KR100263185B1 - Apparatus and method of distributed processing with two-processor in voice mailing system - Google Patents

Abstract

PURPOSE: A Distributed processing method using two processors and a system for the same are provided to process a task which is performed in one line interface part dividedly in two control parts, and deal with many ports in the line interface part. CONSTITUTION: A main control part(211) controls overall operations of a voice processing system. A line interface part(214) is connected between a switching center(217) and a VME bus. The part(214) codes the voice data by compression, which is received when call connection is made, in order to reduce the amount of the data. The part(214) recovers the compressed voice data to original data. An SCSI control part(216) is connected between the VME bus and a message storing part(215). The SCSI control part(216) supplies a control signal for recording and reading the voice data to the part(215). The message storing part(215) is connected to the SCSI control part(216) via an SCSI bus. The message storing part(215) stores the voice data under the control of the SCSI control part(216).

Description

Distributed processing apparatus and method using two processors {APPARATUS AND METHOD OF DISTRIBUTED PROCESSING WITH TWO-PROCESSOR IN VOICE MAILING SYSTEM}

The present invention relates to a speech processing system, and more particularly, to an apparatus in which a line interface unit is distributedly controlled by two control units.

The operation of a general speech processing system is as follows.

When the call is connected through the line interface unit, the call connection is notified to the MPU (Master Processing Unit (hereinafter referred to as a main processing unit)), and the main processing unit that recognizes this transfers an appropriate command to the line interface unit. Then, the line interface unit records voice, fax, or general data applied through the line to the hard disk under the control of the main processing apparatus, or transmits the data recorded on the hard disk through the line. . The hard disk is controlled by the SCSI controller, which is responsible for data transfer from the line interface unit to the hard disk, or from the hard disk unit to the line interface unit.

Here, the detailed configuration and operation of the line interface unit will be described with reference to FIG. 1.

The line interface unit includes a plurality of input / output devices (I / O devices) 114 that handle voice processing, dual tone multi frequency (DTMF), and tone processing, a control unit (CPU) 111 for controlling them, and the main processing device; And a common memory 113 for command interface, a memory 112 for storing simple voice messages and service data, etc., and an E1 interface unit 115 for interfacing with the exchanger.

Referring to FIG. 1, the control unit 111 analyzes the data processed by the plurality of input / output devices 114 and transmits the data to the upper main processing device through the common memory 113 or inputs data transmitted from the main processing device from the common memory 113. It receives and delivers to the input and output device 114.

As described above, the line interface unit according to the prior art has a limit of capacity that can be processed by one board because one controller controls all input / output devices. In other words, the number of ports (subscriber ports) that can be processed in the board cannot be increased. Nowadays, the trend requires a large number of ports in one board, so it is urgent to present a structure that can handle more ports in a board.

Accordingly, an object of the present invention is to provide an apparatus capable of dividing and processing work performed in one line interface unit.

Another object of the present invention is to provide an apparatus capable of processing a larger number of ports in one line interface unit.

A voice processing system for storing voice data from an exchange on a disc and transmitting voice data from the disc to the exchange for achieving the above objects comprises: a main control unit controlling the operation of the voice processing system, and voice data; A plurality of line interface units for compressing a message storage unit, voice data connected from the exchanger, storing the compressed voice data, storing the compressed voice data, and outputting the compressed voice data to the exchange unit; A separate second processor for processing only a first processor and a first processor for controlling a general operation to distribute and process work, to record voice data in the message storage and to read voice data from the message storage. It characterized in that it comprises a SCSI control unit for controlling.

It is still another object of the present invention to provide a method for distributing and processing voice processing using two processors by a line interface unit of a voice processing system.

A first processor for processing various tones and DTMF for achieving the above object, a second processor dedicated to processing the voice data, and compression and decompression of the voice data under the control of the first processor and the second processor And a plurality of slave DSPs to process, an E1 interface unit responsible for E1 frame interfacing with the exchanger, and a SCSI switching unit responsible for switching the E1 frame between the plurality of slave DSPs and the E1 interface unit. In the voice processing system, the call processing method, when the call is connected, the E1 interface unit notifies the first processor of the call reception, and the first processor allocates a dormant port among the plurality of slave DSPs to receive data. And the frame switching unit connects the frame received from the E1 interface to the pause port. Setting up a call, and notifying that the first processor has established a call to a higher main controller after setting up the call, delivering a predetermined voice message to the slave DSP by the second processor, and transmitting the voice data. Receiving the slave message by restoring the voice message and transmitting the voice message to the exchange through the SCSI switching unit and the E1 interface unit; and if there is DTMF detection after the voice message is transmitted, the slave DSP analyzes the message and performs the first message. And notifying, by the first processor, the first processor to transmit the analysis data to the upper control unit again, and the upper main control unit reads the corresponding voice data from the disc in response to the analysis data and delivers the analyzed data to the second processor. The second processor transfers the voice data to a predetermined slave DSP. It reaches reconstructing the voice data, characterized by comprising the step of outputting to the exchange voice data to the slave DSP is the recovered through the switching unit and the Scar four E1 interfaces.

