比现有技术中的集中式矢量 DSL系统 使用更小的带宽。
步骤 907、 集中矢量处理器对通过步骤 906得到的用户数据进行外部 矢量化处理, 得到外部处理数据。
在本发明实施例中,由于通过步骤 906得到的用户数据只是各个单板 上的且与其他单板的用户线路强相关的用户数据,故得到的用户数据的数 量小于或等于 Nn个, 若得到的用户数据小于 Nw个则将没有获取到的用户 数据置为 0, 从而仍可以用一 Nwxl的列向量来表示所有的用户数据。
所述外部矢量化处理具体为,集中矢量处理器生成一 NW X NW串扰消除 处理矩阵 W2 , 需要说明的是, 由于辅助矢量处理器已经处理了该辅助矢 量处理器所在单板的用户之间的远端串扰,故该矩阵 W2的对角线位置(左 上至右下的位置) 的 N个") ^的方阵为 N个零矩阵, 且不参与之后的矩 阵乘法运算; 将矩阵 W2和各个单板的用户数据所构成的 Nw X 1的列向量做 矩阵乘法运算, 得到一 Nwxl的列向量作为外部处理数据。 由于置为 0 的 用户数据表示该用户和其他单板的任一用户之间都不是线路强相关,故在 进行矩阵运算的时候, 也可以将与置为 0的用户数据相关的计算省略, 从 而可以在保证不影响最终结果的前提下减少集中矢量处理器的运算量。
步骤 908、 集中矢量处理器将得到的外部处理数据回传至各个单板的 后期处理单元。
此步骤具体为,集中矢量处理器将非 0的外部处理数据回传至各个单 板的后期处理单元。
上行数据的出口带宽与入口带宽有相同的带宽需求,与本发明实施例 步骤 906的有益效果相同, 利用此步骤可以减小集中矢量处理器与单板之 间的数据传输带宽。
需要说明的是, 对于只有部分(至少一个)单板中包含辅助矢量处理
器的矢量 DSL系统例如图 5所示的矢量 DSL系统而言,
上述步骤 S 1为, 所述包含所述辅助矢量处理器的单板还用于选择单 板上的与其他单板的用户线路强相关的用户。
例如对于图 5所示的矢量 DSL系统, 此步骤具体为, 该系统中的单 板 1还用于选择单板 1上的与其他单板 (除去单板 1之外的 N- 1个单板 ) 的用户线路强相关的用户,同样单板 2还用于选择单板 2上的与其他单板 (除去单板 2之外的 N-1个单板) 的用户线路强相关的用户。
这一步骤的目的是为了只将包含辅助矢量处理器的单板上的与其他 单板的用户线路强相关的用户数据传送给集中矢量处理器,这样就可以节 省单板与集中矢量处理器之间数据传输的带宽。
步骤 S2为, 所述集中矢量处理器还用于选择每个包含辅助矢量处理 器的单板上的与其他单板的用户线路强相关的用户;
例如对于图 5所示的矢量 DSL系统, 此步骤具体为, 该系统中的集 中矢量处理器还用于选择单板 1上的与其他单板(除去单板 1之外的 N-1 个单板)的用户线路强相关的用户, 同样用于选择单板 2上的与其他单板 (除去单板 2之外的 N-1个单板) 的用户线路强相关的用户。
这一步骤的目的是为了让集中矢量处理器控制选择出的用户所在的 单板将该用户的用户数据传送给集中矢量处理器, 也就是说, 集中矢量处 理器需要获取包含辅助矢量处理器的单板上的与其他单板的用户线路强 相关的用户数据,这样就可以在一定程度上节省单板与集中矢量处理器之 间数据传输的带宽。
实施例四:
本发明实施例提供的方法仍是基于图 6所示的矢量 DSL系统, 其中 单板的个数为 N ( N≥2 ), 分别是单板 1、 单板 2 单板 N; 每个单板 的用户数为 n ( n>2 ) , 且单板 1 的用户为 X ^ X" , 单板 2 的用户为 X„ + 1〜X2„, 单板 N的用户为 X(w— 1 +1〜 ; 且每一单板都包含一辅助矢 量处理器。 参考图 10 , 针对 CO 端下行数据给用户的情况, 本实施例提
供的方法包括:
步骤 1001、 CO端将其需要下行给用户的数据发送到各用户所在的单 板上;
即 co端将需要下行给用户 χ^χ»的数据发送到单板 1, 将需要下行 给用户 Χ„ + 1〜Χ2„的数据发送到单板 2, 将需要下行给用户 X(W— 1 +1〜 的 数据发送到单板N。 具体地, 是发送给各单板的先期处理单元。
步骤 1002、 每个单板上的先期处理单元将其单板接收到的 CO 端发 送的用户数据进行先期处理;
此步骤中的先期处理包括:各单板上的先期处理单元将其单板接收到 的用户数据经过编码器的编码处理。经过编码处理后的用户数据可以做如 下表示: 对于单板 1的用户 Χ^Χ»的数据表示为 = { ,..., }; 对于单板 2的用户 Χ„ + 1〜Χ2„的数据表示为 = {χ„+1,χ„+2,...,χ2„}; 同样地, 直至对于单板 Ν的用户 Χ(Ν— 〜ΧΛ¾的数据表示为 = {x(N ' ' ' * ' 。
步骤 1003、 每个单板的先期处理单元将该单板的用户数据传送给该 单板上的辅助矢量处理器;
