WO1999014865A1

WO1999014865A1 - System and method for data rate adjustments in modem communications

Info

Publication number: WO1999014865A1
Application number: PCT/US1998/019171
Authority: WO
Inventors: Daniel L. Moore
Original assignee: Diamond Multimedia Systems, Inc.
Priority date: 1997-09-15
Filing date: 1998-09-14
Publication date: 1999-03-25

Abstract

A method in accordance with the invention allows the error-correcting protocol utilized by the modem (104) to select the data rate that will maximize data throughput. To do so, the receiving modem (104) counts the number of valid data frames (130) it receives during a given time interval, and the receiving modem (104) also counts the number of data frames the modem (104), via the error-correcting protocol, requests to be retransmitted (132) and/or are ignored during the same time interval. The ratio (136) of the count of the good frames (130) received to the count of those that were ignored/retransmitted (132) provides a measurement of the data bandwidth being lost due to error correction. The ratio is then compared (140) to a threshold value (138) and, if the ratio (136) exceeds the threshold value (138), the data rate is adjusted accordingly.

Description

SYSTEM AND METHOD FOR DATA RATE ADJUSTMENTS IN MODEM COMMUNICATIONS

Field of Invention The invention generally relates to the transmission of information between modems or other communication devices, and specifically the invention relates to determining the best data rate for a particular connection.

Background

In modem communications, modems transmit information between one another at a data rate, typically measured in bits per second (bps). Often, however, it is desirable to change the data rate during a particular connection, particularly to reduce the bit error rate or to improve data throughput (the number of good bits received per unit time). Newer modulation protocols

(e.g., ITU-T V.32 or V.34) utilized by modems will permit a data rate change mid-connection.

To determine the data rate for a connection, most commonly the received carrier signal is examined and a bit error rate is determined, often in the form of a "Quality Measure" (QM), a unitless indicator of the quality of a connection. Calculation of QM will vary among modem vendors but will be generally known to those of skill in the art. Typically, however, the received carrier signal is examined and various mathematical functions are applied to determine a QM. For instance, some vendors utilize quadrature amplitude modulation, which includes examining the sine wave carrier, periodically phase shifting the sine wave, and plotting the amplitude at various times in a polar coordinate system. These calculations will generally yield a statistical approximation of the bit error rate of the signal.

The ultimate goal of producing a Quality Measure is to give a numerical representation of the quality of a particular connection. For instance, some vendors use a QM range of 0 (representing a perfect connection with no errors) to 127 (indicating a large number of bit errors and that the connection is about to be disconnected). Further, the modem vendor often provides a threshold QM value, e.g., a QM of 30. The threshold QM indicates that if a QM of higher than the threshold is measured for the connection, the data rate should be lowered to reduce bit errors. If a QM below the threshold is measured, the data rate for the connection should be increased or remain the same.

Unfortunately, using QM, i.e., using the carrier to derive a statistical approximation of the bit error rate, does not allow for the selection of the most optimum data rate for the connection in terms of data throughput. First, being a statistical approximation, QM does not always accurately reflect the bit error rate. Second, with error correcting protocols such as ITU-T V.42 (LAPM) or ITU-T V.42 alternative protocol, a data rate with a relatively high error rate may be selected using QM because the error correcting protocols correct the errors. Nonetheless, a connection that maximizes data throughput is generally more desirable over one that just minimizes bit errors at a particular data rate. For instance, with some connections, the optimum data throughput may occur when a data rate with a lower bit error rate is maintained, while with other connections the optimum throughput may occur at a higher data rate with a higher bit error rate. The goal is ultimately to get the most good data received in any given time interval. Third, many modem vendors choose a fairly low threshold QM, e.g., 25-30. Nonetheless, the inventor of the present invention has discovered that often using a data rate yielding a QM well above the threshold QM results in a higher throughput that still has acceptable performance. Thus, Quality Measure is not a reliable indicator for achieving optimum data throughput rate. Moreover, calculating QM can consume a noticeable percentage of

DSP resources. In many cases, the percentage will be significant enough to require more powerful DSPs to obtain the desired performance, adding to the modem cost.

Therefore, it is desirable to determine an improved method for finding a data rate which will optimize data throughput for a given modem connection.

Summary of the Invention Rather than use the carrier signal to determine the need for data rate adjustments, a method in accordance with the invention takes advantage of the error correction protocol utilized by the modem to determine a measurement of the data bandwidth used for error correction and compares that measurement to a threshold value. The comparison results in an indication of whether data rate needs to be adjusted to maximize throughput. Specifically, to determine a measurement of the data bandwidth used for error correction, one embodiment of the invention counts the number of good, or valid, frames received during a given time interval. During the same time interval, the number of frames that are ignored and/or are required to be retransmitted for error correction are also counted. A ratio of the number of good frames to the number of ignored/retransmitted frames serves as the measurement of data bandwidth used for error-correction. The measurement is compared to a threshold value. In one embodiment, the threshold value is predetermined. In a second embodiment, the threshold value is the difference between the present data rate and a possible new data rate. If the measurement exceeds the threshold value a change in data rate is indicated.

