JP3540225B2 - A device for measuring the fluctuation component of delay time in a packet switching network. - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a delay time fluctuation component measuring apparatus in a packet switching network that measures a delay time fluctuation component (jitter or fluctuation) in a transmission path based on a difference between a transmission time and a reception time of a test packet. In particular, the present invention relates to a delay time fluctuation component measuring apparatus in a packet switching network capable of accurately measuring a delay time fluctuation component regardless of the accuracy error of each of the clock timers on the transmission side and the reception side of a test packet.
[0002]
[Prior art]
Since the delay time and its fluctuation component in the packet switching network are used to measure the congestion state and capacity of the transmission path, a technique for accurately measuring these is desired. In the conventional delay time measurement, a test packet is transmitted from a transmitting terminal connected to one end of a transmission path to be measured. At this time, the transmission time Ts measured by the clock timer of the transmission side terminal is recorded in the test packet. Upon receiving the test packet, the receiving terminal measures the reception time Tr with its own clock timer, and based on the difference (Tr−Ts) between the transmission time Ts registered in the test packet and the reception time Tr. Then, the delay time and its fluctuation component are obtained.
[0003]
[Problems to be solved by the invention]
The time Ts by the time timer on the sending side terminal and the time Tr by the time timer on the receiving side terminal do not always match, and the time given by each time timer has a time difference, and the time difference changes with the passage of time. Therefore, as in the prior art described above, the time difference between the transmission time Ts registered at the transmitting terminal transmitting the test packet and the reception time Tr measured at the receiving terminal receiving the test packet includes a delay time. Not only the error component due to the accuracy error of each time timer is included. Therefore, there is a problem that the delay time cannot be accurately represented by the measured time difference.
[0004]
SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems of the prior art, and to set a transmission time registered in a test packet on a test packet transmission side and a reception time of the test packet measured on a test packet reception side. Accordingly, it is an object of the present invention to provide a delay time fluctuation component measuring device in a packet switching network that can accurately measure the delay time of a transmission path based on the above.
[0005]
[Means for Solving the Problems]
In order to achieve the above object, the present invention provides a test packet sending means including a sending side time timer, registering the sending time in a plurality of test packets, and sending the same to a transmission line to be measured. A test packet receiving means for receiving the test packet from the transmission path to be measured and detecting a reception time thereof; a transmission time extracting means for extracting a transmission time registered in each of the received test packets; Total delay time detecting means for detecting a total delay time based on a time difference between the extracted transmission time and the detected reception time for each test packet, and a total delay calculated for each test packet. Based on the time, of the total delay time, a time error component and a measurement object caused by a relative accuracy error between the sending side and the receiving side time timers. A non-variable component function detecting means for detecting a time function of a non-variable component that is a sum of fixed delay components specific to the transmission path; and a total delay time based on the total delay time and the time function of the non-variable component. And a fluctuation component extracting means for extracting a fluctuation component included in the data.
[0006]
According to the above feature, even if there is a time error between the time timer of the sending terminal and the time timer of the receiving terminal, the non-variable component of the delay time including the time error is obtained as a non-variable component function. If the calculated value of the non-fluctuation component function is subtracted from each total delay time, the fluctuation component of the delay time in the transmission path, that is, the jitter, can be accurately measured regardless of the time error.
[0007]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of a first embodiment of a delay time fluctuation component measuring apparatus to which the present invention is applied.
[0008]
In the sending terminal 2, the test packet sending unit 21 includes a sending time timer 21 a that independently counts the current time. When sending a plurality of test packets TP to the transmission path 1 to be measured, the test packet sending unit 21 counts the time. The current time is registered in each test packet TP (i) as the transmission time Ts (i).
[0009]
In the receiving terminal 3, the test packet receiving unit 31 includes a receiving time timer 31 a that independently counts the current time, and when each of the test packets TP (i) is received from the measurement target transmission line 1, The current time is detected as the reception time Tr (i) for each test packet TP (i). The transmission time extracting unit 32 extracts the transmission time Ts (i) of each test packet TP (i) registered in the received test packet TP (i).
[0010]
The total delay time detector 33 calculates the time difference [= reception time Tr (i) −transmission time between the transmission time Ts (i) and the reception time Tr (i) extracted for each test packet TP (i). Ts (i)] is detected as the total delay time D (i) of the transfer path 1 to be measured. As described in detail later, the total delay time D (i) includes a fixed delay time which is determined almost fixedly for each transmission path 1, a jitter component which fluctuates according to the congestion state of the transmission path, and the like. And a time error component caused by the time difference between the time timers 21a and 31a on the receiving side.