步骤 1004、 辅助矢量处理器对该单板的用户数据进行内部预编码处 理, 得到内部处理数据;
此步骤中的所述内部预编码处理具体为,每个单板的辅助矢量处理器 生成一 X 预编码处理矩阵 P 1 , 该矩阵 P 1与该单板的用户数据所构成 X 1 的列向量进行矩阵乘法运算, 得到 w X 1的列向量作为该单板的内部处理数 据。 需要说明的是, 每个单板的辅助矢量处理器所生成的矩阵 P1是为了消 除该单板的用户之间的会出现的串扰而生成的,故对于每个单板的矩阵 P1 是不同的。
步骤 1005、 辅助矢量处理器输出内部处理数据至后期处理单元; 处理器;
步骤 1007、 集中矢量处理器对通过步骤 1006得到的用户数据进行外
部矢量化处理, 得到外部处理数据;
在此步骤中的用户数据是各个单板的所有用户的数据, 可以用一
Nw x l的歹 'J向量来表示。
所述外部预编码处理具体为,集中矢量处理器生成一 NW X NW预编码处 理矩阵 P 2 , 需要说明的是, 由于辅助矢量处理器已经处理了该辅助矢量 处理器所在单板的用户之间的远端串扰, 故该矩阵 P 2的对角线位置 (左 上至右下的位置) 的 N个") ^的方阵为 N个零矩阵, 且不参与之后的矩 阵乘法运算; 将矩阵 P 2和各个单板的用户数据所构成的 Nw x l的列向量做 矩阵乘法运算, 得到一 Nw X 1的列向量作为外部处理数据。
步骤 1008、 集中矢量处理器将得到的外部处理数据回传至各个单板 的后期处理单元;
需要说明的是上述步骤 1003〜1005和步骤 1006〜1008可以是并行的两 个过程, 而没有先后顺序之分。
步骤 1009、 每个单板的后期处理单元接收该单板的外部处理数据, 并结合内部处理数据进行后期处理。
此步骤中的后期处理是指,每个单板的后期处理单元将该单板的各个 用户的内部处理数据和外部处理数据相加,得到该用户进行预编码处理后 的用户数据; 当然, 后期处理还可以包括, 将每个用户进行预编码处理后 的用户数据进行反向傅里叶变换, 之后发送给各个用户。
由于在本实施例提供的集中矢量处理器的矩阵乘法运算中,排除了矩 阵 P 2中对角线位置的 N 个《χ «的方阵的矩阵运算, 此过程具体最多需要 进行 4(N * ")2 _4 *"2个乘法运算及 Ν _3w*«2个加法运算, 故与现有技 术相比可以减少集中矢量处理器的运算量。 换言之。 集中矢量处理器的数 据处理能力一定的情况下, 可以增大矢量 DSL系统的用户容量。
需要说明的是, 若对于例如图 5所示的只有部分(至少一个)单板中 包含辅助矢量处理器的矢量 DSL 系统而言, 上述各步骤都只是针对至少 一个包含辅助矢量处理器的单板的用户数据进行的,而对于不包含辅助矢
量处理器的单板的用户数据可以采用现有技术的方法就可以实现,故在此 不力口赘述。
实施例五:
为了减少单板和集中矢量处理器之间的数据传输带宽,本发明实施例 对实施例四中针对图 6所示的矢量 DSL系统的方案进一步地优化。 相较 于实施四中的方案而言, 本发明提供的方案在进行步骤 1006之前还包括 步骤 Q1 或步骤 Q2 , 以得知各个单板上与其他单板线路强相关的用户, 并只将这些用户的数据传送给集中矢量处理器, 以达到减少带宽的目的。
步骤 Q1和步骤 Q2是两个并列的方案, 选择其一即可。
步骤 Ql、 每个单板选择该单板上的且与其他单板的用户线路强相关 的用户。 他单板的用户线路强相关的用户。
步骤 Q2、 集中矢量处理器选择每个单板上与其他单板的用户线路强 相关的用户;
此步骤可以为,集中矢量处理器根据经验值或者其他方式选择每个单 板上与其他单板的用户线路强相关的用户。 若利用此步骤, 则各个单板在 集中矢量处理器控制下,将该单板上与其他单板线路强相关的用户数据传 送给集中矢量处理器进行后续的数据处理。
此后, 进行步骤 1006〜1008与实施例四有所不同, 其他步骤可参考实 施例四中的描述。
步骤 1006、 各个单板将该单板上的且与其他单板的用户线路强相关 的用户数据传送给集中矢量处理器;
例如, 单板 1上的且与其他单板 (除单板 1之外的单板)的用户线路 强相关的用户有 个, 则单板 1将这 个用户的数据传送给集中矢量处理 器; 单板 2上的的且与其他单板(除单板 2之外的单板)的用户线路强相 关的用户有《2个, 则单板 2将这《2个用户的数据传送给集中矢量处理器;
同样地, 直至单板 N上的且与其他单板 (除单板 N之外的单板) 的用户 线路强相关的用户有 个, 则单板 N将这 个用户的数据传送给集中矢 量处理器; 且 ηλ < η ·, η2 <η , … , ηΝ≤η。
本发明实施例提供的矢量 DSL系统在每个单板与集中矢量处理器之 间的带宽需要, 参考实施例三中的计算方法得到