1 is a detailed configuration diagram of a line interface unit of a speech processing system according to the prior art.

2 is an overall configuration diagram of a speech processing system according to the present invention;

3 is a detailed configuration diagram of a line interface in the configuration of FIG.

4 is a view for explaining an example of distributed processing by two controllers according to an embodiment of the present invention;

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, in adding the reference numerals to the components of each drawing, the same components have the same reference numerals as much as possible even if displayed on different drawings. In describing the present invention, when it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

2 is an overall configuration diagram of a speech processing system according to an embodiment of the present invention.

The main control unit 211 is a master processing unit (MPU) that controls the overall operation of the speech processing system. The line interface unit 214 is connected between the switch 217 and the VME bus, and is in charge of the connection with the switch 217 to compress and encode the voice data received during the call connection to reduce the amount of information, or to convert the compressed data into the original voice data. Restore it and print it out. The SCSI controller 216 is connected between the VME bus and the message storage unit 215 and generates a control signal for recording and reading voice data in the message storage unit 215. The message storage unit 215 is connected to the SCSI controller 216 through a SCSI bus and stores voice data received under the control of the SCSI controller 216.

The line interface unit 214 performs distributed processing using two processors, and thus can accommodate more ports than in the prior art. In the present embodiment, one line interface unit 214 can accommodate 32 ports, so the total number of ports that the voice processing system can accommodate is 256 (32 ports x 8 line interface units).

Referring to the operation of the voice processing system as shown in FIG. 2, when a call is connected through the line interface 214 from the switch 217, the main control unit 211 is notified to the main control unit 211. Output to the line interface unit 214. Then, the line interface unit 214 outputs voice, facsimile or general data received from the exchange 217 on the VME bus under the control of the main controller 211, or receives data on the VME bus and outputs the data through the SCSA bus. do. The SCSI controller 216 is connected to the VME bus and is responsible for data transfer from the line interface unit 214 to the message storage unit 215 or from the message storage unit 215 to the line interface unit 214.

3 is a detailed block diagram of the line interface unit 214 in the configuration of FIG. 2.

The line interface unit 214 includes a control unit (or first processor) 312 for controlling the overall operation of the line interface unit, a memory 313 in which a program of the control unit 312 is stored, and a common controller for communicating with an upper control unit (main control unit 211). A memory 1, a digital signal processor (DSP) 315 for allocating and processing the work of the controller 312, and a common memory 2 for the digital processor 315 to communicate with the upper controller; And a plurality of slave digital processing units (SLAVE DSP1..8) 318 for processing data input and output from respective ports, and a plurality of slave digital processing units used by the control unit 312 and the digital processing unit 315 to control them. Buffer (CPU buffer, DSP buffer), the E1 interface unit 320 for interfacing with the E1 line, and the E1 buffer received from the E1 interface unit 320. It consists of a SCSI switching unit (SCSA SC4000) to switch the frame to the corresponding port.

The main function of the line interface unit 214 is to process voice, tone, and DTMF to communicate with the upper control unit (main control unit 211). At this time, the data requiring the most processing among the processed data is voice data. Therefore, using two control units (control unit 312, digital processing unit 315), one control unit (digital processing unit 315) is in charge of only voice processing, and the other control unit (control unit 312) processes all the other one board (line interface unit). To allow more ports to be handled. In addition, in the present invention, one slave digital processing unit 318 processes four ports per one, so that the line interface unit 214 can process all 32 ports.

Here, if the description of the above configuration based on the signal flow as follows.

The slave digital processor 318 compresses and expands the voice input and output to each port, and detects TONE / DTMF and notifies the controller 312. In addition, the slave digital processing unit 318 processes four ports per one, and therefore, the eight digital processing units 318 according to the embodiment of the present invention all process 32.

The digital processor 315 receives the command of the controller 312 and transmits the compressed data input from the slave digital processor 318 to the main controller 211, or transmits the voice data input from the main controller 211 to the slave digital processor 318.

The controller 312 recognizes the TONE / DTMF processed and transmitted by the slave digital processor 318 and controls the entire line interface unit 214 to operate.

The E1 interface 320 converts the data received from the switch 217 into pulse code modulation (PCM) data, or converts the data received from the slave digital processor 318 into a signal that the switch 217 can receive and transmit. In addition, when a call request is made, the controller 312 is notified or a command of the controller 312 is received to inform the switch 217 that a call request is made.

The SCSI switching unit 319 performs a switching role between the E1 interface 320 and the slave digital processor 318.

Hereinafter, an operation according to an embodiment of the present invention based on the configuration of FIG. 3 will be described.

When a call is received from a user who wants to use the voice processing system, the E1 interface 320 detects that there is a call request and notifies the controller 312, and the controller 312 idles among all ports of the slave digital processor 318. Allocate an in port and get ready to receive a signal. Accordingly, the controller 312 sets up a call by connecting the slave digital processor 318 to the E1 interface 320 through the casa switching unit 319.