A system and method in accordance with the invention is advantageous in that it produces a more accurate and cost effective method of determining an optimum data rate for maximum data throughput than conventionally used methods.

Brief Description of the Drawings The present invention is described with respect to particular exemplary embodiments thereof and reference is accordingly made to the drawings, which are not necessarily drawn to scale, in which:

Fig. 1 illustrates in a generalized representational block diagram of data transmission between a transmit and a receive modem, including a block diagram representation of a frame structure; and Fig. 2 illustrates the steps to be performed in accordance with one embodiment of the present invention.

Detailed Description of the Invention A method in accordance with the invention allows the error-correcting protocol utilized by the modem to select the data rate that will maximize data throughput. To do so, the receiving modem counts the number of valid data frames it receives during a given time interval, and the receiving modem also counts the number of data frames the modem, via the error-correcting protocol, requests to be retransmitted and/or frames that are ignored during the same time interval. The ratio of the count of the good frames received to the count of those that were retransmitted and/or ignored provides a measurement of the data bandwidth being lost due to error correction. The ratio is then compared to a threshold value and, if the ratio exceeds the threshold value, the data rate is adjusted accordingly. More specific details of the invention are discussed below.

Figure 1 shows a transmit modem 102 and a receive modem 104. In practice, each modem is capable of both receive and transmit functions, but for the convenience of this discussion are here referred to separately as a transmit and a receive modem. Information is generally sent between the modems 102, 104 in accordance with a standardized protocol such as V.42, propounded by the ITU-T and incorporated by reference herein. V.42 is an error-correcting protocol. The error-correcting function of V.42 is also commonly referred to as Link Access Procedure for Modems or "LAPM'. An alternative function to LAPM and also specified in V.42 is referred to herein as the "V.42 alternative protocol." Both LAPM and V.42 alternative protocol require data and other information to be transmitted in a structure called a frame.

Fig. 1 shows a generic V.42 frame 110. Such a frame contains an opening flag field 112, an address field 114, a control field 116, an information field 118, a frame check sequence (FCS) field 120, and a closing flag field 122. The flag fields, 112 and 122, delimit the frame by using a predetermined unique bit pattern, "01111110." The address field 114 generally identifies the error- correcting connection and the error-correcting entity associated with the connection. The FCS field 120 is used to guard against bit errors in transmission and contains a Cyclic Redundancy Check (CRC) polynomial used to evaluate if an error has occurred. The control field 116 is generally used to distinguish between different frame types (e.g. , information frames, supervisory frames, and unnumbered frames^'). The information field 118 is the field into which the data or command information is inserted.

In V.42, different frame types are used to send different types of information. "Information" frames are used to send data packets. "Supervisory" frames are used to transmit control information. "Unnumbered" frames are used to transmit additional information (data, control signals) but are not used for general data packet transmission. As used herein, the term "frame" (unless indicated otherwise by context) generally refers to sequentially numbered frames, and, in the context of the description, will most commonly be information frames.

Information and Supervisory frames are assigned sequence numbers that appear in control field 116. The sequence numbers aid in keeping track of data transmitted and/or received. As defined in V.42, the sequence numbers range from 0-127. After receiving a frame numbered 127, the sequence numbers for the next frames are 0, 1, 2.... Such sequence numbers are thus sometimes referred to as "modulus" sequence numbers. Keeping track of such modulus sequence numbers is well understood in the art.

In conventional systems, the number of transmitted and/or retransmitted frames is generally only known to the transmitting modem and not the receiving modem (i.e., the receiving modem has no way of determining if a frame is sent if the frame is not received). Further, if a frame does contain an error, the receiving modem can no more make sense of such a frame than if the receiving modem had received noise. Such erroneous frames are simply ignored by the receiving modem. Nonetheless, in accordance with the invention the receiving modem determines both how many frames received are good as well as how many frames had to be retransmitted and/or were ignored, and the receiving modem does so by taking advantage of a sequential numbering system such as that used by LAPM in one embodiment of the invention. Referring to Figure 2, as a frame is received, step 210, the frame is checked for errors by running a calculation on a CRC polynomial in the FCS field 120 of the frame 110 (Figure 1) as is generally known in the art. If the frame contains an error, the frame is ignored, step 222 and the process returns to step 210 to receive the next frame. When a frame is ignored, the frame is simply discarded and no acknowledgment of the frame is sent to the transmitting modem as is done when good frames are received. If the received frame is good, step 220, the sequence number of the frame is verified, step 223. If the sequence number is the expected sequence number, then a counter 130, representing a "good count" increments by one, step 225. The counter can be implemented in software, hardware, and/or firmware as is known in the art. In step 223, if a frame is discovered missing, i.e., the expected sequence number is not received (such as when a frame has been ignored), the modem 104 requests retransmission of the missing frame or frames, step 235, as is known in the art, e.g., in accordance with V.42.

