FR2814018A1 - METHOD AND DEVICE FOR MANAGING VOICE STAMPS IN DYNAMIC BANDWIDTH CIRCUIT CIRCUIT EMULATION SERVICES - Google Patents

Abstract

According to a SAR (segmentation and reassembly) buffer management method which operates in DBCES (dynamic bandwidth circuit emulation service) mode, first of all an active channel buffer limit is determined. This depends on the number of active channels in the virtual circuit. Then the amount of buffering for each active channel is set to be equal to CDV (cell delay variation) plus the active channel buffer limit. This prevents the occurrence of underflow and overflow when the number of active channels suddenly changes.

Description

FIELD OF THE INVENTION

  The present invention relates to telecommunications and more particularly to Dynamic Bandwidth Circuit (DBCES) emulation services in a cell relay network such as ATM.

BACKGROUND OF THE INVENTION

  Asynchronous Transfer Mode (ATM) is a service that transports data in small packets or small cells of fixed size over a packet-switched network. The cells are statistically multiplexed on fixed physical links between network nodes and establish virtual circuits on the network between endpoints. ATM includes a number of specifications among which AAL1 (ATM Tuning Layer 1) establishes a standard for carrying time-sensitive data such as voice or video information on virtual circuits between endpoints. . This is known as Circuit Emulation (CES) because it provides a number of voice channels that appear to the user to be similar to

  example of times division multiplexed channels.

  The DBCES mode is a dynamic bandwidth usage mode in an ATM network based on the detection of the time windows of a given TDM route that are active and those that are inactive. When an idle state is detected in a specific time window, the time window is removed from the next ATM structure and the bandwidth it was using can be reused for other services. The DBCES mode is described in detail in ATM

  Forum specification: af-vtoa-0085.000 (July 1997).

  Devices known as SAR devices (Segmentation and Reassembly) convert incoming data into cells and vice versa. SARs include buffers to store

  cells to be processed.

  When operating in DBCES (Dynamic Bandwidth Circuit Emulation) mode, if the number of voice channels carried by the AAL1 cell stream is changed, the rate at which the cells arrive at the receiver varies. This variation of

  arrival flow may generate an overflow or overflow

  overflow into the voice buffer at the receiver that receives incoming cells unless special precautions are taken. In order to address this problem, it has been proposed to add a

  additional buffer allocation to prevent

  overflows and insert an extra delay into the buffer to prevent over-overflows. Alternatively, another solution is to start the buffer with a specific number of voice channels that are active to establish an adequate buffer fill but the buffer can not be restored automatically to the buffer.

following slip errors.

  Finally, another solution is simply to let overflow and overflow occur; however, this should cause receivers to receive incorrect data, which is transposed according to pinging sounds at the

received audio information.

OBJECT AND SUMMARY OF THE INVENTION

  According to a broad aspect, the invention proposes a method allowing

  manage buffers in a SAR device (segmentation and reorganization)

  assembly) that operates in a dynamic bandwidth circuit emulation (DBCES) mode, comprising: determining an active channel buffer limit for a virtual circuit based on the number of active channels; and setting the buffering amount for each active channel such that it is equal to the CDV (cell delay variation) plus the active channel buffer limit. The amount of buffering is typically represented by the distance between the write pointer and the read pointer for the circular buffers. According to the invention, for any current value of active voice channels, either N, a defined minimum number of active voice channels, or MIN, and a desired cell delay variation tolerance, or CDV, a pitch spacing of write pointer (slip pointer) relative to the read pointer is

  calculated. Following a start or error (that is, a sub-

  overflow or overflow), this desired spacing is used to create a new value for the write pointer. Authorizing various definitions of the MIN allows the system user to compromise the delay in the buffer for the bandwidth in the cell stream. Recalculating the desired spacing as N varies allows the buffer to manipulate any number of voice channels up to MIN. A restore from errors is performed automatically (without intervention of the central processing unit or user CPU) through the use of the correct buffer spacing. The cell stream can be started with any initial value, N, of channels

active vocals.

  The inclusion of the Cell Delay Variation (CDV) in the calculation means that any cell delay variation in the received cell stream does not generate slip errors (i.e.

  overflow or overflow).