ml +m2 +--- + mN = 524 *(ηι +η1 +··· + ηΝ) Mbits/s 而利用现有技术中的集中式矢量 DSL 系统, 下行数据时所需要的入 口带宽为 524*N*w Mbits/s; 将本发明实施例提供的矢量 DSL系统与现有 技术中的集中式矢量 DSL系统相比, 由于 ^^+^+…+ ^ ^^ ", 故只要任一单板的至少一个用户与其他单板的用户不是线路强相关,则利 用本发明实施例提供的矢量 DSL 系统就可以比现有技术中的集中式矢量 DSL系统使用更小的带宽。
步骤 1007、 集中矢量处理器对通过步骤 1006得到的用户数据进行外 部矢量化处理, 得到外部处理数据。
在本发明实施例中, 由于通过步骤 1006得到的用户数据只是各个单 板上的且与其他单板的用户线路强相关的用户数据,故得到的用户数据的 数量小于或等于 个, 若得到的用户数据小于 个则将没有获取到的用 户数据置为 0, 从而仍可以用一 Nwxl的列向量来表示所有的用户数据。
所述外部矢量化处理具体为,集中矢量处理器生成一 NWXNW串扰消除 处理矩阵 P2 , 需要说明的是, 由于辅助矢量处理器已经处理了该辅助矢 量处理器所在单板的用户之间的远端串扰,故该矩阵 P2的对角线位置(左 上至右下的位置) 的 N个") ^的方阵为 N个零矩阵, 且不参与之后的矩 阵乘法运算; 将矩阵 P2和各个单板的用户数据所构成的 Nwxl的列向量做 矩阵乘法运算, 得到一 Nwxl的列向量作为外部处理数据。 由于置为 0 的 用户数据表示该用户和其他单板的任一用户之间都不是线路强相关,故在 进行矩阵运算的时候, 也可以将与置为 0的用户数据相关的计算省略, 从 而可以在保证不影响最终结果的前提下减少集中矢量处理器的运算量。
步骤 1008、 集中矢量处理器将得到的外部处理数据回传至各个单板 的后期处理单元。
此步骤具体可以为,集中矢量处理器将非 0的外部处理数据回传至各 个单板的后期处理单元。
下行数据的出口带宽与入口带宽有相同的带宽需求,与本发明实施例 步骤 1006的有益效果相同,利用此步骤可以减小集中矢量处理器与单板之 间的数据传输带宽。
需要说明的是, 对于例如图 5所示的只有部分(至少一个)单板中包 含辅助矢量处理器的矢量 DSL系统矢量 DSL系统而言,
上述步骤 Q1为, 所述包含所述辅助矢量处理器的单板还用于选择单 板上的与其他单板的用户线路强相关的用户。
上述步骤 Q2为, 所述集中矢量处理器还用于选择每个包含辅助矢量 处理器的单板上的与其他单板的用户线路强相关的用户。
同样针对只有部分 (至少一个) 单板中包含辅助矢量处理器的矢量 DSL系统矢量 DSL系统而言, 也可以减少单板与集中矢量处理器之间数 据传输的带宽。
实施例六:
如图 11所示, 本发明实施例还提供了一种应用于图 5或图 6所示的 矢量 DSL系统的单板, 该单板包括:
先期处理单元 1101 , 用于接收并先期处理所述单板的用户数据; 第一传送单元 1102 , 用于将所述单板的用户数据传送给该单板上的 辅助矢量处理器;
辅助矢量处理器 1103 , 用于对所述单板的用户数据进行内部矢量化 处理, 并输出内部处理数据;
第一接收单元 1104 , 用于接收由辅助矢量处理器输出的内部处理数 据;
第二传送单元 1105 , 用于将所述单板的用户数据传送给集中矢量处
理器进行外部矢量化处理;
第二接收单元 1106 , 用于接收由集中矢量处理器进行外部矢量化处 理后输出的外部处理数据;
后期处理单元 1107 , 用于将所述第一接收单元的内部处理数据和所 述第二接收单元的外部处理数据进行后期处理。
图 11所示的单板结构框图, 在图 5所示的矢量 DSL系统中, 只是应 用于单板 1和单板 2; 在图 6所示的矢量 DSL 系统中, 应用于全部的 N 个单板。
所述矢量化处理包括预编码处理和 /或串扰消除处理。 在用户上行数 据给 CO端的情况下, 所述矢量化处理是指串扰消除处理; 在 CO端下行 数据给用户的情况下, 所述矢量化处理是指预编码处理。 也就是说, 所述 单板可以只用于上行或者只用于下行, 当然, 可以既用于上行也用于下行 (在用户上行数据给 CO端的情况下, 但单板用于上行; 在 CO端下行数 据给用户的情况下, 该单板用于下行)。
所述集中矢量处理器可以集成在所述单板上,或者所述集中矢量处理 器独立于所述单板基板而单独存在。
进一步地, 如图 12所示所述单板还可以包括:
选择单元 1108 , 用于选择该单板上的与其他单板的用户线路强相关 的用户;
此时, 所述第二传送单元 1105 , 具体用于将该单板上的与其他单板 的用户线路强相关的用户数据传送给集中矢量处理器进行外部矢量化处 理;
所述第二接收单元 1106 , 具体用于接收该单板上的与其他单板的用 户线路强相关的用户数据由集中矢量处理器进行外部矢量化处理后输出 的外部处理数据。