The controller 312 notifies the main controller 211 that a call is formed and requests the first service voice data to be sent. Then, the main controller 211 immediately sends the first voice data through the voice memory 2, and the controller 312 confirms that the voice data has arrived and notifies the digital processor 315 of the arrival of the voice data.

Then, the digital processor 315 transfers the voice data of the voice memory 314 to the slave digital processor 318. In this case, the slave digital processor 318 decompresses the voice data, converts the voice data into PCM data, and sends it to the E1 interface 320 through the SCSI switching unit 319. The E1 interface 320 transmits the PCM data to the switch 217. .

At this time, when the user listens to the first voice data and inputs a DTMF (button), it is analyzed by the slave digital processor 318, and the analysis data is notified to the controller 312. The controller 312 transfers the data to the main controller 211 through the common memory 1. Therefore, the main controller 211 transmits the voice data corresponding to the input DTMF again.

Here, the operation of accessing the slave digital processing unit 318 by the controller 312 and the digital processing unit 318 is performed by 16msec TIME SHARING.

4 is a diagram illustrating a method in which the controller 312 and the digital processor 315 access the slave digital processor 318.

Referring to FIG. 4, the timer interrupt 316 allows the digital processing unit 315 to access slave digital processing units 1 to 4 and the control unit 312 to slave digital processing units 5 to 8 during the first 16ms interrupt, while the digital processing unit during the next 16ms interrupt. 315 interrupts the slave digital processor 5 through 8 and the controller 312 accesses the slave digital processor 318 1 through 4.

Here, the audio compression method performed by the slave digital processor 318 processes at 32 KBPS using ADPCM, which requires a processing speed of 64 bytes per 16 msec. Since each port of the slave digital processor 318 is accessed every 32 msec, the amount of data processed in 32 msec is 128 bytes, i.e., 128 bytes per 32 msec. In the case of assuming that the port to be processed within 16 msec is 16 ports as in the present embodiment, the total throughput is 128 × 16 × 2 KBYTES. Therefore, it is enough to process 2KBYTES ADPCM data within 16msec.

In addition, since it takes 500 nsec / 1BYTES for the digital processor 315 to transfer data to the slave digital processor 318, it is sufficient to move 2 KBYTES within 16 msec. In addition, since the controller 312 accesses and processes the slave digital processor 318, the amount of data is not so large. Therefore, it is not a matter of time for the controller 312 to access and process the slave digital processor 318.

As described above, the present invention implements two control units (DPUs) in the line interface unit (line card) of the voice processing system to divide the work in one line board to process more subscriber ports. Can be. That is, it provides an advantage to maximize the system capacity.

Claims (4)

A voice processing system for storing voice data from an exchange on a disc and transmitting voice data from the disc to the exchange, A main control unit for controlling an operation of the voice processing system; A message storage unit for storing voice data, A plurality of line interface units for compressing the voice data connected from the exchanger, storing the compressed voice data in the message storage unit, restoring the compressed voice data, and outputting the compressed voice data to the exchange unit; It has two processors and a first processor that controls the general operation to distribute work And a SCSI control unit for recording voice data in the message storage unit and controlling the reading of the voice data from the message storage unit. A line interface device provided in a voice processing system, for compressing voice data from an exchanger, storing the compressed data on a disc, and restoring the compressed data to the exchanger. A first processor which controls the operation of the line interface and processes various tones and DTMFs; A second processor dedicated to processing the voice data; A plurality of slave DSPs for compressing and restoring the voice data under the control of the first processor and the second processor; An E1 interface unit in charge of E1 frame interfacing with the exchanger; And a SCSI switching unit in charge of switching the E1 frame between the plurality of slave DSPs and the E1 interface unit. A first processor for processing various tones and DTMF, a second processor dedicated to processing the voice data, and a plurality of slave DSPs for compressing and restoring the voice data under the control of the first processor and the second processor Call processing in a voice processing system including an E1 interface unit for handling an E1 frame interface with the switch, and a SCSI switching unit for switching the E1 frame between the plurality of slave DSPs and the E1 interface unit. In the method, The E1 interface unit notifies the first processor of the call reception when the call is connected, and the first processor allocates an idle port among the plurality of slave DSPs and prepares to receive data; Setting up a call by connecting the frame received from the ESC interface to the pause port by the SCSI switch; After the call setup, the first processor notifies that a call is formed to a higher main controller, and the second processor transmits a predetermined voice message to the slave DSP; Receiving, by the slave DSP, the voice message and reconstructing the voice message and transmitting the voice message to the exchange through the SCSI switching unit and the E1 interface unit; If there is DTMF detection after transmitting the voice message, the slave DSP analyzes it and notifies the first processor, and the first processor transmits the analysis data to the upper control unit again, and the upper main control unit transmits the analysis data to the analysis data. In response to reading the voice data from the disc and delivering the voice data to the second processor; Transmitting, by the second processor, the voice data to a predetermined slave DSP to restore the voice data, and outputting, by the slave DSP, the restored voice data to the exchange through the casa switching unit and the E1 interface unit. Characterized in that. The method of claim 3, And the first processor and the second processor share and access the plurality of slave DSPs at predetermined intervals.

Images