  Typically, the spacing is set so that write pointer = read pointer + CDV + (47 / MIN) - (47 / N) o 47 is used because each cell in ATM AAL1 contains

  approximately 47 bytes of voice data.

BRIEF DESCRIPTION OF THE DRAWINGS

  The invention will now be described in more detail by way of example only with reference to the accompanying drawings, in which: Figs. 1a and 1b are functional diagrams which show the circular buffers according to SAR for both SDT modes of operation and DBCES of a SAR; and Fig. 2 is a block diagram showing the re-assembly control structure for a SAR in a preferred implementation.

  DETAILED DESCRIPTION OF THE INVENTION

  A structured data transfer mode (SDT) is an AAL1 data transfer mode in which data is first structured into blocks which are then segmented into cells for transfer. In a typical application, TDM data is transferred on an ATM virtual circuit between a transmitting or

  segmentation (TXSAR) and a reception or re-

  assembly (RXSAR) which, as noted above, convert data according to ATM cells and vice versa. The virtual circuit

  carries a number of channels emulating a TDM service.

  The RX_SAR SDT is responsible for writing data in

  from cells received in circular buffers

  while the TDM reads the data from the buffers and outputs them as TDM channels. However, some buffer control is required because the RX_SAR SDT and TDM operate at different clock rates (the data is placed in the buffers based on the arrival rate of the cells (which varies) while the data is read from the buffers at a constant rate that is determined by the clock frequency

  TDM output (eg 2048 MHz)).

  In the SDT and DBCES modes of operation, one of the

  The main tasks of the RX_SAR SDT is to ensure that no

  Overflow or no overflow occurs between the RX_SAR SDT and a TDM module that converts the data to time-division multiplexed channels. Overflow occurs when there is insufficient data in the

  buffer to power the TDM module at TDM rate. An over-

  Overflow occurs when the buffer overflows because the cells arrive faster than the data is

read by the TDM module.

  The RX_SAR SDT is responsible for maintaining a relationship between the RX_SAR SDT write pointer and the TDM read pointer. This command is provided by a subroutine which, whenever the RX_SAR SDT is about to write to the circular buffer, determines the distance between the write pointer and the read pointer. Normally, each write is performed when the previous write is complete and the write pointer simply advances as the bytes are written. However, when starting or when the RX_SAR SDT algorithm has decided that a

  slip error (ie overflow or underflow)

  overflow) occurred or during a transition from zero active channel to at least one active channel, the RX_SAR SDT uses an internally calculated write pointer value as a value of

correct writing pointer.

  In these cases, regardless of the value of the original write pointer, the write pointer is set to the pointer value

  slip and data are written on the circular buffer of

  assembly at the location pointed by the slip pointer. The purpose of using the slip pointer is to establish a known relationship between the read pointer and the write pointer, to eliminate the risk of multiple successive slips (eg a displacement of the write pointer that would distance it more of the read pointer to prevent overflows from occurring again and again). The invention can be implemented in any suitable SAR device. An appropriate SAR device is, for example, the Mitel MT90528 device. Figure la represents a circular buffer 10 for an RXSAR in an SDT mode of operation. Data is stored at individual addresses 12 in the buffer. A

  Write pointer 16 determines where incoming cells are written.

  A read pointer 14 determines where the outgoing TDM data is

read.

  Fig. 1b shows a similar arrangement for a mode

DBCES operation.

  Slip pointer values in SDT and DBCES operation modes A predetermined spacing is maintained between the read pointer and the write pointer. An explanation of the calculations is presented

  below. The calculations use an excess rounding.

  If the virtual circuit is configured for a basic SDT (Figure la), the slip pointer is assigned a value such that it is equal to TDM + CDV read pointer (CDV is called "avg lead" or advance average and is equal to the value "Maximum_Lead / 2" programmable by the user). CDV represents the cell delay variation in which cells will arrive on the virtual channel at the

reception SAR.