本发明实施例提供的单板中由于辅助矢量处理器分担了一部分的数 据处理量, 故在系统用户数量不变的情况下, 可以减少集中矢量处理器的
数据处理量, 在集中矢量处理器的数据处理能力一定的情况下, 可以增大 矢量 DSL 系统的用户容量; 当然, 若集中矢量处理器的数据处理能力可 以得到提升, 那么运用本发明提供的方案可以进一步增大矢量 DSL 系统 的用户容量。
实施例七:
本发明实施例还提供了一种 DSLAM 设备, 该设备包括至少两个单 板, 其中至少一个所述单板如图 11所示, 包括:
先期处理单元 1101 , 用于接收并先期处理所述单板的用户数据; 第一传送单元 1102 , 用于将所述单板的用户数据传送给该单板上的 辅助矢量处理器;
辅助矢量处理器 1103 , 用于对所述单板的用户数据进行内部矢量化 处理, 并输出内部处理数据;
第一接收单元 1104 , 用于接收由辅助矢量处理器输出的内部处理数 据;
第二传送单元 1105 , 用于将所述单板的用户数据传送给集中矢量处 理器进行外部矢量化处理;
第二接收单元 1106 , 用于接收由集中矢量处理器进行外部矢量化处 理后输出的外部处理数据;
后期处理单元 1107 , 用于将所述第一接收单元的内部处理数据和所 述第二接收单元的外部处理数据进行后期处理。
例如, 图 5所示的矢量 DSL系统的 DSLAM设备上, 只有单板 1和 单板 2 的结构框图可以参考图 11 , 其他的单板与现有技术中不包含辅助 矢量处理器的单板相同或相似, 在此不加贅述; 图 6所示的矢量 DSL系 统的 DSLAM设备上, 全部的 N个单板的结构框图都可参考图 11。
进一步的, 所述单板如图 12所示还可以包括:
选择单元 1108 , 用于选择该单板上的与其他单板的用户线路强相关 的用户;
此时, 所述第二传送单元 1105 , 具体用于将该单板上的与其他单板 的用户线路强相关的用户数据传送给集中矢量处理器进行外部矢量化处 理;
所述第二接收单元 1106 , 具体用于接收该单板上的与其他单板的用 户线路强相关的用户数据由集中矢量处理器进行外部矢量化处理后输出 的外部处理数据。
图 5所示的矢量 DSL系统的 DSLAM设备中, 集中矢量处理器用于 对包含辅助矢量处理器的单板的用户数据进行外部矢量化处理,并对不包 含辅助矢量处理器的单板的用户数据进行内部矢量化处理和外部矢量化 处理。
本发明实施例提供的 DSLAM设备中的包含辅助矢量处理器的单板, 通过辅助矢量处理器分担了一部分的数据处理量,故在系统用户数量不变 的情况下, 可以减少集中矢量处理器的数据处理量, 在集中矢量处理器的 数据处理能力一定的情况下, 可以增大矢量 DSL系统的用户容量; 当然, 若集中矢量处理器的数据处理能力可以得到提升,那么运用本发明提供的 方案可以进一步增大矢量 DSL系统的用户容量。
通过以上的实施方式的描述,所属领域的技术人员可以清楚地了解到 本发明可借助软件加必需的通用硬件的方式来实现, 当然也可以通过硬 件, 但很多情况下前者是更佳的实施方式。 基于这样的理解, 本发明的技 术方案本质上或者说对现有技术做出贡献的部分可以在矢量 DSL 系统的 硬件架构的基础上以软件产品的形式体现出来,该计算机软件产品存储在 可读取的存储介质中, 如计算机的软盘, 硬盘或光盘等, 包括若干指令用 以使得一台计算机设备 (可以是个人计算机, 服务器, 或者网络设备等) 执行本发明各个实施例所述的方法。
以上所述, 仅为本发明的具体实施方式, 但本发明的保护范围并不局 限于此, 任何熟悉本技术领域的技术人员在本发明揭露的技术范围内, 可 轻易想到的变化或替换, 都应涵盖在本发明的保护范围之内。 因此, 本发
明的保护范围应以所述权利要求的保护范围为准。
Claims
1、 一种矢量 DSL的方法, 其特征在于, 包括:
至少一个单板接收并先期处理各自单板的用户数据; 所述单板将该单板的用户数据传送给该单板上的辅助矢量处理器进行 内部矢量化处理, 得到内部处理数据;
所述单板将其用户数据传送给集中矢量处理器进行外部矢量化处理, 得到外部处理数据;
所述单板接收该单板的外部处理数据, 并结合内部处理数据进行后期 处理。
2、 根据权利要求 1所述的方法, 其特征在于, 在所述单板将其用户数 据传送给集中矢量处理器进行外部矢量化处理, 得到外部处理数据之前, 还包括:
所述单板选择该单板上的与其他单板的用户线路强相关的用户;或者, 所述集中矢量处理器选择所述单板上的与其他单板的用户线路强相关 的用户;
所述单板将其用户数据传送给集中矢量处理器进行外部矢量化处理, 得到外部处理数据为:
所述单板将该单板上的与其他单板的用户线路强相关的用户数据传送 给集中矢量处理器进行外部矢量化处理, 得到外部处理数据。