Referring again to Figure 1, suppose (for example) that frames sequentially numbered 10-14 are transmitted by modem 102 and received by modem 104, where both modems operate in accordance with V.42. Modem 104 performs a calculation on the CRC polynomial in the FCS field 120 of each frame 10-14 as each frame is received. If modem 104 determines that frames 10-13 are error-free, but that frame 14 contains a CRC error, the receive modem 104 will ignore frame 14. Upon receipt of frame 15, the receive modem will recognize that a frame is missing and request retransmission of frame 14 following procedures specified by V.42, step 230 (Figure 2).

The number of frames retransmitted by the transmitting modem 102, however, depends on whether the modems are in selective reject (RET) mode or selective repeat (SREJ) mode, both of which are alternative optional modes of operation specified by V.42. In the selective repeat mode (SREJ), or "retransmit N" mode, if frame 14 is requested to be retransmitted, the transmit modem retransmits frame 14 only.

Alternatively, in REJ mode, or "go back N" mode, the transmitting modem transmits the frame requested to be retransmitted, e.g., frame 14, plus all frames following frame 14 that were already transmitted, e.g., frames 15-17. As will be appreciated, because of time required to receive a particular frame (network transmission delays), time required to calculate if an error exists and discover a missing frame, and time required to request retransmission of a frame, several frames subsequent to the erroneous frame (frame 14) may have already been transmitted by the transmitting modem 102 by the time receiving modem 104 can request retransmission. If, for instance, frames 15-17 have been transmitted, and if the modems are operating in REJ mode, frames 14-17 will be retransmitted. In accordance with one embodiment of the invention, the receiving modem 104 counts the number of retransmitted frames, step 235 (Fig. 2), in a counter 132. Counter 132 can be implemented in software, hardware, and/or firmware as is known by those of skill in the art. If the modems are operating in REJ mode, the receiving modem 104 counts in counter 132 the frame that caused the retransmission request and any other frames received until the desired frame, e.g., frame 14, is received. In the above example, counter 132 would have counted four retransmitted frames (frames 14-17). If the modem is operating in SREJ mode, then the receiving modem simply counts, in counter 132, the number of times it sends a retransmission request. Of course, similar methods of counting can be implemented using other error correcting protocols, such as V.42 alternative protocol, and the scope of the invention should not be construed to be limited to LAPM or even V.42.

After a predetermined time interval (e.g., 5 seconds) has passed as determined by a timer 134, step 240, the number of good frames received, as indicated by counter 130, and the number of retransmitted frames received, as indicated by the counter 132, are compared in the form of a ratio 136 to a threshold value 138. The threshold 138 can be fixed or configurable in different embodiments of the invention. The comparison in one embodiment of the invention is made by comparator 140. If the ratio 136 of good frames received to retransmitted frames exceeds the threshold 138, then the data rate is adjusted accordingly, step 260. In some embodiments of the invention, the ratio 136 exceeds a threshold (threshold 1) if the ratio is greater than the threshold, step 250, and the data rate is lowered accordingly, step 260. In other embodiments of the invention, the ratio exceeds a threshold (threshold 2) if the ratio is less than the threshold 260, and the data rate will be raised in step 260. Still other embodiments provide for both of the comparisons of steps 250 and 252. If the appropriate time interval has not passed, step 240, the process returns to step 210 to receive the next frame. If in steps 250 and/or 252 the retransmission-count to good-count ratio does not exceed the threshold value, or once the data rate has been adjusted in step 260, the respective counters are reset, step 262, and the process returns to step 210. As with counters 130, 132, any of elements 134, 136, 138, and 140 can be implemented in hardware, software, and/or firmware as is known in the art.

The actual method of adjusting the data rate, step 260, will depend upon the modulation method being utilized for the connection. For instance, V.34 supports rate renegotiations, retrains, and seamless rate changes, each of which is known in the art.

A method in accordance with the invention also accounts for the situation where an erroneous frame is received, but no frames follow (i.e., no missing frames are identified in step 223). In many embodiments of the invention, such a situation will be of no consequence: if frames are not being sent, there is generally no need to adjust the data rate. Nonetheless, as is known in the art, the transmitting modem contains a timer. If the transmitting modem does not receive an acknowledgment of the missing frame during a given time period, the transmitting modem will query the receive modem, ultimately prompting the retransmission of the frame. Thus, some embodiments of the invention will include "ignored" frames in the retransmission count by monitoring these queries. An alternative embodiment of a method in accordance with the invention is to compare the data bandwidth being used for error correction to the change in the data rate that will occur if the data rate is lowered. In other words, the change in data rate is used as a threshold value. If the bandwidth being used for error correction exceeds the change in data rate that will occur, then the data rate can be adjusted to increase data throughput.