  On average, the cells should arrive at the RXSAR SDT at a steady rate (Fig. La) so that the cell data is always written to the circular buffer at the TDM + CDV read pointer time. However, due to cell delay variation, it is possible for cells to arrive more slowly or more rapidly than average. Therefore, sufficient buffering must be provided in the circular buffers so that if a cell arrives late, there is enough (previously received) data in the buffer to ensure that the TDM can continue reading. from the circular buffer at a rate

  regular without undergoing overflow.

  If the virtual channel is configured for DBCES and the DBCES_Control bit <2> is not set, the DBCES version of the RX_SAR SDT drag pointer is the same as the SDT version of the slip pointer (ie to say that it is equal to read pointer

CT + CDV).

  The problem in this situation lies in the fact that when the number N of active channels arriving on a virtual circuit varies, the arrival rate of the cells also varies. For example, in the case of two channels, each cell consumes two bytes of each TDM frame so that about 24 frames are consumed to fill the payload of a cell. A VC or 4-channel virtual circuit consumes 4 bytes of each TDM frame so that only 12 frames are consumed to fill a cell. In the case of a single channel, only one byte of each frame is consumed so that a total of 47 TDM frames are consumed before a cell is full and ready to be transmitted. If it is assumed that a cell arrives and is written to the circular buffer at the location pointed by the slip pointer (i.e., a mean distance from the read pointer), no sub- Overflow should not occur. However, if the number of active channels on the virtual circuit decreases, a longer amount of time will be consumed for a cell to be formed at the transmitter. As a result, the arrival rate of the cells will decrease. As a result, the TDM module will continue to perform a

  reading from the circular buffer at its regular rate and will be subdivided

  overflow because there will not be enough data in the buffer to allow the TDM to continue reading until

what the next cell arrives.

  Figure 2 shows a control structure for the DBCES format. The DBCES_Control field determines the amount of buffered data that the user wants to store in each circular channel reassembly buffer. If this field is set, the bit <2> controls whether this variable is used or not. Bits <1: 0> represent the desired minimum delay buffering for each channel; "00" = 47 bytes (therefore, the minimum number of active channels in the VC will be 1); "01" = 24 bytes (therefore, the minimum number of active channels in the VC will be 2); "10" = 16 bytes (therefore, the minimum number of active channels in the VC will be 3); "11" = 12 bytes (therefore, the minimum number of

  Active channels in the VC will be 4).

  In accordance with the principles of the invention, if the virtual circuit is configured for DBCES and the <2> bit of DBCES_Control <2> (see Fig. 2) is set, the write pointer of RX_SAR SDT is assigned to be equal to TDM read pointer + CDV + buffer limit

  Active channel CDV for the virtual circuit.

  The active channel buffer limit for the VC is calculated as the number of bytes of the DBCES buffering desired by the user (as configured by setting the DBCESControl field in the DBCES re-assembly command structure for the VC) minus the maximum number of bytes that can be written in each circular re-assembly buffer SDT for the VC, given the activity profile of the virtual circuit (for example, if there are 2 active channels in the virtual circuit, a maximum of 24 bytes of data can be written on any circular buffer following

the arrival of a cell).

  DBCES_Control is used to ensure "clean" buffering when the number of active channels in a VC can vary in large amounts. When the user sets the DBCES_Control field, it sets the minimum number of buffering bytes that the user always wants to have available in each circular SDT reassembly buffer for the VC. This extra buffering adds delay in terms of the data that is being sent

  output on the TDM bus but is useful for

  buffer overflows when the number of active channels in a

VC varies.

  A specific example will now be considered.

  It is assumed that the user has determined that the expected CDV between cells on the virtual circuit is 625 microseconds (5 bytes of data in the buffer). Furthermore, the user has decided that it will allow the number of active channels in the virtual circuit to decrease to a minimum equal to unity (which requires storage

  47 bytes of additional data in the buffers).

  Now, it is assumed that there are four active channels in the virtual circuit to boot. Therefore, using the formula mentioned above: distance to read pointer = CDV + active channel buffer limit = 5 bytes + (47 bytes - 47 bytes / 4) = 40 bytes, write pointer will start writing (following startup or following an error condition) in a location that is 40 bytes in front of the read pointer. After receiving a cell, a quarter of the cell, that is 12 bytes, will be added to the buffer so that the write pointer will then be 52 bytes in front of the pointer

reading.