3、 根据权利要求 1或 2所述的矢量 DSL的方法, 其特征在于, 所述 矢量化处理包括预编码处理和 /或串扰消除处理。
4、 一种单板, 其特征在于, 包括:
先期处理单元, 用于接收并先期处理所述单板的用户数据;
第一传送单元, 用于将所述单板的用户数据传送给该单板上的辅助矢 量处理器;
辅助矢量处理器, 用于对所述单板的用户数据进行内部矢量化处理, 并输出内部处理数据; 第一接收单元, 用于接收由辅助矢量处理器输出的内部处理数据; 第二传送单元, 用于将所述单板的用户数据传送给集中矢量处理器进 行外部矢量化处理;
第二接收单元, 用于接收由集中矢量处理器进行外部矢量化处理后输 出的外部处理数据;
后期处理单元, 用于将所述第一接收单元的内部处理数据和所述第二 接收单元的外部处理数据进行后期处理。
5、 根据权利要求 4所述的单板, 其特征在于, 还包括:
选择单元,用于选择该单板上的与其他单板的用户线路强相关的用户; 所述第二传送单元, 具体用于将该单板上的与其他单板的用户线路强 相关的用户数据传送给集中矢量处理器进行外部矢量化处理;
所述第二接收单元, 具体用于接收该单板上的与其他单板的用户线路 强相关的用户数据由集中矢量处理器进行外部矢量化处理后输出的外部处 理数据。
6、 一种数字用户线路接入复用器 DSLAM设备, 其特征在于, 至少一 个所述单板包括:
先期处理单元, 用于接收并先期处理所述单板的用户数据;
第一传送单元, 用于将所述单板的用户数据传送给该单板上的辅助矢 量处理器;
辅助矢量处理器, 用于对所述单板的用户数据进行内部矢量化处理, 并输出内部处理数据;
第一接收单元, 用于接收由辅助矢量处理器输出的内部处理数据; 第二传送单元, 用于将所述单板的用户数据传送给集中矢量处理器进 行外部矢量化处理;
第二接收单元, 用于接收由集中矢量处理器进行外部矢量化处理后输 出的外部处理数据;
后期处理单元, 用于将所述第一接收单元的内部处理数据和所述第二 接收单元的外部处理数据进行后期处理。
7、 根据权利要求 6所述的 DSLAM设备, 其特征在于, 所述单板还包 括: 选择单元, 用于选择该单板上的与其他单板的用户线路强相关的用户; 所述第二传送单元, 具体用于将该单板上的与其他单板的用户线路强 相关的用户数据传送给集中矢量处理器进行外部矢量化处理;
所述第二接收单元, 具体用于接收该单板上的与其他单板的用户线路 强相关的用户数据由集中矢量处理器进行外部矢量化处理后输出的外部处 理数据。
8、根据权利要求 6或 Ί所述的 DSLAM设备,其特征在于,所述 DSLAM 设备还包括:
集中矢量处理器, 用于对包含辅助矢量处理器的单板的用户数据进行 外部矢量化处理, 并对不包含辅助矢量处理器的单板的用户数据进行内部 矢量化处理和外部矢量化处理。
9、一种矢量 DSL系统,其特征在于,包括: DSLAM设备;所述 DSLAM 设备包括至少两个单板, 其中至少一个所述单板包含一辅助矢量处理器, 该辅助矢量处理器用于对其所在单板的用户数据进行内部矢量化处理, 并 且包含所述辅助处理器的单板的用户数据由集中矢量处理器进行外部矢量 化处理;
所述集中矢量处理器用于对包含辅助矢量处理器的单板的用户数据进 行外部矢量化处理, 并对不包含辅助矢量处理器的单板的用户数据进行内 部矢量化处理和外部矢量化处理。
10、 根据权利要求 9所述的矢量 DSL系统, 其特征在于, 所述包含所 关的用户, 并将该单板上的与其他单板的用户线路强相关的用户数据传送 给所述集中矢量处理器。
1 1、 根据权利要求 9所述的矢量 DSL系统, 其特征在于, 所述集中矢 量处理器还用于选择每个包含辅助矢量处理器的单板上的与其他单板的用 户线路强相关的用户, 并控制该用户所在单板将该用户的用户数据传送给 所述集中矢量处理器。
12、 根据权利要求 9〜11任一项权利要求所述的矢量 DSL系统, 其特 征在于, 所述集中矢量处理器设置在所述 DSLAM设备中, 或者独立于所 述 DSLAM设备。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP11782945.7A EP2672661B1 (en) | 2011-05-18 | 2011-05-18 | Vectored-digital subscriber line method, system, single board and digital subscriber line access multiplexer device |
CN201180000602.1A CN102217240B (zh) | 2011-05-18 | 2011-05-18 | 矢量dsl的方法、系统及一种单板和dslam设备 |