[0011]
The non-fluctuation component function detection unit 34 calculates the total delay time D (i) obtained for each test packet TP (i) and the transmission time Ts (i) of each test packet TP based on the relationship. As described in detail below, a time function F (i) of the non-variable component included in the total extension time D (i) is obtained.
[0012]
FIG. 2 is a distribution diagram plotting the relationship between the total delay time D (i) obtained for each test packet TP (i) and the transmission time Ts (i) of each test packet TP.
[0013]
Normally, the fixed delay time of the transmission line 1 is almost constant, and the jitter component fluctuates irregularly. Therefore, if the time of each of the time timers 21a and 31a on the transmission side and the reception side is exactly the same, , The plots should be distributed almost horizontally. On the other hand, in the present embodiment, since the plots are distributed to the upper right, the receiving-side time timer 31a advances faster than the sending-side time timer 21a.
[0014]
In addition, since the total delay time D (i), which normally cannot show a negative value, shows a negative value immediately after the start of measurement, and then shows a positive value after showing almost zero. The relative time of the time timers 21a and 31a is that the sending time timer 21a has advanced at first, but the receiving time timer 31a advances faster than the sending time timer 21a as described above. It indicates that the side time timer 21a advances and the time difference between the two increases.
[0015]
Here, the total delay time D (i) varies according to the fixed delay time Kt fixed to the transmission line 1 and the congestion state of the transmission line 1 as shown in Expression (1). A jitter component Z (i) and a time error component ΔTt (i) of each of the time timers 21a and 31a described above are included.
D (i) = Kt + Z (i) + ΔTt (i) (1)
Of these, the time error component ΔTt (i) uniformly increases or decreases. Therefore, if the plot distribution shown in FIG. 2 is represented by a linear function f (i), the slope of the plot distribution indicates the measured time of each of the time timers 21a and 31a. As a representative of the ratio, a deviation amount of the total delay time D (i) from the linear function f (i) becomes a jitter component Z (i). Since the jitter component Z (i) cannot take a negative value, the regression line along the lower limit of the plot distribution shown in FIG. 2 is defined as the time function F (i) of the non-variable component. For example, the difference [D (i) -F (i)] between the time function F (i) and the total delay time D (i) corresponds to the jitter component Z (i).
[0016]
Therefore, in the present embodiment, the time function F (i) in FIG. 2 is obtained by using the following algorithm. First, the basic concept of the present algorithm will be described with reference to FIG. 3. Procedure 1: The total delay time D (1) having the minimum transmission time Ts (i) is set to the left end, and the other total delay times D (2 ) To D (13), D (j) having the minimum inclination among all the straight lines having the right end.
[0017]
Step 2: D (j) obtained in step 1 is set as a new right end,
Step 3: Repeat step 2 until the right end reaches the last point.
[0018]
Step 4: Among the straight lines obtained as described above, the straight line having the largest time difference between the left end and the right end is defined as the above-described non-variable component time function F (i).
[0019]
Next, the algorithm according to the present embodiment will be described with reference to the flowchart in FIG.
[0020]
In step S1, "0" is set as an initial value in a "longest time difference ΔTmax" (corresponding to the maximum value of the time difference between the left end and the right end) described later. In step S2, "1" is set to "left end time L" that defines the left side of the two plots whose inclinations are to be detected.
[0021]
In step S3, a value obtained by adding "1" to the left end time L is set to "right end time R" defining the right side of the two plots whose inclinations are to be detected. Then, a slope A of a straight line connecting the two plots D (L) and D (R) is obtained, and this is set as a provisional minimum value Amin of the slope A. Further, the value of R is set as an initial value at "minimum right end time Rmin" defining the right side of the two plots whose inclination indicates the minimum value Amin, and the L is set at "minimum left end time Lmin" defining the left side. Is set as the initial value.
[0022]
In step S4, a slope A of a straight line connecting two plots D (L) and D (R) to be detected this time is obtained. In step S5, a tentative minimum value Amin up to the present regarding the slope A is compared with the current slope A. If the current slope A is not smaller than the tentative minimum value A, in step S6, the detection target The right end time R is incremented in order to update the right side plot of. In step S7, it is determined whether or not a new right plot exists, and if so, the steps S4 to S4 are performed to determine the slope of a straight line connecting the new right plot and the left plot D (L). Each process of S7 is repeated.
[0023]
At this time, when the gradient A smaller than the tentative minimum value Amin is detected in the step S5, the right end time R and the left end time L at that time are respectively changed to the minimum right end time Rmin and the minimum right end time Rmin in the step S8. It is registered as the minimum left end time Lmin.