Several other alternative embodiments of a method in accordance with the invention exist. For instance, rather than measuring the bandwidth used for error-correction over a single time interval, a running average over several time intervals could be utilized to compare with the threshold. In addition, several thresholds could be utilized, i.e., comparison to several thresholds could indicate an adjustment of data rates up or down by more than one data rate step.

It should also be apparent to those of skill in the art, that while measurement of the number of good data frames to ignored/re-transmitted data frames is described, a method in accordance with the invention could be utilized to count good bytes to retransmitted/ignored bytes, or even good bits to retransmitted/ignored bits. In other words, the information unit, i. e. , the unit of measure for delineating received information, should not be construed to limit the scope of the invention. Finally, while the described embodiment of the invention operated during data transfers, other embodiments of the invention may be utilized during modem idle time. For instance, when no data is being sent, several frames could be sent solely for a data rate adjustment determination. Thus, a cost-effective and accurate technique has been disclosed for evaluating the data rate of a given connection. Using a measurement of the bandwidth that is used by the error-correction protocol to decide if the data rate should be adjusted will increase the performance of modems relative to those not using the technique described above but will not tax DSP resources.

It should be understood that the particular embodiments described above are only illustrative of the principles of the present invention, and various modifications could be made by those skilled in the art without departing from the scope and spirit of the invention. Thus, the scope of the present invention is limited only by the claims that follow.

Claims

CLAIMS What is claimed is: 1. A method, for use by a receiving modem, of determining a data rate for optimizing data throughput, comprising the steps of:

receiving a plurality of information units at a data rate;

determining, by said receiving modem, a measurement of bandwidth used for error correction;

evaluating, by said receiving modem, whether said measurement of bandwidth exceeds a threshold value; and

adjusting the data rate.

2. The method of claim 1, wherein said step of determining includes:

determining, by the receiving modem, the number of information units said plurality of information units that are good; and

determining, by the receiving modem, the number of information units in said plurality of information units that require retransmission.

3. The method of claim 2, wherein said step of evaluating includes:

comparing a ratio of the number of information units required for retransmission and the number of good information units to a threshold value.

4. The method of claim 3, wherein an information unit is a data frame.

5. The method of claim 4, wherein each data frame is in a V.42 frame structure.

6. The method of claim 3, wherein said threshold value is predetermined.

7. The method of claim 3, wherein said threshold value is configurable.

8. The method of claim 3, wherein said threshold value is a difference in said data rate and a new intended data rate.

9. The method of claim 1, wherein the step of evaluating includes evaluating whether said bandwidth has exceeded said threshold by being more than said threshold value.

10. The method of claim 1, wherein the step of evaluating includes evaluating whether said bandwidth has exceeded said threshold by being less than said threshold value.

11. A method for use by a receiving modem of determining a data rate for optimizing data throughput, comprising the steps of:

receiving a plurality of data frames at a data rate;

comparing a ratio of good data frames and retransmitted data frames to a threshold value; adjusting said data rate.

12. The method of claim 11, further including the steps of:

counting a number of good data frames received in said plurality of data frames;

counting a number of data frames requested by said receiving modem to be retransmitted; and

forming said ratio with said number of good data frames and said number of data frames requested.

13. The method of claim 12, wherein said plurality of data frames are in a V.42 frame structure.

14. A modem, comprising:

an interface coupled to receive a plurality of data frames at a data rate;

a bandwidth determiner operable to determine a measurement of data bandwidth used for correcting errors with said plurality of data frames;

an evaluator operable to determine whether said data bandwidth exceeds a threshold value; and a data rate adjuster operable to adjust said data rate in response to said evaluator.

15. The modem of claim 14, wherein said bandwidth determiner includes:

a good-frame counter to obtain a good-count;

a retransmitted-frame counter to obtain a retransmission-count.

16. The modem of claim 15, wherein said evaluator includes:

a comparator coupled to receive a ratio of said good-count and said retransmission-count and coupled to receive said threshold value.

17. A modem, comprising: ΓÇó

an interface coupled to receive a plurality of data frames at a data rate;

a good-frame counter to obtain a good-count;

a retransmitted-frame counter to obtain a retransmission-count;

a comparator to compare a ratio of said good-count and said retransmission-count to a threshold value;

a data rate adjuster operable to adjust the data rate at which the modem is transmitting in response to said comparator.