PCT/CN2011/074253 WO2011144028A2 (zh) | 2011-05-18 | 2011-05-18 | 矢量dsl的方法、系统及一种单板和dslam设备 |
US14/082,398 US9008163B2 (en) | 2011-05-18 | 2013-11-18 | Vectored-DSL method and system, board, and DSLAM device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2011/074253 WO2011144028A2 (zh) | 2011-05-18 | 2011-05-18 | 矢量dsl的方法、系统及一种单板和dslam设备 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/082,398 Continuation US9008163B2 (en) | 2011-05-18 | 2013-11-18 | Vectored-DSL method and system, board, and DSLAM device |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2011144028A2 true WO2011144028A2 (zh) | 2011-11-24 |
WO2011144028A3 WO2011144028A3 (zh) | 2012-04-26 |
Family
ID=44746746
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2011/074253 WO2011144028A2 (zh) | 2011-05-18 | 2011-05-18 | 矢量dsl的方法、系统及一种单板和dslam设备 |
Country Status (4)
Country | Link |
---|---|
US (1) | US9008163B2 (zh) |
EP (1) | EP2672661B1 (zh) |
CN (1) | CN102217240B (zh) |
WO (1) | WO2011144028A2 (zh) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015123815A1 (zh) * | 2014-02-19 | 2015-08-27 | 华为技术有限公司 | 信号处理方法、装置及系统 |
WO2018001451A1 (en) * | 2016-06-27 | 2018-01-04 | Huawei Technologies Co., Ltd. | Parallelization of a tomlinson-harashima-precoder for mimo transmission in a dsl system |
CN106407140B (zh) * | 2016-09-05 | 2019-06-11 | 华为技术有限公司 | 数据处理方法及装置 |
WO2019069128A1 (en) * | 2017-10-02 | 2019-04-11 | Lenovo (Singapore) Pte. Ltd. | METHOD AND APPARATUS FOR USING A DETERMINED COMPRESSION MATRIX TO FORM A SET OF COMPOSITE BEAMS |
Family Cites Families (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1109328B1 (en) * | 1999-12-14 | 2004-02-25 | STMicroelectronics S.A. | DSL transmission system with far-end crosstalk compensation |
US7688884B2 (en) | 2005-06-10 | 2010-03-30 | Adaptive Spectrum And Signal Alignment, Inc. | Vectored DSL nesting |
CN101047459B (zh) * | 2006-05-19 | 2011-05-04 | 华为技术有限公司 | 用于动态管理xDSL频谱的方法及其装置 |
CN101136659B (zh) | 2006-08-28 | 2012-01-25 | 华为技术有限公司 | 一种xDSL多线对串扰消除的方法、系统及设备 |