[0024]
As described above, when the right side plot D (Rmin) whose slope shows the minimum value when the left side plot is D (1) is obtained, in step S9, the transmission time difference ΔT [= Ts (Rmin) − Ts (Lmin)].
[0025]
In step S10, the transmission time difference ΔT at the time when the gradient is the minimum with respect to the current left end plot is compared with the longest time difference ΔTmax of the transmission time difference up to the present time. In S11, the value of the minimum right end time Rmin is set to the left end time L in order to set the next left end plot to the current right plot position. In step S12, it is determined whether or not a plot exists on the right side of the new left plot, and if so, the process returns to step S3 and the above-described processes are repeated.
[0026]
On the other hand, in step S10, if the current time difference ΔT is longer than the longest time difference ΔTmax so far, the process proceeds to step S13. In step S13, the "longest left end time Lmax" defining the left side of the two plots having the longest transmission time difference is updated to the value of the minimum left end time Lmin, and the "longest right end time Rmax" defining the right side is updated. The value is updated to the value of the minimum right end time Rmin, and the longest time difference ΔTmax is updated to the value of the transmission time difference ΔT.
[0027]
Returning to FIG. 1, when the left plot D (Lmax) and the right plot D (Rmax) whose slope shows the minimum value and the time difference shows the longest value are obtained as described above, the non-variable component function detection unit 34 Then, a linear function passing through each plot D (Lmax), D (Rmax) is obtained, and this is registered as the time function F (i) of the non-variable component. The fluctuation component (jitter) extraction unit 35 subtracts the operation value of the time function F (i) from each of the total delay times D (i), as shown in FIG. 5, to extract only the jitter component Z (i). Selectively detect.
[0028]
As described above, according to the present embodiment, even if there is a time error between the time timer 21a of the sending terminal 2 and the time timer 31a of the receiving terminal 3, the transmission timer registered in each test packet TP (i). Based on the time Ts (i) and the reception time Tr (i) of the test packet measured on the receiving side of the test packet, the delay time fluctuation component, that is, the jitter Z (i) in the transmission path 1 is accurately determined. Can be measured.
[0029]
FIG. 6 is a block diagram of a second embodiment of the present invention, wherein the same reference numerals as those described above denote the same or equivalent parts.
[0030]
In the first embodiment described above, the non-variable component function F (i) of delay time and the jitter Z (i) are obtained based on all the measured total delay times D (i). The total delay time D (i) also includes illegal data that has been specifically and significantly delayed due to an accidental delay cause, and it is desirable to remove such illegal data in advance. Therefore, in the second embodiment of the present invention, the measurement result selection unit 36 is added to remove invalid data based on the correlation between the detected total delay times D (i).
[0031]
FIGS. 7 and 8 are diagrams schematically showing a method of eliminating invalid data by the measurement result selection unit 36. In this example, two arbitrary continuous total delay times D (j-1) and D (j ) And remove the improper data. In the present embodiment, the limit value at which the sending-side time timer 21a can be delayed with respect to the receiving-side time timer 31a and the limit value at which the receiving-side time timer 31a can be delayed with respect to the sending-side time timer 21a have slopes of ±. It is set in advance as a linear function of Dm.
[0032]
Here, as shown in FIG. 7, the total delay time D (j) is different from the total delay time D (j-1) by the allowable delay time ΔDref [= ± Dm · ｛between the time timers 21a and 31a. Ts (j) -Ts (j-1)}], the total delay time D (j) can be regarded as invalid data.
[0033]
Similarly, as shown in FIG. 8, if the total delay time D (j-1) is delayed from the total delay time D (j) by more than the allowable delay time ΔDref, the total delay time D (j-1 ) Can be considered invalid data.
[0034]
Therefore, in the present embodiment, two consecutive total delay times D (j−1) and D (j) or two arbitrary total delay times D (i) and D (j) are determined according to the algorithm shown in the flowchart of FIG. The invalid data is excluded based on whether or not the correlation shown in (j) is the relationship shown in FIGS. In the flowchart of FIG. 9, an example will be described in which invalid data is excluded based on the correlation between two consecutive total delay times D (j−1) and D (j).
[0035]
In step S31, an initial value 2 is set to a variable j for defining two samples D (j-1) and D (j) to be sorted. In step S32, the difference [Ts (j) between the coefficient Dm and the sending time of each of the samples D (j-1) and D (j) in consideration of the time difference between the time timers 21a and 31a on the sending and receiving sides. ) −Ts (j−1)] to calculate the allowable delay time ΔDref.