  In the case where a cell with N = 1 (i.e. only one channel) is sent, 47 TDM frames will be consumed for the cell to be assembled for reception in the RX_SAR receive buffer. During this time, 47 bytes of data will be read from the buffer by the read pointer. Therefore, when the next cell arrives, the write pointer will still be 5 bytes to the front of the read pointer so that no overflow

will not happen.

  distance to read pointer = CDV = 5 bytes + active buffer buffer limit = 5 bytes + (47 bytes - 47 bytes / 1) = 5 bytes However, without DBCES_Control we should have:

  distance to reading pointer = CDV = 5 bytes.

  In this case, if the number of active channels in the virtual circuit drops to unity, the TDM module will read the data stored in the buffer in 5 frames and then start reading old data because new cells will not be no arrivals. Therefore,

  an overflow should occur.

  A user programming capability of the desired buffering can be provided through the use of a DBCES_Control area in each DBCES re-assembly control structure, as can be seen in FIG. user has the option to either enable or disable this additional buffering by using the <2> bit in the DBCESControl field. The <1: 0> bits of the DBCES_Control field are used to select the amount of

  additional buffering required in the circular buffer of

  assembly of each channel inside the VC. The options are provided to the user based on the minimum number of active channels that will be transported on the VC. When less active channels are transported, the amount of buffering that will be required will be greater since the cell arrival rate decreases when there are fewer channels in the VC (longer time is consumed to collect enough data to fill a cell). According to the example, the user has 4 options: "00" = minimum of 1 active channel; "01" = minimum of 2 active channels; "10" =

  minimum of 3 active channels; "11" = minimum of 4 active channels.

  Various methods can be used to present or program CDV and MIN (for example, allows MIN = 1 channel to 4 channels and allows CDV = Maximum_Lead / 2 (CDV is approximately between 125 microseconds and 64 milliseconds)). The method can be used in conjunction with a

  overflow and overflow detection algorithm.

  The invention therefore proposes an efficient way

  to prevent the problem of over-overflows and under-

  overflows that can occur when the number of channels is

modified in a DBCES service.

Claims (10)

  A method of managing buffers in a SAR device (segmentation and re-assembly) that operates in a dynamic bandwidth circuit emulation (DBCES) mode, characterized in that it comprises: determining a limit active channel buffer for a virtual circuit based on the number of active channels; and setting the buffering amount for each active channel so that it is equal to the VDC (variation in delay of   cell) plus the active channel buffer limit.   Method according to claim 1, characterized in that each buffer comprises a write pointer and a read pointer and the distance between the write pointer and the read pointer   determines the amount of buffering.   Method according to claim 1 or 2, characterized in that the active channel buffer limit is equal to the number of bytes of payload data per cell divided by the minimum number of allowed channels minus the number of bytes of payload data   per cell divided by the number of current active channels.   4. Method according to claim 3, characterized in that the amount of buffering is recalculated when the number of channels varies so as to ensure adequate buffering the next time o   a change in the number of channels occurs.   Method according to claim 4, characterized in that said method can be activated by setting a control bit in a   control structure in the buffers.   6. Method according to claim 4, characterized in that   said buffers are circular buffers.   7. Method according to claim 1, characterized in that said   DBCES is implemented in an ATM network.   8. A control data structure for a SAR device (segmentation and re-assembly) for implementing a dynamic bandwidth circuit emulation service (DBCES), characterized in that it comprises: a control zone; a command with a bit that determines whether or not buffer management is active; and bits that determine the minimum buffering for each channel; and in that: the buffering amount for each active channel is set to be equal to the CDV (cell delay variation) minus one   active channel buffer limit, which depends on the number of active channels.   The control data structure according to claim 8, characterized in that the active channel buffer limit is equal to the number of bytes of payload data per cell divided by the minimum number of allowed channels minus the number of bytes of payload data per cell divided by the number of current active channels. The control data structure according to claim 8, characterized in that the active channel limit is 47 / MIN, where MIN is the minimum number of N channels considered, minus 47 / N o N is the number of current active channels.