CN101166130B (zh) * | 2006-10-17 | 2010-05-19 | 中兴通讯股份有限公司 | 一种二层组播功能的分布式实现方法 |
EP1914905B1 (en) | 2006-10-20 | 2011-03-16 | Ericsson AB | Vectored digital subscriber line system having modular vectoring arrangements |
US8515017B2 (en) * | 2007-03-07 | 2013-08-20 | Ikanos Communications, Inc. | Systems and methods for loop termination detection based on per-port calibration |
EP2301216A4 (en) * | 2008-06-10 | 2016-06-22 | Ikanos Communications Inc | VECTED AD-ADDITIONAL ADDRESSING |
JP5406315B2 (ja) * | 2009-02-27 | 2014-02-05 | イカノス テクノロジー リミテッド | 自己誘導遠端クロストークを緩和するためのシステムおよび方法 |
EP2271021B1 (en) * | 2009-06-24 | 2012-01-04 | Alcatel Lucent | Joint signal processing across a plurarity of line termination cards |
CN103125104B (zh) * | 2010-07-22 | 2015-10-21 | 伊卡诺斯通讯公司 | 用于操作矢量化vdsl线路组的方法 |
CN103181092B (zh) * | 2010-10-15 | 2015-08-05 | 伊卡诺斯通讯公司 | Dsl外部噪声降低方法、dsl线卡设备及向量dsl系统 |
US8605567B2 (en) * | 2010-12-02 | 2013-12-10 | Adtran, Inc. | Apparatuses and methods for enabling crosstalk vectoring in expandable communication systems |
US8369485B2 (en) * | 2010-12-07 | 2013-02-05 | At&T Intellectual Property I, L.P. | Methods, apparatus, and articles of manufacture to trigger preemptive maintenance in vectored digital subscriber line (DSL) systems |
US8537655B2 (en) * | 2011-01-28 | 2013-09-17 | Alcatel Lucent | Multiplicative updating of precoder or postcoder matrices for crosstalk control in a communication system |
EP2678997A4 (en) * | 2011-02-23 | 2017-11-15 | Ikanos Communications, Inc. | System and method for partitioning dsl vector cancellation |
-
2011
- 2011-05-18 EP EP11782945.7A patent/EP2672661B1/en active Active
- 2011-05-18 CN CN201180000602.1A patent/CN102217240B/zh active Active
- 2011-05-18 WO PCT/CN2011/074253 patent/WO2011144028A2/zh active Application Filing
-
2013
- 2013-11-18 US US14/082,398 patent/US9008163B2/en active Active
Non-Patent Citations (2)
Title |
---|
None |
See also references of EP2672661A4 |
Also Published As
Publication number | Publication date |
---|---|
EP2672661A2 (en) | 2013-12-11 |
CN102217240B (zh) | 2014-02-19 |