[0036]
In step S33, a difference value obtained by subtracting the sample D (j-1) from the sample D (j) is compared with the allowable delay time ΔDref. If the difference value exceeds the allowable delay time ΔDref, each sample D ( j-1) and D (j) are determined to be as shown in FIG. 7, and in step S34, the sample D (j) is excluded as invalid data.
[0037]
In step S35, the difference value obtained by subtracting the sample D (j) from the sample D (j-1) is compared with the allowable delay time ΔDref. If the difference value exceeds the allowable delay time ΔDref, each sample is compared. It is determined that the relative relationship between D (j-1) and D (j) is as shown in FIG. 8, and in step S36, the sample D (j-1) is excluded as invalid data.
[0038]
In step S37, the variable j is incremented. In step S38, it is determined whether or not the determination process for all the samples has been completed. If not, the process returns to step S32 and the above-described processes are repeated.
[0039]
As described above, when the elimination of the improper total delay time D (i) is completed, the non-variable component function detection unit 34 determines the non-variable component function F (i) based on the remaining total delay time D (i). , And the fluctuation component extraction unit 35 extracts the jitter Z (i).
[0040]
As described above, according to the present embodiment, data that has been incorrectly delayed due to an accidental delay cause is removed in advance, so that a fluctuation component of the delay time can be obtained at high speed.
[0041]
【The invention's effect】
According to the present invention, even if there is a time error between the time timer of the sending terminal and the time timer of the receiving terminal, the sending time registered in the test packet on the sending side of the test packet and the receiving time of the test packet Based on the reception time of the test packet measured on the side, the fluctuation component of the delay time in the transmission path, that is, the jitter can be accurately measured.
[Brief description of the drawings]
FIG. 1 is a block diagram of a first embodiment of the present invention.
FIG. 2 is a diagram plotting a time distribution of a total delay time.
FIG. 3 is a diagram schematically showing a method of detecting a non-variation component function F (i) of a delay time.
FIG. 4 is a flowchart illustrating a method of detecting a non-variable component function F (i).
FIG. 5 is a diagram showing a detection result of a jitter component.
FIG. 6 is a block diagram of a second embodiment of the present invention.
FIG. 7 is a diagram for explaining a method for eliminating illegal data.
FIG. 8 is a diagram for explaining a method for eliminating illegal data.
FIG. 9 is a flowchart showing a method for eliminating illegal data.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Transmission path, 2 ... Sending terminal, 3 ... Receiving terminal, 21 ... Test packet sending part, 21a ... Sending time timer, 31 ... Test packet receiving part, 31a ... Receiving time timer, 32 ... Sending time extraction Section, 33: total delay time detecting section, 34: non-variable component detecting section, 35: variable component extracting section, 36: measurement result selecting section

Claims (4)

A test packet sending means including a sending side time timer, registering the sending time in a plurality of test packets, and sending the registered time to the transmission path to be measured;
A test packet receiving means including a receiving side time timer, receiving the test packet from the transmission line to be measured, and detecting the reception time,
Sending time extracting means for extracting the sending time registered in each received test packet;
For each received test packet, a total delay time detecting means for detecting a total delay time based on a time difference between the extracted transmission time and the detected reception time,
Based on each total delay time obtained for each test packet, a time error component and a measurement target transmission of the total delay time due to a relative accuracy error between the sending side and receiving side time timers. Non-variable component function detecting means for detecting a time function of a non-variable component that is a sum of fixed delay components specific to the road;
A variable component extracting means for extracting a variable component included in each total delay time based on the total delay time and a time function of the non-variable component. . Comparing a first total delay time determined for one test packet with a second total delay time determined for another test packet, the first total delay; Measuring time selecting means for judging the first total delay time as an invalid measurement result when the time is not in a predetermined relationship with the second total delay time;
The non-variable component function extracting unit and the variable component extracting unit determine the non-variable component function and the variable component, respectively, based on each of the total delay times not determined as the invalid measurement result. 2. The delay time fluctuation component measuring device in the packet switching network according to 1. 3. The delay in the packet switching network according to claim 2, wherein the schedule relationship is determined based on a limit value of a value predicted as a time error between the time timers on the receiving side and the sending side. Time fluctuation component measurement device. The non-variable component function detecting means sets a regression line representing a lower limit of a time distribution of each total delay time obtained for each of the test packets as a time function of the non-variable component. Item 4. An apparatus for measuring a fluctuation component of delay time in a packet switching network according to any one of Items 1 to 3.

Images