US20140079105A1 (en) | 2014-03-20 |
EP2672661B1 (en) | 2014-09-10 |
EP2672661A4 (en) | 2014-01-08 |
WO2011144028A3 (zh) | 2012-04-26 |
CN102217240A (zh) | 2011-10-12 |
US9008163B2 (en) | 2015-04-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5335137B2 (ja) | 複数の回線終端カード間のジョイント信号処理 | |
CN103891154B (zh) | Xdsl中调整的传送 | |
WO2008067765A1 (fr) | Système de traitement de signal, puce, carte externe, dispositif de filtre et procédé de traitement de signal | |
JP2002084260A (ja) | 多重入力・多重出力dslシステムのための並列処理 | |
CN103828253B (zh) | 串扰测量期间的功率调整避免 | |
US8964884B2 (en) | Power control in linear precoder design for MIMO DSL transmission | |
US9020145B2 (en) | MIMO mechanism for strong FEXT mitigation | |
CN107454995B (zh) | 用于数字用户线的多用户多输入多输出的相关方法和装置 | |
WO2008031353A1 (fr) | Procédé, système et dispositif de suppression de diaphonie xdsl | |
WO2011144028A2 (zh) | 矢量dsl的方法、系统及一种单板和dslam设备 | |
WO2014107857A1 (zh) | 一种dsl系统信号处理方法、装置及系统 | |
CN103229445A (zh) | 一种线路的初始化方法及设备 | |
Ginis et al. | Alien crosstalk cancellation for multipair digital subscriber line systems | |
EP2986062B1 (en) | Power control method, device and system | |
JP6250182B2 (ja) | 低パワーモードのためのベクタリング | |
WO2012109853A1 (zh) | 通过预编码消除远端串音的方法和装置 | |
WO2017045207A1 (zh) | 信号处理方法、装置和系统 | |
WO2008145537A1 (en) | Method and device for data processing and communication system comprising such device | |
CN111835383B (zh) | 回音与近端串扰消除系统 | |
CN113838474B (zh) | 通信系统啸叫抑制方法和装置 | |
US10069533B2 (en) | Methods and systems for reducing crosstalk using sequential non-linear vectoring | |
CN106716853B (zh) | Dsl向量化系统中的方法和布置 | |
CN109644019B (zh) | 通过金属线对传输数据的方法和装置 | |
CN107852385B (zh) | 一种线路速率提升方法和装置 | |
CN108347393B (zh) | 交叉线序的处理方法、设备和系统 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201180000602.1 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2011782945 Country of ref document: EP |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11782945 Country of ref document: EP Kind code of ref document: A2 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |